Renewable and Sustainable Energy Reviews

Renewable and Sustainable Energy Reviews 155 (2022) 111955
Contents lists available at ScienceDirect
Renewable and Sustainable Energy Reviews

journal homepage: www.elsevier.com/locate/rser
An overview of the challenges of solar power integration in isolated

industrial microgrids with reliability constraints
Louis Polleux a, b, *, Gilles Guerassimoff a, Jean-Paul Marmorat a, John Sandoval-Moreno b,
Thierry Schuhler b
a
Center for Applied Mathematics, Mines Paristech PSL, Sophia Antipolis, France
b
Research & Development, Total Energies, Paris, France
A R T I C L E I N F O A B S T R A C T
Keywords: Reducing carbon emissions and electricity costs in industry is a major challenge to ensure competitiveness and
Microgrids compliance with new climate policies. Photovoltaic power offers a promising solution but also brings consid
Renewable integration erable uncertainties and risks that may endanger the continuity and quality of supply. From an operational point
Industry
of view, large-scale integration of solar power could result in unmet demand, electrical instabilities and
Power quality
equipment damage. The performance and lifetime of conventional fossil equipment are likely to be altered by
repeated transient operations, making it necessary to adopt specific modeling tools. Control strategies and sizing
methodologies must be adapted to account for the strong reliability constraint while dealing with significant
production uncertainties. In addition, conventional mitigation technologies, such as storage and load flexibility,
have limited potential in these applications and may result in high investments or penalties if they are not
properly assessed. This study provides an overview of these challenges by providing a transversal analysis of the
scientific literature from fossil engine thermodynamics to control system theory applied to industrial systems.
The main characteristics of reliability-constrained microgrids are identified and a conceptual definition is pro
posed by analyzing state-of-the art studies of various industrial applications and taking oil-and gas microgrids as
an enlightening example. Then follows a review of the challenges of accounting for dynamical behavior of fossil
equipment, PV and storage systems, ending with the identification of several research gaps. Finally, applicable
control strategies are presented.
these applications, microgrids will play a key role.

1. Introduction One of the main advantages of microgrids is undoubtably the ability
to manage renewable energy resources as well as storage and conven
The beginning of the 2020 decade, subject to a combination of tional fossil generation to ensure the right trade-off between costs,
climate change mitigation and poverty alleviation, faces the world with reliability and sustainability [7,8]. Microgrids now cover a wide variety
a double challenge that needs to be tackled as fast as possible [1,2]. of uses, from grid-connected systems able to sell and buy electricity
Establishing an efficient low-carbon industry is a priority to support depending on the market price, to robust and isolated systems ensuring
socio-economic development while meeting Paris Agreement targets. continuous power for mining or military facilities. Examples of such
In 2016, the industry sector accounted for 36% of global CO2 systems can be found in Refs. [9,10].
emissions, mainly due to its high energy-intensive processes [3]. The IEA Microgrid studies mainly focus on either investment planification or
continues to highlight that power facilities operating with low-carbon operational management with an optimization philosophy in order to
electricity constitute one of the three main levers to achieve a sustain find a compromise between three main objectives [11]: providing
able transition [4]. electricity at the lowest cost or carbon footprint possible; ensuring the
Facilities located in developed countries can substantially reduce safe supply of electricity to deliver all consumers at the required time;
their GHG emissions using process electrification with green electricity and guaranteeing the quality of supply at the delivery point.
purchase agreements [5]. However, in numerous applications, plants are This optimization paradigm is the cornerstone of a microgrid’s per
either connected to an unreliable grid or totally isolated from it [6]. In formance, from early developments to the plant’s real-time
* Corresponding author. Center for Applied Mathematics, Mines Paristech PSL, Sophia Antipolis, France.
E-mail address: louis.polleux@mines-paristech.fr (L. Polleux).
https://doi.org/10.1016/j.rser.2021.111955
Received 9 February 2021; Received in revised form 3 November 2021; Accepted 24 November 2021
Available online 1 December 2021
1364-0321/© 2021 Elsevier Ltd. All rights reserved.
L. Polleux et al. Renewable and Sustainable Energy Reviews 155 (2022) 111955
List of abbreviations IEA International energy agency

ISO International Organization for Standardization
ASAI Average Service Availability Index LCOE Levelized costs of electricity
ASUI Average Service Unavailability Index LNG Liquefied natural gas
CAIDI Customer Average Interruption Duration Index LP Linear programming
CCGT Combined cycle gas turbine MG Microgrid
CFD Computational fluid dynamics MILP Mixed integer linear programming
DC Direct Current MPC Model Predictive control
EENS Expected Energy Not Supplied O&G Oil and gas
EMS Energy management system O&M Operation and maintenance
GHG Greenhouse gases PV Photovoltaic
GHI Global horizontal irradiance R&D Research and development
GT Gas turbine TSO Transmission system operator
HIL: Hardware in the loop VIGV Variable inlet guide vane
ICE Internal combustion engine WVM Wavelet variability model
management, and must be able to handle phenomena that take place at - Size: The installed power capacity or total load served generally
various time-scales. The long-term planning (or sizing) of micro-grids exceeds several hundreds of kW (and kWh) and often reaches several
involves several development steps and aims at finding the optimal ar MW (and MWh).
chitecture to optimize the plant in the long run (typically 20 years) - Types of consumer: Consumers are mainly operators of industrial
[12–18]. Energy management aims at balancing energy flows between equipment such as electric drivers for torque production, pumping
producers and consumers. Different strategies may be used to minimize and compression. The proportion of resistive load may be slightly
the costs under constraints of reliability or environmental footprint. This lower than on the residential microgrid. The daily profile follows the
is typically carried out within intervals of 10 min to several days production schedule and therefore differs significantly from a resi
[19–21]. Controlling power ensures good quality of supply and safe dential and tertiary load profile.
operation, and prevents large fluctuations and instabilities by balancing - Ownership and operation: The generating units and transmission
power flows in real time. This is generally assessed with a timescale from system within the industrial microgrid can be financed, owned and
a few milliseconds to seconds [22–24]. operated by different entities, but assets are rarely separated or
Due to the relatively small size of microgrids compared to utility managed individually. Consuming equipment is considered as a
grids, their sensitivity to power fluctuations is significantly increased, single client.
with the result that developers now have to account for power dynamics,
flexibility and production uncertainties [25,26]. Since microgrids are 1.1.2. Reliability as a watchword for planning and operation
local energy systems, technological choices may differ from one user It is needless to say that ensuring the continuity of electricity supply
case to another, such as commercial, residential, military or industrial is a matter of profitability for industrial facilities, but it can also be a
plants. Unlike for commercial and residential applications, the literature matter of safety. Loss of electricity supply can cause leakage in pipes,
only scantly covers industrial microgrids [27,28]. However, specific tanks and even process bolting in reactors. This is confirmed in
constraints on industrial power systems modify the above-mentioned numerous accident reports (such as [35]), proving that the power system
optimization paradigm and call for a deeper understanding of the dy must be resilient and carefully designed.
namic behavior of their components, such as renewable systems and The reliable design of power systems has been extensively covered in
fossil generation. This brings considerable research challenges to the Ref. [36] with definitions of concepts as well as golden rules for plan
microgrid field involving numerous problematics, varying from classic ning, operation and maintenance. The overall reliability of a power
control system theory to fossil generator thermodynamics and renew plant refers to its ability to continuously supply electricity. Power plant
able resource signal treatment. unavailability causing electricity outage is mainly due to (1) the un
availability of power producers, (2) their disconnection due to electrical
instability or (3) the loss of the whole grid supply after a large instability.
1.1. The concept of industrial off-grid microgrids
To ensure a power plant’s highest availability, the mean time be
tween failures of power suppliers and the mean time to recover normal
1.1.1. General definition
operation must be evaluated [37]. One part of the problem consists in
It is difficult to establish a proper definition, since needs, size and
optimizing the maintenance planning of each power device. However,
resources may substantially differ from one case to another. Based on a
unplanned events, measured by the failure rate of equipment (proba
review of existing projects (see Table 1), studies and industrial surveys,
bility of unexpected loss), need to be minimized. In conventional
it is possible to propose some general characteristics:
Table 1
Examples of studies and projects focusing on industrial microgrids.
Reference [29] [30] [31] [32] [33] [34]
Application Mining Offshore oil Industry Industry Water treatment Manufacturing

Energy Electricity Electricity Heat/Electricity Electricity Electricity Electricity
Power 26 MW 40 MW 14 MW MW-scale 3.2 MW 20–40 MW
On grid/off grid Off-grid Off-grid On-grid Not specified On grid On grid
Renewable resource Solar Wind Wind – Solar Wind Solar Wind Solar
Fossil generation Diesel CCGT Not specified Diesel Diesel –
Storage – – Yes
Comment Installed plant Study Study Study Study Study
2
fossil-based systems, the major risk is fossil generator breakdown, but inertia reduce the system’s resiliency to power fluctuations. However,
the integration of PV systems introduces new parameters into the reli the literature only scantly covers this aspect of industrial microgrids,
ability assessment: despite the fact that it remains critical for their long-term development
[47, p. 58], [48]. When planning and operating their systems, de
- Risk of equipment failure in the PV power plant (PV strings, velopers must now evaluate the maximum allowable PV penetration
inverters) that will still satisfy the reliability target.
- Loss of transmission system (lines and converters) between the main
bus and the PV plant. 1.1.3. Focus on oil and gas applications
Oil and gas applications (O&G) offer a wide variety of examples for
Additionally, PV power fluctuations due to cloud passage can cause industrial microgrids, since O&G has used on-site isolated electrical
considerable power quality fluctuation that has the same impact as loss generation for decades. Many operating facilities, studies and future
of equipment. This is a new aspect in the reliability assessment as it is not projects can provide useful feedback on experience, operational data,
related to the reliability of the equipment, but to the site’s weather and even methodological insights. Some key characteristics and chal
conditions. This will therefore be addressed in detail in this paper. lenges are shared with numerous other applications, such as hospitals,
When such events occur, load shedding can be triggered to maintain military facilities, high-tech manufacturing (manufacture of electronic
frequency and voltage (see part 3.1.1), which impacts the profitability of chips), and highly sensitive urban areas.
the system since some equipment is not fed. However, if the system Due to the level of reliability required and the isolated location of
becomes unstable, generators and load may disconnect in cascade to O&G facilities, fossil generation is generally the only solution to produce
protect themselves from overspeed, leading to a complete blackout of electricity, and this participates in “well-to-wheels” emissions [49]. The
the system [38] and therefore potentially an industrial accident. Fig. 1 increasing pressure of global warming has led the O&G industry to study
illustrates the reliability aspects of industrial microgrids. renewable integration in its power generation units [50].
If the facility is connected to a stable grid, switching from off-grid to Table 2 gives an overview of studies addressing the problem of
on-grid configuration may be sufficient. However, both in weak-grid hybrid fossil-renewable power for O&G. Wind integration and offshore
(subject to outages) and off-grid configurations, the continuity of sup platforms are the most widely covered subjects, while fewer studies have
ply relies only on the self-producing capacity of the plant [39,40]. This addressed the problem of onshore facilities and solar PV systems from a
motivates integrating redundancy constraints in fossil generation plan technical point of view. The studies highlight that electrical instabilities
ning [41] and installing storage capacity to boost the flexibility of the due to renewable resources constitute the main technical limiting factor
plant. The reliability challenges of integrating more and more renewable and that storage is generally necessary to allow large-scale integration.
resources into large-scale systems have been covered by island micro Design optimization, such as carried out by Ref. [30], shows the interest
grid studies [42] such as those performed for the islands of La Réunion of renewable integration for fuel and CO2 reductions. However, no
[43] and El Hierro [44]. In Refs. [45,46] it is shown that a stochastic parametric study including storage has been proposed to date and it is
variation of renewable resources and a reduction of the mechanical expected that the additional capital expenditure would reduce the
Fig. 1. Summary of reliability aspects of industrial microgrids from risks to consequences. λ refers to the failure rate (see Ref. [36]) and ΔP refers to the probability of
power variation in the equipment.
3
Table 2
Review of recent studies addressing O&G microgrids.
Reference [52,53] [30] [54] [51,55] [56–58]
Scope Upstream & downstream Offshore oil Offshore oil Offshore oil Offshore oil
Energy Electricity, Heat Electricity Electricity Electricity Electricity
Power range 40 MW 40 MW MW
On grid/off grid Off-grid Off-grid Both Off-grid
Renewable Solar Wind Wind Wind Wind

resource
Fossil generation – CCGT CCGT Gas turbine Gas turbine
Storage – – Battery Battery

Focus area Large-scale potential for Method for optimizing Power quality challenges. Example with Operational & energy Electrical stability
solar design Hywind project optimization study
financial benefits. Furthermore [51], concludes that renewables will solar variation on continuity of supply is much more challenging to
impact the fatigue and lifetime of fossil generation, which could be a evaluate.
major issue when reliability is a key factor. - At short time-scales, the variability of PV power remains subject to
Based on this brief literature review, the following specificities of significant uncertainties, and is the main reason for power imbal
O&G facilities can be identified: ances in the grid. Methods for quantifying and forecasting short-term
ramps are still under development, as well as models that account for
• Isolated location: O&G rigs are often located in isolated areas or in the effect of a plant’s geographic dispersion. As the impact of solar
developing countries with an unreliable grid. Facilities can be variability increases with the penetration rate, accurate models are
located either onshore or offshore. an absolute necessity to reliably plan industrial microgrids.
• High power demand: The power delivered is rarely below 10 MW - Flexibility levers used to compensate renewable power variations are
and can exceed hundreds of MW. crucial to ensure grid reliability. Storage systems are the subject of
• Low fuel price: As the fuel is available on-site and sometimes not increasing investigation, while less attention is being paid to fossil-
valued for business, its financial value is very low (until recently, fuel technologies due to their apparent maturity. But their
CO2 emissions usually incurred no cost). response to successive ramps remains unknown and may signifi
• Proportion of torque, electricity and thermal demand: except for cantly alter their lifetime and fuel consumption. Very few models
refineries, which consume more heat than electricity, O&G facilities account for their dynamical behavior or integrate it in a long-term
almost exclusively use torque and electricity. assessment.
• Reliability and quality specifications: Strict limitations are estab - A large number of studies now address the control and management
lished for frequency and voltage fluctuations as well as equipment of microgrids to smartly interconnect all power devices on the grid.
availability. This involves making the right trade-off between economic perfor
• Low load flexibility: For safety and economic reasons, demand must mance and robustness. Reviews of existing technologies should take
be met at all times and no shift in load can occur during the day due this point of view in order to identify the appropriate strategies to
to unavailability of renewable resources. implement.
• Fossil generation to ensure continuous power: The architecture fea
tures gas turbines or internal combustion engines with redundancy This paper intends to contribute to filling these gaps as follows. First,
constraints to prevent generator loss from causing a system failure. we present research questions related to the integration of renewables
into reliability-constrained industrial microgrids. For the sake of suc
This overview shows the interest of O&G microgrids to identify the cinctness, only solar-PV applications are covered in this paper. We then
challenges and to highlight success stories of other industrial applica provide a review of the short-term dynamical assessment of PV systems
tions in isolated areas. In a research perspective, O&G microgrids pro and fossil generation. Next follows an investigation of real-time control
vide interesting case studies and make it possible to compare results and strategies that ensure reliability and power quality within industrial
validate them in a similar environment before considering their adap microgrids. Fourthly, we propose a discussion of the main findings and
tation to other applications. compare them with previous literature review studies. The final section
concludes and suggests future avenues for research.
1.2. Research challenges and paper summary
2. Review on power supply technologies
From the review presented above, the following research challenges
are identified: 2.1. The challenge of solar-power variability
- Due to their relatively small size compared to large grids, industrial Integrating PV systems in industrial power plants brings additional
power systems are more vulnerable to electrical instabilities [54]. risks for the continuity of supply and may therefore reduce the reliability
However, industrial facilities cannot afford unplanned blackouts and of the power plant. Reference [59] provides an overview of reliability
electricity shortages. This calls for the development of a method to assessment methods for PV inverters, modules, transmission systems,
ensure an equilibrium of active and reactive power with a high de and overall distribution systems based on fault analysis. Insights on the
gree of reliability. impact of aging and weather conditions on reliability over the lifetime of
- Reliability constraints make it necessary to consider unplanned a plant are provided in Ref. [60], showing that high average ambient
power fluctuations brought about by the integration of PV systems. temperature is likely to increase failure rates (almost 10% after 5 years
Equipment breakdowns can easily be integrated in the reliability in the case of a PV system in Arizona against 0% in Denmark). This factor
assessment process by following the same procedure as fossil gen must therefore be considered in reliability analyses. In Ref. [59], solar
eration breakdown. However, the impact of short-term stochastic variability is mentioned as one of the main future challenges for distri
bution systems; however no insights are given on its impact on
4
reliability. This is why a special focus on solar variability assessment is 2.1.1.1. A variability indicator to accurately account for perturbations. It
provided in this study. is possible to assess the level of variability over an entire day in order to
Solar photovoltaic (PV) power is traditionally assessed with a quasi- determine the frequency and magnitude of perturbations. The vari
static framework, as only hourly variations are of interest for economic ability index as defined in Ref. [64] is the ratio between the length of
evaluation [15]. As previously mentioned, the vulnerability of global irradiance series and the length of clear sky irradiance over a
large-scale isolated power systems motivates a study of sub-hourly defined time interval (the study proposes to calculate the variability
phenomena that might impact electrical stability. Hence, short-term index over 1 day). In Ref. [65], 6 metrics are compared to evaluate the
variability is a key element to ensure the balance between production solar variability on 31 test days. The study proposes a new indicator
and consumption and therefore a high level of reliability. Eq. (1) shows based on integrating the cumulative density function of solar increment.
the relationship between the power produced by a PV panel and the According to the study, these metrics can be used to characterize the
solar irradiance [61]. variability at both high and low frequencies and classify days depending
on the perturbations.
GHIplane of
PPV = Pinstalled * *[1 − Kθ (θ − 25)] (1)
array
1000
2.1.1.2. Climate-dependent variability. Location plays a role in solar
where Kθ is the temperature sensitivity coefficient, Pinstalled the rated variability due to climate type, orography and vegetation (large forests
capacity of the panel in ISO conditions, and GHIplane of array the Global also generate clouds during daylight). A solar variability map of the US
Horizontal Irradiance calculated in the plane of the panel. The actual drawn up using high-resolution production data [66] shows significant
amount of energy collected by the panel will depend on additional differences between desertic-arid areas and islands like the Lau islands
factors [62] such as the clearness of the sky, the performance of the PV and Hawaii where the highest level of variability is observed. In
panel (efficiency, temperature sensitivity, aging etc.), shade and fouling Ref. [67], the author studied the relationship between the clearness
due to the surrounding environment, etc. However, only GHI variation index and solar variability for locations in different climate zones. The
significantly impacts the system at short time-scales. authors concluded that climate zone and weather-driven clouds may
Cloud passage is the main contributor to short-term variability and have less impact on variability that the orography of a site. It is expected
needs to be addressed carefully. Fig. 2 shows an example of solar irra that the small, fast-moving clouds formed by the relief may have a bigger
diance variability over two days with a 1-s time-step. impact on short-term variability.
The following sub-sections will focus on the approaches imple
mented to explain, measure and forecast short-term solar variability in 2.1.1.3. Impact on PV systems. A PV system’s short-term variability is
order to integrate the risk of power imbalance in an industrial microgrid. different from the irradiance variability observed from a single sensor.
In Ref. [68] it is concluded that the power profile entirely follows the
2.1.1. Different approaches to address PV variability irradiance profile for time ranges greater than 10 min. However,
Solar variability affects the irradiance measured by sensors as well as short-term variability is affected by the size, shape and distribution of a
the power produced by one or several PV panels. Two time-scales are plant. For plants of several megawatts, 1-s, 10-s, and 1-min ramps can be
important for microgrid operation and planning: approximately 60%, 40%, and 10% smaller, respectively, than those
measured by a pyranometer. These results are confirmed in Refs. [69,
- Solar power range: difference between the minimum and maximum 70].
solar power output within a time interval (typically 15 min to 1 h).
- Solar ramps: the change in solar power or irradiance in a short time 2.1.2. Modeling solar system variability
interval (within the range of the sampling interval: 1 s to 1 min) that
dispatchable units will have to instantly compensate. 2.1.2.1. Smoothing irradiance data to obtain PV power. Since PV power
transients are smoother than irradiance transients, considerable work
Solar range and solar ramps are necessary to calculate the amount of has been done to obtain a realistic power output from the data
spinning reserve, the ramping capacity of dispatchable units, and the employing a single irradiance sensor. A dispersion factor is used in
storage capacity requirements. In high-reliability applications like in Ref. [71] to characterize different types of layout (crowded or spacious),
dustrial systems, extreme ramp events are also of interest to make a and the irradiance variability is smoothed to obtain the power produc
robust assessment. tion in several configurations. In Ref. [72], this concept is used to
calculate the extreme ramp rate of a power plant. In Ref. [70], the
Fig. 2. Two irradiance profiles captured with 1-s sampling [63].
5
authors proposed a transfer function with empirical coefficients to cost-effectiveness, reliability and emission regulations compliance need
address variability smoothing and compare the results with six to be guaranteed.
multi-megawatt power plants. As an example, Northern Ireland has identified the need for flexi
A Wavelet Variability Model (WVM) is proposed in Ref. [73] to bility as a major issue for a large-scale renewable integration [92]. The
obtain a PV power time-series from irradiance data. The WVM is potential wind curtailment was assessed using hourly wind ramp anal
compared to three other methods: linear scaling, moving average, and ysis. The study shows the great potential of thermal generation flexi
averaging on every sensor available. When compared to the 2 MW bility for guaranteeing energy balances.
output power, the WVM outperformed the three methods, especially The main objectives for thermal generation in industrial applications
when evaluating the maximum ramp rates. can be summed up as follows:
2.1.2.2. Modeling solar variability by generating stochastic time series. 1. Adapt the power output to match the power needs.
Assessing solar variability is a big challenge if no high-resolution irra 2. Provide enough primary response to ensure frequency and voltage
diance data are available on-site. One solution is to generate an irradi stability.
ance profile based on numerical weather prediction or satellite 3. Maintain the highest level of availability when operated in extended
predictions. In Ref. [75], methodologies based on Markov chains are transient mode.
proposed to generate irradiance and a clear-sky index profile [76]. re 4. Ensure cost-efficiency in the long run with regards to O&M costs and
produces high-frequency patterns of historical data to increase the aging.
temporal resolution of satellite prediction. However, these methodolo 5. Guarantee compliance with emission regulations in transient
gies require a representative training dataset of high-resolution ground operations.
measurements, which requires to have closely located sensors available.
Recently [77], paved the way for generating 1 min data without ground As performed in Ref. [93], numerous commercial studies attempt to
measurements, but concluded that additional research needs to be car compare fossil generation flexibility for hybrid generation purposes. To
ried out in order to obtain finer resolution. ensure that such comparisons are creditable and accurate, performance
assessment needs shift from steady-state, off-design correlations to dy
2.1.3. Forecasting of solar irradiance and variability namic modeling at a short time-scale.
Solar forecasting has been a growing topic in recent years since it For project developers, making a choice based on a generator’s
provides valuable information for microgrid operations and is used for dynamical performance is not an easy task:
various purposes, such as market trading, reserve scheduling, genset
planning, and storage management [78]. Its potential to optimize bat • Ramp-up capacities are rarely provided by manufacturers for small
tery sizing and lower the LCOE of ramp-constrained multimegawatt time-scales (seconds).
power plants is highlighted in Ref. [79]. Capturing solar variability with • Manufacturers usually provide ramp-up capacities for fast start-ups
forecasts may consist in providing either very short-term production and shut-downs, which are meant to be occasional.
estimations or an indicator of the variability to be expected for a large • There is no guarantee that the thermodynamic parameters (pressure,
horizon (15 mn to 1 h) [80,81]. As an example [82], proposes a meth temperature, etc.) will remain within acceptable limits.
odology to estimate the largest ramp rate by analyzing cloud shadow • No insights are provided on the machine’s performance: fuel con
velocity and irradiance sensor measurements. sumption, emission, fatigue.
Satellite irradiance forecasts and numerical weather predictors are
now widely used and can be accessed online [83]. Due to their large time These reasons call for a deeper understanding of fossil generator
and space resolutions, a downscaling approach is necessary to predict dynamics and more detailed modeling. The next part presents some
the variability at 10-s or 1-min scales. Such solutions are proposed in theoretical aspects as well as the state-of-the-art on dynamic modeling of
Refs. [84–86], but due to the geographical dependency of the variability, fossil assets.
no generic method has been proposed to downscale satellite prediction
without high-resolution data at the specific location. 2.2.1. Gas turbine
Sky imagers have been developed and commercialized in recent
years with the intention of providing very short-term predictions for 2.2.1.1. Working principles. Gas turbine (GT) theory has been widely
spinning reserve management (diesel load margin, storage capacity, developed in past decades and detailed in a large number of references
etc.) [87,88]. The development of these techniques will be a key factor [94–96]. GT performance predictions have been one of the main con
of success for the reliable management of industrial microgrids. cerns in order to evaluate differences between ISO parameters provided
Finally, in-situ measurements can enable forecast processing with a by manufacturers and user’s operating conditions. Fig. 3 and Fig. 4 show
very high resolution [89,90]. One criticism of this technique is that the the main components of single-shaft and twin-shaft gas turbines.
sensor network has to map a large area or be carefully placed depending As shown in Fig. 5, the compressor crosses iso-efficiency lines during
on the general cloud cover. If not, clouds may be inaccurately described transient operation and therefore increases the turbine’s fuel con
and, consequently, solar variability and ramp events will be sumption and particle emissions and impacts its mechanical fatigue [95,
misevaluated.
2.2. Fossil generation
With the penetration of variable and uncertain renewable power,

fossil generation (mainly gas turbines and internal combustion engines)
is about to dramatically change from a steady-state operational para
digm with a fixed load and a well-known production plan to a constantly
changing demand subject to stochastic perturbation. This obliges man
ufacturers and users to ensure operational flexibility [91] with an
increasing number of start-ups, shut-downs and fast load changes, and
increases the need for power-quality control. Meanwhile, Fig. 3. Single-shaft gas turbine [94].
6
Fig. 4. Twin-shaft turbine [94].
Fig. 6. The GAST model as presented in Ref. [101]. The capacity of the
generator to provide power as a function demand is modeled using time con
stants Ti and saturations V.
perform a gas turbine diagnostic under transient operation. Reference

[103] proposes a dynamic modeling of a single-shaft turbine featuring
VIGV (Variable Inlet Guide Vanes). The simulation time is typically 10
min.
In references [102,103], a model is calibrated with steady-state data
using an advanced iterative technique to minimize the error between the
model’s results and a benchmark provided by verified data (e.g.
PROOSIS or ThermoFlow software). These methodologies may provide
tools to represent the behavior of gas turbines. The parameters vary from
one turbine to another. For a 0-D representation, a compressor map and
a turbine map are necessary. This information has to be made available
by manufacturers, which creates a strong barrier to study and compare
their performances.
Fig. 5. Compressor transient running line. Two references point to an interesting compromise between data
requirements and accuracy, with a “component-based” approach using
p. 268]. Acceleration and deceleration are also constrained by surge and the Modelica environment [106,107]. The models should be easy to
flameout limits that need to be taken into account to assess the ramping adapt from one gas turbine to another, and results show good accuracy
capacity of the turbine. compared to experimental data.
To date, gas turbines have rarely been used for renewable microgrids
despite their significant potential [108]. Based on their model [102],
2.2.1.2. Representation of dynamical behavior. The thermodynamic re
show that hybrid gas/wind power plants can reduce both fuel con
lationships of the components can be used to simulate the steady-state
sumption and NOx emissions by at least 40%. However, NOx reduction is
“off-design” points due to part-load operations, special weather condi
debatable, as pointed out in Ref. [109], where an increase in NOx
tions, etc. This methodology can be found in the literature and is used by
emissions from large-scale, gas-fired power plants has been observed for
commercial software like ThermoFlow [97] and Proosis [98] but only
shares of renewables below 30%. Due to increasing pressure from
represents static conditions.
emissions regulations, the question of emissions modeling for transient
The development of dynamic models for the performance prediction
operations and the integration of selective catalytic reduction will
of gas turbines is still the subject of considerable research. Various
undoubtably be a key challenge for researchers in the future.
methodologies with a growing degree of complexity can be used [99].
Computational-Fluid-Dynamics (CFD) models are probably the most
2.2.2. Internal combustion engines
accurate, but the level of complexity may go beyond the needs of
renewable integration studies. Moreover, these models require large
computing capacities and perfect knowledge of the components’ 2.2.2.1. Working principle. Research on diesel engine modeling has
geometry. mainly focused on propulsion applications [110] and only a few refer
Some studies propose “black-box” models based on artificial intelli ences study its behavior in power systems [111,112]. The principle of a
gence in order to learn and reproduce the behavior of an asset based on diesel generator is basically the same as an automotive diesel engine:
operational data [100]. Zero-D (0-D) modeling appears to be a good combustion within the pistons drives the crank shaft, which provides
trade-off between a comprehensive approach and complex calculations. mechanical torque to the electric generator (see Fig. 8). This process
Several levels of complexity exist here as detailed in Ref. [101], which follows the Carnot thermodynamic cycle.
compares different modeling techniques for power quality studies. The
paper shows that significant differences are obtained if physical pa 2.2.2.2. Representation of dynamical behavior. Unlike for gas turbines,
rameters are ignored, such as ambient temperature. manufacturers tend to provide more insights on the short-term dynam
A similar 0-D representation is introduced in Refs. [102–104] (which ical behavior of their engines since this is a key selling argument (see
also introduces the inter-component volumes method). Blocks are ob Fig. 9). Nevertheless, the impact of a repeated load increase remains
tained by implementing a Laplace transformation of thermodynamic unknown and a detailed thermodynamic modeling is necessary.
equations (see Figs. 6 and 7).
Reference [102] proposes a MATLAB/Simulink model of a dual-shaft 2.2.2.3. Thermodynamic approach. The thermodynamics of the tran
turbine. This reference provides a control scheme for Gas Turbines (GT) sient operation of ICE are widely covered in Refs. [114,115]. Two main
for the purpose of a wind-GT hybrid power plant and simulates 10 h of modeling approaches can be used considering either a continuous phe
operation. The same modeling technique is applied in Ref. [105] to nomenon (mean - value model) or discrete events for combustions
7
Fig. 7. Dynamic model of the engine and its regulation system [102].
Fig. 8. Principle of ICE power conversion from combustion to electricity supply.
based on thermodynamic expressions, enabling a fairly accurate study of

the engine’s step load response. The parameters needed for the simu
lation are generally not available in the literature and the author rec
ommends calibrating the model using experimental data.
Finally, reference [117] provides a simple methodology to model the
transient behavior of a diesel engine. The model is calibrated with
experimental data and requires accurate turbocharger parameters to
run.
2.2.2.4. Reduction to a time constant. From a control-science perspec

tive, the transient modeling of a diesel engine mainly consists in
describing 4 main components: the regulator (controller), the actuator,
the engine’s combustion, and the synchronous generator (see Fig. 11). In
Refs. [118,119] dynamic models are proposed for the frequency control
(taking actions on the angular velocity) and therefore represent the
short-term transient response of the engine.
The above-mentioned references are based on a mathematical
Fig. 9. Recommended load increase of Wartsila W32 according to the product modeling of the machine’s processes and components. Other modeling
guide [113]. techniques consist in “learning” the engine’s response using neuro-
computing. In Ref. [120], an HIL test bench is combined with a real
within the pistons (discrete-event models) diesel engine in order to learn the frequency’s transient behavior and
Both methods aim at describing the change in the pistons’ input control. A dynamic model is then built to conduct stability analysis.
parameters (air mass, fuel mass and exhaust gas recirculation) when a However, this approach gives a very poor understanding of the ma
torque command is given to the system as described in Fig. 10. chines, which makes it very difficult to adapt to other engines with, for
In [112, p. 748], several modeling approaches are described, many of example, a different number of pistons.
which are based on highly complex fluid mechanics. However, a lack of Studies [121,122] implement a promising methodology using
data on the engine’s geometry makes it impossible to implement such a GT-SUITE software with a good understanding of components and
model. Moreover, this study intends to be adaptable to multiple types of available component libraries for dynamic modeling.
engine. The different methods implemented for assessing the engine’s flexi
Reference [116] presents a simple mathematical model of an engine bility potential have many advantages. Firstly, they extend compre
hension of the engine and may allow more accurate stability studies. For
example in Ref. [123], a sensitivity analysis of the engine injection delay
is carried out. The results show that a short time delay significantly
improves the grid’s frequency stabilization. The study recommends a
low injection time delay (mostly found in high rotational speed engines)
in order to improve the penetration rate of renewables.
A good comprehension of the thermodynamical limitation is also a
very important factor in renewable integration. In Refs. [124–126], it is
Fig. 10. Principle of torque regulation.
8
Fig. 11. Conventional diesel engine representation [118].
shown that a high penetration rate can be enhanced by low-loading

engines featuring dedicated pre-heating and clutching systems.
2.2.3. Aging and maintenance considerations related to PV variability

smoothing
Is has been observed that fossil generators will increasingly be
operated in a transient state with rapid variation to smooth out solar
variability. The evolution of fatigue and fault probability due to tran
sient operation is still poorly addressed by the scientific community
[127]. evaluates the impact of an extended dynamic operation of CCGT,
and indicates that the lifetime reduction may be up to 10% greater than
in reference steady-state cases. Further analysis would be necessary to
evaluate the failure rates of typical fossil technologies in industrial
microgrids. Solar variability modeling may once again prove crucial for
such an evaluation (see Fig. 12). Fig. 13. Buffer storage system.
2.3. Storage systems provide services with reaction times varying from a few seconds to
several hours. Li-Ion batteries, flywheels and supercapacitors seem to be
Storage systems may achieve two different objectives in microgrids: the most suitable options for these applications (see Table 3) [129,130],
energy shifting (see Fig. 12) and system services. System services are depending on the final use of the energy (for instance, heat storage may
applications dedicated to enhancing a system’s reliability when subject be the best option if the final service requires heat).
to unplanned events. For example, a storage system can provide power if
the renewable output suddenly decreases or if a fossil generator trips
3. Operation of industrial microgrids
(Fig. 13). The storage device is therefore used as a buffer to compensate
for the start time and ramp of the replacing unit.
Most studies of microgrids tackle the problem of operation and
Depending on its objectives, a storage system will have technical
planning by handling a single aspect of the microgrid (control, energy
specifications that require carefully choosing the technology [128].
management or sizing). However, due to the specific characteristics of
gives an outstanding overview of storage technologies in microgrids and
industrial microgrids, each should be subject to careful attention. As a
provides examples for the five main types of storage technology listed in
matter of fact, the size of a facility impacts power quality problems and
Fig. 14.
endangers the continuity of operations. Secondly, operational costs are
Thanks to the variety of storage technologies available, numerous
part of product profitability and must be optimized. Finally, as industrial
applications are covered, from very small uninterruptible power systems
power plants require huge investments, reliable sizing is crucial. Con
to utility scale systems such as pumped hydro. Fig. 15 shows the general
trol, energy management, and sizing are three steps in a single process
trends of technologies according to their rating capacities and discharge
that aims at finding the best operational scenario for the microgrid.
times (additional features for technology comparison can be found in
Fig. 16 shows the different time-scales for the main actions and phe
Ref. [129]).
nomena that take place during the microgrid’s lifetime.
Considering the characteristics of industrial microgrids and their
production units, storage systems must contain several MWs in order to
3.1. Short-term power control
The performance of a microgrid is closely related to its control

scheme since it is supposed to execute the production plan, while pro
tecting the power system and devices when instabilities occur. Micro
grids ensure a reliable electricity supply by following power quality
specifications and protect the devices from going out of their operational
range. In industrial microgrids, special attention is paid to the frequency
and voltage control.
3.1.1. Fundamentals of frequency and voltage regulation

In a synchronous electricity network, the value of the frequency
needs to be the same at every point to avoid generator de-
synchronization and allow proper use of the equipment. This means
that all of the synchronous generators feeding the system must share the
Fig. 12. Peak-shaving storage system. same rotor angle velocity ω as this is directly linked to the frequency (Eq.
9
Fig. 14. Categories of storage technology [129].
reactive capacities through the successive sizing steps. Finally, unlike for
active power, reactive power can be produced by all devices, which
gives more flexibility to the system. This is why assessing the impact of
variable energy resource penetration on voltage stability can be
considered less important than assessing frequency stability problems
[141,142].
3.1.1.1. The role of inertia in passive regulation. As seen in Fig. 17, the
kinetic energy delivered in the first few seconds is much greater than the
primary reserve and frequency-dependent load reduction. When a sud
den power imbalance occurs, mechanical torque takes place as both
sides of the alternator shafts cause a deceleration. The inertia softens the
shaft’s deceleration by returning the kinetic energy stored in the rotating
mass [140]. The inertia constant H (sec) accounts for the kinetic energy
stored in the synchronous machine. It is calculated as a function of the
Fig. 15. Discharge time and power ratings of storage technologies [128,131]. angle velocity ω (rad/sec), moment of inertia J (kg/m2) and power
rating of the unit (MVA) (see Eq (5)).
(2)). When a global imbalance of active power occurs in the system, the
swing equation (Eq. (3)) expresses the fluctuation of the frequency 1 J ω2 *10− 6
H= * (5)
[140]. 2 MVApu rating
ω = 2πf (2) The inertia constant of a system is the sum of the individual inertia
( )
∑
pu constant H of each generating unit i Htotal = Hi . Due to the
dΔf Ppu
gen − Pload i
= f0 (3) increasing penetration rate of renewables, industrial microgrids tend to
dt 2H
have less inertia in their systems. This is due to the fact that the inertia
Similarly, local voltage fluctuation can occur due to reactive power
values of diesel generators and gas turbines are lower than those of large
imbalances. The network’s topology plays a major role and stability
conventional fossil units (see Table 4), and because units may be shut
assessments require more information on the system. The general
down when renewable power feeds the system.
sensitivity of the voltage and rotor angle (ΔV, Δθ) on active and reactive
In a microgrid, maintaining enough inertia is crucial to ensure the
power (ΔP, ΔQ) is expressed in Eq (4) [140] (J is the Jacobian matrix
grid’s stability. This aspect must be considered with the same care as
expressing the sensitivity of the system):
active control actions. This is pointed out in Ref. [144], where a
[ ] [ ][ ]
ΔP J JPV Δθ microgrid’s frequency response is evaluated for different inertia con
= Pθ (4)
ΔQ JQθ JQV ΔV stant values.
Although PV systems do not provide inertia to the grid, power
Voltage stability can be handled at relatively low costs with available
electronics and a fast response storage system may help to synthetize
commercial technologies. In addition, reactive power compensation
inertia and therefore improve the system’s resiliency [23,145].
capacities (such as capacitor banks) are more affordable than active
power compensation capacities (which are new production units, such
3.1.1.2. Control actions on dispatchable units. Typical control strategies
as storage or fast fossil generation). This means that the techno-
for microgrids under various conditions are detailed in Refs. [7,146] and
economic balance will be less impacted by making adjustments to
an outstanding review is provided by Ref. [23]. Fig. 17 shows the typical
Table 3
Review of the most relevant technologies for reliability-constrained industrial microgrids.
Technology References Advantages Drawback Comment
Li-ion batteries [128,132, High energy density, high cycling efficiency, Lifecycle degradation due to Market leader. Drawn by synergies with automotive
133] rapid response time, low self-discharge cycling and thermal effects. applications
Applicable to other uses (energy shifting) Recent economies of scale.
Flywheels [134,135]. Quasi-infinite number of cycles Low energy density Low maturity, high costs
High self-discharging rate
Supercapacitors [136–138] High power density, long lifetime and limited High self-discharge rate (up to 40% Suitable only for very short-term applications. Low
aging. a day). maturity in large-scale applications
10
Fig. 16. Different time-scales for planning and operation of power systems inspired by [139].
different droop characteristics participate in the frequency regulation.

The active power control is a common rule applied by all units that
takes place in the primary response. However, controllers must embed
additional features in order to prevent the equipment from going out of
their operational range. This will be presented in the next sections.
3.1.2. Control of fossil engines

Fossil engines are naturally the main actors of power quality regu
lation. In most stability studies, the regulation potential is solely
addressed using droop Rgen . But fossil units remain limited by their
flexibility in terms of ramp and load factor (which justifies the study of
their dynamical behavior as presented in part 2.2).
The control of fossil units and its simulation for stability studies will
depend on the availability of information regarding the unit’s transient
behavior as well at its technical limitations (ramping range, power
Fig. 17. Evolution of frequency after a sudden load step [143].
range, maximum fuel mass-flow, temperature limits, emissions limits,
etc.). In order to model and study the electrical stability, only a high
Table 4 level of abstraction is generally possible due to lack of public informa
Value of inertia constant for several production units [143]. tion (see Fig. 19).
Type of production unit Inertia constant (s)
The control scheme may consist of a succession of saturation and
time delays accounting for constraints and transient behavior, such as in
600 MW steam turbine 13.7
the reduced order system frequency response model in Ref. [148] shown
11 MW average speed diesel engine 3
21 MW average speed diesel engine 5.2 in Fig. 20. This accounts for communication and injector time delays as
30 MW gas turbine 2.5 well as the physical limits of minimum and maximum power.
Photovoltaic generator 0 In Fig. 20, and similarly to the control of conventional large-scale
units, the power output is only driven by the droop characteristics and
the frequency shift. However, in normal operations, the fossil generator
three-layer hierarchical control used in most systems to adjust the power
needs to supply the power required by the user. The control signals of
and correct any deviations.
the load demand and frequency regulation are therefore summed to
Among all of the primary controls developed in the literature, the
obtain the final power need by the generator [140] (see Fig. 21).
droop-control method is by far the most commonly used and the easiest
to implement. In large power systems, it is used to control units involved
in primary regulation [140,143]. Its use in microgrids is justified and 3.1.2.1. Control of gas turbines. In ref [95], it is highlighted that a gas
detailed in Refs. [22,147]. turbine’s control system adjusts the power output while protecting the
Any power imbalance is automatically compensated by the genera engines from excessive exhaust gas temperatures (turbine overheating),
tion proportionally to the difference in active power ΔP = Δ(Pgen − excessive speed and stress of rotating parts, and stalling and surging due
Pload ). Each generator participates proportionally to its droop value Rgen to a high compressor ratio. Fig. 19 gives a direct example for gas tur
bines: the fuel command depends on the frequency deviation, and is then
and its contribution is therefore ΔPgen = Rgen *Δf (note that the genera
saturated by the surge limit and flameout limit. This is applied in
tion remains limited by its ramp rate). Fig. 18 shows how two units with
Ref. [102] to control the mass flow, which is a good methodology to
11
Fig. 18. Control of two units with different droop characteristics [140].
system’s frequency.
As battery energy systems are increasingly common in microgrid
applications, the literature covers a wide variety of control strategies
[149]. introduce the concept of state-of-charge weighted droop control
in order to adapt the regulation to the amount of energy remaining in the
battery. Similarly to fossil engines, battery control systems include
protections and saturation to take chemical dynamics and components’
electrical limitations into account [151]. In Ref. [152], an ingenious
control scheme is proposed to coordinate the frequency control of fossil
units and storage systems.
Since control schemes must take the specificities of each technology
into account, specific strategies are developed for flywheels [153] and
super-capacitors [154].
Fig. 19. Principle of a multi-constrained fossil generation frequency controller. 3.1.4. Control of load
In high-reliability industrial applications, the continuity of supply is
include the physical limits of components in gas turbine power control. a priority. This leaves little potential for taking action on load in order to
However, as pointed out in part 2.2, the knowledge required for correct instabilities. Underfrequency load-shedding procedures must
modeling such control is rarely available and sometimes a simplified only be used in emergency situations in order to protect the grid from
model has to be implemented (e.g., use of maximum and minimum ramp blackout [140]. It is however important to carefully design the load
rates). shedding to provide significant stability improvement [155]. If a sig
nificant share of load is considered as non-critical, it is possible to use it
3.1.2.2. Control of diesel engines. Diesel control strategies and chal as a flexibility potential and regulate the frequency and voltage de
lenges regarding the plant’s protection are stated in Ref. [24]. Similarly viations [51,156]. However, accurate load models are necessary to
to gas turbines, the power demand of a diesel unit must take into account assess the potential of flexibility and stability improvements [157]. This
its limitations, such as the minimum loading factor, or its maximum solution is one of the most advantageous since no additional investment
achievable ramp rate [118,119]. propose a diesel engine frequency is necessary to integrate a large share of renewables.
control and show how an engine’s characteristics may impact grid sta
bility. In Ref. [118], the advantages of a highly flexible control for diesel 3.1.5. Participation of PV systems
allowing low loading are presented. The results show an improved sta Unlike for wind generators, which have higher inertia and show
bility margin for frequency and highlight the potential of low-load diesel some capacity for frequency and voltage regulation, PV generators are
technology to reduce storage capacity investments. As fossil units tend
to be flexibility providers more than prime energy suppliers, low-load
and highly flexible fossil unit control will probably be a key issue in
the future.
3.1.3. Control of storage systems

Fast-discharging storage technologies are widely used for short-term
regulation of power imbalances. An overview of control strategies for
storage systems in given in Ref. [149]. In Ref. [150], an experimental
study shows how droop-controlled storage can be used to regulate the Fig. 21. Coupled frequency-load setpoint regulation.
Fig. 20. Reduced order frequency model accounting for gas turbine constraints and dynamics [148].
12
more limited when no storage is associated [158]. Depending on the PV systems and storage). Thanks to power quality simulation, proba
inverter technology, the power factor can be adjusted and therefore bilities of electricity outages (underfrequency load-shedding or grid
participate in the voltage regulation. Using curtailment, it is also blackouts) can be identified and thus integrated in the reliability
possible to regulate overfrequency events by lowering the power output assessment. This issue has been partly addressed in a previous work
of the inverter. This is detailed in Ref. [159], which presents an overview based on the duration of overfrequency and underfrequency regimes
of the active power control of PV systems. The potential of PV plant over one day [165]. Deeper investigations need to be carried out to
curtailment has also been pointed out in Ref. [160] to lower the risks of propose a methodology to evaluate the reliability indicators related to
short-term drops due to solar variability. By reducing the number of power quality problems. Fig. 22 proposes such a procedure for the
connected panels and therefore the available production, the potential reliability assessment of an industrial microgrid with renewable
solar drop is also reduced. This is however subject to high uncertainties integration.
related to the forecast of short-term variations. In addition, a significant
part of the solar production is lost, which reduces the environmental and
economic performance of the plant. This solution must therefore be put 3.2. The energy management approach
into perspective with other mitigation levers, such as storage systems,
and should be evaluated at the sizing step. Although power control is very important for grid stability and
When associated with a sufficient energy buffer with a very fast robustness, it does not deal with economic objectives or handle forecasts
response time, grid-forming inverters can enhance the grid’s stability of load and renewable resources. The economic performance of the
with synthetic inertia. This also allows black start capability when a power system involves ensuring the global equilibrium of production
large share of renewables feed the system. An overview of grid-forming and consumption while minimizing the overall operational costs and
PV inverters from an operator, manufacturer and research perspective is meeting important constraints such as emissions reduction. An Energy
provided in Ref. [161]. Management System (EMS) ensures the proper allocation of production
units by monitoring data and collecting forecasts. Considering industrial
3.1.6. Modeling of short-term control and power quality microgrid characteristics, centralized EMS generally seems more
Modeling the short-term power control is mandatory to accurately convenient [8].
assess a system’s potential for renewable integration. Using voltage and The efficient performance of EMS lies in its ability to allocate the
frequency response modeling, the authors of [46] proved that the resources of each device at a minimal cost, which means that its dispatch
maximum renewable penetration limit to ensure grid reliability in algorithm is of paramount importance. In Refs. [19,20], detailed surveys
Indonesia was 31%. In Ref. [162], a transient stability method is used to of energy management methods are proposed and a summary is reported
determine the maximum intermittent power penetration in an isolated in Table 5.
system. Using swing equation equal area criteria, the ability of a system Linear programming (LP) and multi-layer optimization techniques
to properly control active power during transient events is studied in turn out to be very relevant to address reliability and power quality in
Ref. [163]. Numerous commercial software applications now allow industrial energy management problems. This is why a special focus is
stability studies featuring renewable technologies (ETAP, MATLAB, proposed in this paper.
PSCAD, OPEN-DSS) and are widely used in both academic and industrial
environments. Finally, research programs increasingly work on simu 3.2.1. Linear programming, robust and stochastic optimization
lating systems in real time using hardware-in-the loop techniques [164]. EMS protects the grid from unexpected events (severe loss of
By running simulations over a long period of time, it is possible to renewable generation, a generator contingency or an increase in load)
assess the performance of the microgrid in terms of quality of supply by implementing the concept of spinning reserve, the N+1 rule, or even
through reliability indicators [32] such as EENS, CAIDI, ASAI, and ASUI. an aggregated indicator for a frequency shift. In Ref. [148], the chal
This study gives an example of how to assess the reliability level in lenges of a frequency-constrained model are developed. The possible
several plant configurations (islanded vs grid-connected, various sizes of options to directly integrate frequency shift mitigation lead to either a
non-linear problem or a sub-optimal solution (use of minimum ramping
Fig. 22. Example of a procedure for reliability assessment over a microgrid layout (where decisions are made on PV and storage capacities).
13
Table 5
Comparative review of energy management optimization techniques.
Optimization family Principle References Advantages Drawbacks
Rule-based & dynamic Logical decisions & flow diagrams [166,167] Comprehensive framework Lack of flexibility and reproducibility
programming Close description of expert knowledge Hard to implement for complex
systems
Artificial intelligence Machine learning based on large datasets [168,169] Large number of situations handled No comprehensive description
Lack of reliability certification
Meta-heuristics Optimization using bee-colony, genetic [170–172] Comprehensive optimization framework with No guarantee of optimality
algorithms, etc. non-linear modeling capabilities
Linear programming Optimization based on linear description of [148, Guarantee of optimality Requires relatively high level of
objective functions and constraints 173–179] Easy implementation abstraction
Fast-running commercial solvers Lower performance of non-linear and
non-convex description
Multi-layer optimization Predictive control model [55, Ability to enable different levels of Need to build a specific framework
180–185] abstraction with high fidelity model for each application
High modularity and flexibility to describe
the problem
capacity, minimum value of inertia, etc.).

Deterministic LP approaches such as described above are easy to
implement, yet insufficient to handle reliability and power quality, since
transients are caused by unplanned renewable losses and generator
contingencies. A stochastic approach can deal with probabilistic fore
casts and is therefore convenient to address renewables’ prediction
uncertainties. However, the complexity of the problem is significantly
increased, and this approach does not guarantee a fully reliable solution.
A stochastic approach is most relevant when a tradeoff can be made
between reliability penalties and operational savings. The authors in
Ref. [179] propose a two-stage stochastic MILP formulation to calculate
both unit commitment and reserve scheduling with 0-Hz and ± 10 MHz
of tolerance for frequency deviation. In Ref. [175], the frequency control
is integrated into the EMS, but this significantly increases the complexity
of the problem. A Benders decomposition is therefore used to solve the
problem.
A robust approach is useful when reliability constraints play an
important role in the operation, and avoids the use of probabilistic Fig. 23. Architecture of model predictive control for energy management.
forecasts thanks to an efficient selection of uncertainty scenarios [174].
Uncertainty is addressed from a conservative point of view involving a
study of the historical performance of the forecast system. This can be a is able to handle forecast uncertainties, simulate dynamic behavior with
challenge if no historical data are available at the plant’s location. accuracy, and improve both renewable penetration and operational
savings.
3.2.2. Model predictive control (MPC) for energy management systems
As pointed out in Ref. [181], a fully-integrated MILP model problem 4. Discussion and main findings
accounting for uncertainties and reliability can be very hard to solve due
to the number of variables and constraints. This justifies the use of In this paper, we detailed the concept of industrial microgrids along
model-predictive control (or a multi-layer approach) to address both with some key characteristics. This brings a new field of application for
control and management of the microgrid. microgrids to add to those previously detailed in Refs. [8,27,28].
As presented in Ref. [183], model predictive control for energy Additionally, we provided a review of renewable integration studies in
management consists of a predictive model that plans optimal energy O&G systems to illustrate the main challenges of industrial microgrids,
exchanges based on forecasts, and a system model that performs the which is a topic that has not been addressed in existing review papers.
dispatch and controls the power flux. The predictive model can be based The reliability of PV systems was found to be widely covered by
on various management techniques such as presented above (see previous literature reviews [59]. However, we identified gaps on the
Fig. 23). The level of accuracy of the system model can be adapted to evaluation of solar variability. Based on previous work gathering studies
available information and may embed a very detailed control scheme. on variability quantifiers [71], and solar forecasting [78], we presented
Finally, MPC reproduces the hierarchical relationship of the energy and state-of-the art techniques to evaluate PV power dynamics. Considering
power management system in real-life conditions. the size of industrial microgrids, we focused on large-scale PV systems,
The predictive model can be divided into two layers in order to stressing the need for geographical smoothing models.
perform a high-level economic evaluation and continuously track the The economic and environmental performances of industrial
optimal operation. In Ref. [185], MPC is used to model a DC low voltage microgrids are closely related to the performances of fossil generators.
microgrid. In Ref. [55], two-layer model predictive control is used to However, while the review studies provide a wide variety of models and
simulate an off-grid O&G platform fed by a wind farm, two gas turbines tools [101,110], they do not focus on the assessment of fossil unit per
and battery storage. formances in dynamic operations and cloud passage. This paper there
In [180,181,184], the high level optimization problem is addressed fore reviewed dynamical modeling techniques to improve the
using linear programming and therefore reproduces the decision of an performance assessment. The impact of dynamic operation on reliability
LP-based energy management system. The results show that the method and lifetime reduction remains an open question after this study as this
14
issue has been poorly addressed in the literature. Finally, the sizing of microgrids is also a topic of paramount
Most of the review studies handle a single aspect of the performance importance and has already been the focus of numerous studies. Sizing
of microgrids by covering either reliability aspects related to short-term that addresses reliability constraints is studied in Refs. [186–188] along
power control [23]or energy management optimization [19]. In this with the uncertainty of solar production [189–192], but a deeper study
study, both concepts have been addressed to draw attention to the link is necessary to identify the most suitable techniques for industrial micro
between these two operational layers. Similarly to Ref. [166], we put the grids. Such an analysis remains a perspective for this work.
focus on the control strategies of each component and also more deeply
cover power quality aspects to provide pathways for a reliability
assessment involving several microgrid configurations and renewable Declaration of competing interest
integration scenarios. Thanks to this analysis, the main goal of power
quality simulation in industrial microgrids may be defined as follows: to The authors declare the following financial interests/personal re
evaluate the risks of loss of electricity supply given a certain level of PV lationships which may be considered as potential competing interests:
penetration whilst properly considering the support role of fossil gen Louis Polleux benefits from a CIFRE PhD grant (No 2018/3264) funded
erators, storage, loads and inverters. by Total Energies S.E, Thierry Schuhler and John Sandoval Moreno are
Lastly, we identified the most relevant energy management tech employed by Total Energies S.E at the R&D department.
niques for industrial microgrids. Previous review studies have provided
a complete overview of energy management techniques and mathe References
matical formulations [19,20]. Since reliability is a key aspect of indus
trial microgrids, this paper provided an overview of reliability and [1] United Nations General Assembly. Political declaration of the high-level political
forum on sustainable development convened under the auspices of the General
power quality constraints formulations, and robust energy management, Assembly. presented at the Sustainable Development Goals Summit, New York.
and highlighted the potential of model predictive control for better Oct. 2019.
ensuring the link between cost optimization and power quality control. [2] UNFCCC Communications. COP25 to Be the launchpad for significantly more
climate ambition | UNFCCC. Nov. 29. 2019. https://unfccc.int/news/cop25-to-be
-the-launchpad-for-significantly-more-climate-ambition-0. [Accessed 9 March
5. Conclusion 2020].
[3] IEA. World energy outlook 2018. IEA Publications; 2018.
[4] IEA. Tracking industry. IEA, Paris. May 2019 [Online]. Available: https://www.
This paper provided a review of the main challenges of solar PV iea.org/reports/tracking-industry. [Accessed 23 April 2019].
integration into industrial microgrids under reliability constraints. The [5] Rissman J, et al. Technologies and policies to decarbonize global industry: review
need for continuous operation of industrial power plants requires high and assessment of mitigation drivers through 2070. Appl Energy May 2020;266:
114848. https://doi.org/10.1016/j.apenergy.2020.114848.
availability of electrical generation. On the other hand, the electrical
[6] Alby P, Dethier J-J, Straub S. Firms operating under electricity constraints in
system is more vulnerable due to its low inertia and low flexibility developing countries. World Bank Econ Rev 2013;27(1):109–32.
compared to large grids or residential systems with electrical loads that [7] Hatziargyriou N, editor. Microgrids: architectures and control. Chichester: Wiley-
IEEE Press; 2014.
are likely to be shed or delayed. The integration of highly variable and
[8] Guerassimoff G. Microgrids : pourquoi, pour qui ? Presses des Mines; 2017
stochastic solar power challenges the performance and reliability of [Online]. Available: https://hal-mines-paristech.archives-ouvertes.fr/hal-
power plants as the rapid power losses involved can cause considerable 01589804.
electrical instabilities. [9] Eigg electric,” the isle of eigg. http://isleofeigg.org/eigg-electric/. [Accessed 10
March 2020].
This calls for a more extensive understanding of the short-term [10] Berkeley Lab, “UCSD | Building Microgrid,” Microgrids at Berkeley Lab.
behavior of renewable systems to anticipate extreme events. Vari https://building-microgrid.lbl.gov/ucsd (accessed Mar. 10, 2020).
ability analysis techniques have been developed to quantify solar ramps [11] Bornard P. Conduite d’un système de production-transport. Tech. Ing. Réseaux
Électr. Transp. Répart., vol. base documentaire : TIB263DUO., no. ref. article :
and can be coupled with a geographic smoothing model in order to d4080. 2000 [Online]. Available: https://www.techniques-ingenieur.fr/base-
assess the power fluctuation of the overall solar plant. The availability of documentaire/energies-th4/reseaux-electriques-de-transport-et-de-repartition-
high resolution time-series remains a challenge, especially in isolated 42263210/conduite-d-un-systeme-de-production-transport-d4080/.
[12] Lian J, Zhang Y, Ma C, Yang Y, Chaima E. A review on recent sizing
areas. In the future, techniques to increase the granularity of satellite methodologies of hybrid renewable energy systems. Energy Convers Manag Nov.
observations and forecasts to capture local and short-term variability 2019;199:112027. https://doi.org/10.1016/j.enconman.2019.112027.
will be key factors of success for isolated microgrids. [13] Anoune K, Bouya M, Astito A, Abdellah AB. Sizing methods and optimization
techniques for PV-wind based hybrid renewable energy system: a review. Renew
A particular feature of large-scale microgrids is their use of fossil
Sustain Energy Rev Oct. 2018;93:652–73. https://doi.org/10.1016/j.
generators. The integration of renewable energy sources therefore re rser.2018.05.032.
quires a change in the operational approach. Consequently, conven [14] Sinha S, Chandel SS. Review of software tools for hybrid renewable energy
systems. Renew Sustain Energy Rev Apr. 2014;32:192–205. https://doi.org/
tional modeling and optimization tools that were designed for static
10.1016/j.rser.2014.01.035.
operations need to evolve, as fuel consumption, fatigue and emissions [15] HOMER Energy. HOMER - hybrid renewable and distributed generation system
could significantly diverge from historic benchmarks. The development design software. https://www.homerenergy.com/. [Accessed 19 March 2019].
of tools to model fossil generation dynamics is crucial to carry out a [16] Berkeley lab, “distributed energy resources - customer adoption model (DER-
CAM) | building microgrid. https://building-microgrid.lbl.gov/projects/der-cam.
proper techno-economic analysis of a plant and ensure availability. [Accessed 19 March 2019].
Control and management schemes must be chosen in order to [17] C. N. R. NRCAN. RETScreen, clean energy management software. Mar. 10. 2010.
correspond to this highly constrained framework. The large number of https://www.nrcan.gc.ca/maps-tools-publications/tools/data-analysis-softwa
re-modelling/retscreen/7465. [Accessed 7 April 2020].
non-linearities (solar ramps, electrical instabilities, fuel consumption, [18] Jayashree S, Malarvizhi K. Methodologies for optimal sizing of battery energy
etc.) as well as the role of the control scheme call for the integration of a storage in microgrids : a comprehensive review. In: 2020 international
detailed description of the electrical system in future techno-economic Conference on computer Communication and informatics (ICCCI); Jan. 2020.
p. 1–5. https://doi.org/10.1109/ICCCI48352.2020.9104131.
studies. Optimization techniques should therefore integrate these phe [19] Zia MF, Elbouchikhi E, Benbouzid M. Microgrids energy management systems: a
nomena into their equations while maintaining the ability to identify the critical review on methods, solutions, and prospects. Appl Energy Jul. 2018;222:
most economical solution. Model predictive control coupled with linear 1033–55. https://doi.org/10.1016/j.apenergy.2018.04.103.
[20] Ahmad Khan A, Naeem M, Iqbal M, Qaisar S, Anpalagan A. A compendium of
programming tends to be the best compromise considering the current
optimization objectives, constraints, tools and algorithms for energy management
state-of-the-art. in microgrids. Renew Sustain Energy Rev May 2016;58:1664–83. https://doi.org/
10.1016/j.rser.2015.12.259.
15
[21] Ossai CI. Optimal renewable energy generation – approaches for managing ageing Herzegovina; Oct. 2018. p. 1–6. https://doi.org/10.1109/
assets mechanisms. Renew Sustain Energy Rev May 2017;72:269–80. https://doi. ISGTEurope.2018.8571673.
org/10.1016/j.rser.2017.01.041. [47] IRENA. Innovation outlook: renewable: mini-grids. 2016. p. 184.
[22] Tayab UB, Roslan MAB, Hwai LJ, Kashif M. A review of droop control techniques [48] Liang X. Emerging power quality challenges due to integration of renewable
for microgrid. Renew Sustain Energy Rev Sep. 2017;76:717–27. https://doi.org/ energy sources. IEEE Trans Ind Appl Mar. 2017;53(2):855–66. https://doi.org/
10.1016/j.rser.2017.03.028. 10.1109/TIA.2016.2626253.
[23] Olivares DE, et al. Trends in microgrid control. IEEE Trans. Smart Grid Jul. 2014; [49] Masnadi MS, et al. Global carbon intensity of crude oil production. Science Aug.
5(4):1905–19. https://doi.org/10.1109/TSG.2013.2295514. 2018;361(6405):851–3. https://doi.org/10.1126/science.aar6859.
[24] Katiraei F, Mauch K, Vandenbergh M, Arribas L. PV hybrid mini-grids: applicable [50] OGCI. A report from the oil and gas climate initiative. Sep. 2018.
control methods for various situations. IEA; 2012. [51] Alves E, Sanchez S, Brandao D, Tedeschi E. Smart load management with energy
[25] Farrokhabadi M, et al. Microgrid stability definitions, analysis, and examples. storage for power quality enhancement in wind-powered oil and gas applications.
IEEE Trans Power Syst Jan. 2020;35(1):13–29. https://doi.org/10.1109/ 2019. p. 15.
TPWRS.2019.2925703. [52] Wang J, O’Donnell J, Brandt AR. Potential solar energy use in the global
[26] Cagnano A, De Tuglie E, Mancarella P. Microgrids: overview and guidelines for petroleum sector. Energy Jan. 2017;118:884–92. https://doi.org/10.1016/j.
practical implementations and operation. Appl Energy Jan. 2020;258:114039. energy.2016.10.107.
https://doi.org/10.1016/j.apenergy.2019.114039. [53] Absi Halabi M, Al-Qattan A, Al-Otaibi A. Application of solar energy in the oil
[27] Parhizi S, Lotfi H, Khodaei A, Bahramirad S. State of the art in research on industry—current status and future prospects. Renew Sustain Energy Rev Mar.
microgrids: a review. IEEE Access 2015;3:890–925. https://doi.org/10.1109/ 2015;43:296–314. https://doi.org/10.1016/j.rser.2014.11.030.
ACCESS.2015.2443119. [54] Ardal AR, Sharifabadi K, Bergvoll O, Berge V. Challenges with integration and
[28] Hirsch A, Parag Y, Guerrero J. Microgrids: a review of technologies, key drivers, operation of offshore oil & gas platforms connected to an offshore wind power
and outstanding issues. Renew Sustain Energy Rev Jul. 2018;90:402–11. https:// plant. In: 2014 Petroleum and chemical industry conference europe, amsterdam,
doi.org/10.1016/j.rser.2018.03.040. Netherlands; Jun. 2014. p. 1–9. https://doi.org/10.1109/
[29] ARENA, “Weipa solar farm - Australian Renewable Energy Agency (ARENA),” PCICEurope.2014.6900054.
Austr Renew Energy Agency. https://arena.gov.au/projects/weipa-solar-farm/ [55] Bigarelli L, Lidozzi A, di Benedetto M, Solero L, Bifaretti S. Model predictive
(accessed Apr. 23, 2019). energy management for sustainable off-shore oil and gas platforms. In: 2019 21st
[30] Riboldi L, Nord L. Offshore power plants integrating a wind farm: design European Conference on power Electronics and applications (EPE ’19 ECCE europe),
optimisation and techno-economic assessment based on surrogate modelling. genova, Italy; Sep. 2019. https://doi.org/10.23919/EPE.2019.8914962. p. P.1-
Processes Dec. 2018;6(12):249. https://doi.org/10.3390/pr6120249. P.10.
[31] Hamedani Golshan ME, Masoum MAS, Derakhshandeh SY. Profit-based unit [56] Hu D, Zhao X. Xu Cai, and Jianfeng Wang, “Impact of wind power on stability of
commitment with security constraints and fair allocation of cost saving in offshore platform power systems. In: 2008 third international conference on
industrial microgrids. IET Sci Meas Technol Nov. 2013;7(6):315–25. https://doi. electric utility deregulation and restructuring and power technologies, nanjing,
org/10.1049/iet-smt.2013.0044. China; Apr. 2008. p. 1688–92. https://doi.org/10.1109/DRPT.2008.4523677.
[32] Nosratabadi SM, Hooshmand R-A, Gholipour E, Rahimi S. Modeling and [57] A. R. Årdal, S. D’Arco, R. E. Torres-Olguin, T. Undeland, and K. Shrarifabadi,
simulation of long term stochastic assessment in industrial microgrids proficiency “Parametric sensitivity of transients in an islanded system with an offshore wind
considering renewable resources and load growth. Simulat Model Pract Theor farm connected to an oil platform,” p. 10.
Jun. 2017;75:77–95. https://doi.org/10.1016/j.simpat.2017.03.013. [58] Anglani N, Salvo SRD, Oriti G, Julian AL. Renewable energy sources and storage
[33] Soshinskaya M, Crijns-Graus WHJ, van der Meer J, Guerrero JM. Application of a integration in offshore microgrids. In: 2020 IEEE international Conference on
microgrid with renewables for a water treatment plant. Appl Energy Dec. 2014; Environment and electrical Engineering and 2020 IEEE Industrial and
134:20–34. https://doi.org/10.1016/j.apenergy.2014.07.097. commercial power systems europe (EEEIC/I CPS europe); Jun. 2020. p. 1–6.
[34] Subramanyam V, Jin T, Novoa C. Sizing a renewable microgrid for flow shop https://doi.org/10.1109/EEEIC/ICPSEurope49358.2020.9160760.
manufacturing using climate analytics. J Clean Prod Apr. 2020;252:119829. [59] Zhang P, Li W, Li S, Wang Y, Xiao W. Reliability assessment of photovoltaic power
https://doi.org/10.1016/j.jclepro.2019.119829. systems: review of current status and future perspectives. Appl Energy Apr. 2013;
[35] Aria. Base aria : perturbation d’alimentation électrique,” ministère du 104:822–33. https://doi.org/10.1016/j.apenergy.2012.12.010.
Développement durable. 2009 [Online]. Available: https://www.aria.developpe [60] Sangwongwanich A, Yang Y, Sera D, Blaabjerg F, Zhou D. On the impacts of PV
ment-durable.gouv.fr/wp-content/uploads/2013/07/09-10-analogies_perturbat array sizing on the inverter reliability and lifetime. IEEE Trans Ind Appl Jul. 2018;
ions_electrique.pdf. 54(4):3656–67. https://doi.org/10.1109/TIA.2018.2825955.
[36] IEEE Std 493-2007 Revis. IEEE Std 493-1997 IEEE recommended practice for the [61] Dobos A. PVWatts version 5 manual. NREL/TP-6A20-62641, 1158421. Sep. 2014.
design of reliable industrial and commercial power systems. Jun. 2007. p. 1–383. https://doi.org/10.2172/1158421.
https://doi.org/10.1109/IEEESTD.2007.380668. [62] Tiwari GN, Tiwari AK, Shyam. Handbook of solar energy: theory, analysis and
[37] Krasich M. How to estimate and use MTTF/MTBF would the real MTBF please applications. Singapore: Springer Singapore; 2016.
stand up?. In: 2009 annual Reliability and maintainability symposium; Jan. 2009. [63] Sengupta M, Andreas A. Oahu solar measurement grid (1-year archive): 1-second
p. 353–9. https://doi.org/10.1109/RAMS.2009.4914702. solar irradiance; oahu, Hawaii (data). 2010. https://doi.org/10.5439/1052451.
[38] Haes Alhelou H, Hamedani-Golshan ME, Njenda TC, Siano P. A survey on power [64] Stein J, Hansen C, Reno MJ. The variability index: a new and novel metric for
system blackout and cascading events: research motivations and challenges. quantifying irradiance and PV output variability. Sandia National Laboratories;
Energies Jan. 2019;12(4). https://doi.org/10.3390/en12040682. Art. no. 4. 2012.
[39] Cole MA, Elliott RJR, Occhiali G, Strobl E. Power outages and firm performance [65] Blaga R, Paulescu M. Quantifiers for the solar irradiance variability: a new
in Sub-Saharan Africa. J Dev Econ Sep. 2018;134:150–9. https://doi.org/ perspective. Sol Energy Nov. 2018;174:606–16. https://doi.org/10.1016/j.
10.1016/j.jdeveco.2018.05.003. solener.2018.09.034.
[40] Pasha HA, Ghaus A, Malik S. The economic cost of power outages in the industrial [66] Lave M. Solar variability zones: satellite-derived zones that represent high-
sector of Pakistan. Energy Econ Oct. 1989;11(4):301–18. https://doi.org/ frequency ground variability. Sol Energy 2017;10.
10.1016/0140-9883(89)90046-7. [67] Lauret P, Perez R, Mazorra Aguiar L, Tapachès E, Diagne HM, David M.
[41] Mashayekh S, et al. Security-constrained design of isolated multi-energy Characterization of the intraday variability regime of solar irradiation of
microgrids. IEEE Trans Power Syst May 2018;33(3):2452–62. https://doi.org/ climatically distinct locations. Sol Energy Feb. 2016;125:99–110. https://doi.
10.1109/TPWRS.2017.2748060. org/10.1016/j.solener.2015.11.032.
[42] Erdinc O, Paterakis NG, Catalão JPS. Overview of insular power systems under [68] Mills A, et al. Dark shadows. IEEE Power Energy Mag May 2011;9(3):33–41.
increasing penetration of renewable energy sources: opportunities and https://doi.org/10.1109/MPE.2011.940575.
challenges. Renew Sustain Energy Rev Dec. 2015;52:333–46. https://doi.org/ [69] Cormode D, Cronin AD, Richardson W, Lorenzo AT, Brooks AE, DellaGiustina DN.
10.1016/j.rser.2015.07.104. Comparing ramp rates from large and small PV systems, and selection of batteries
[43] Maïzi N, et al. Maximizing intermittency in 100% renewable and reliable power for ramp rate control. In: 2013 IEEE 39th photovoltaic specialists conference (PVSC),
systems: a holistic approach applied to Reunion Island in 2030. Appl Energy Oct. tampa, FL, USA; Jun. 2013. p. 1805–10. https://doi.org/10.1109/
2018;227:332–41. https://doi.org/10.1016/j.apenergy.2017.08.058. PVSC.2013.6744493.
[44] Anderson WW, Yakimenko OA. Comparative analysis of two microgrid solutions [70] Marcos J, Marroyo L, Lorenzo E, Alvira D, Izco E. From irradiance to output
for island green energy supply sustainability. In: 2017 IEEE 6th international power fluctuations: the PV plant as a low pass filter. Prog Photovoltaics Res Appl
conference on renewable energy research and applications (ICRERA); Nov. 2017. 2011;19(5):505–10. https://doi.org/10.1002/pip.1063.
p. 245–50. https://doi.org/10.1109/ICRERA.2017.8191274. [71] Hoff TE, Perez R. Quantifying PV power output variability. Sol Energy Oct. 2010;
[45] Favuzza S, et al. An analysis of the intertial response of small isolated power 84(10):1782–93. https://doi.org/10.1016/j.solener.2010.07.003.
systems in presence of generation from renewable energy sources. In: 2018 IEEE [72] van Haaren R, Morjaria M, Fthenakis V. Empirical assessment of short-term
4th international forum on research and technology for society and industry variability from utility-scale solar PV plants: assessment of variability from
(RTSI), palermo; Sep. 2018. p. 1–6. https://doi.org/10.1109/ utility-scale solar PV plants. Prog Photovoltaics Res Appl May 2014;22(5):
RTSI.2018.8548401. 548–59. https://doi.org/10.1002/pip.2302.
[46] Banjar-Nahor KM, Garbuio L, Debusschere V, Hadjsaid N, Pham T-T-H, [73] Lave M, Kleissl J, Stein JS. A wavelet-based variability model (WVM) for solar PV
Sinisuka N. Study on renewable penetration limits in a typical Indonesian power plants. IEEE Trans Sustain Energy Apr. 2013;4(2):501–9. https://doi.org/
islanded microgrid considering the impact of variable renewables integration and 10.1109/TSTE.2012.2205716.
the empowering flexibility on grid stability. In: 2018 IEEE PES innovative smart [75] Bright JM, Smith CJ, Taylor PG, Crook R. Stochastic generation of synthetic
grid technologies conference europe (ISGT-Europe), sarajevo, Bosnia and minutely irradiance time series derived from mean hourly weather observation
16
data. Sol Energy May 2015;115:229–42. https://doi.org/10.1016/j. California, USA, nov; 2013. https://doi.org/10.1115/IMECE2013-64102. p.
solener.2015.02.032. V04AT04A050.
[76] Fernández-Peruchena CM, Gastón M. A simple and efficient procedure for [105] Tsoutsanis E, Hamadache M, Dixon R. Real-time diagnostic method of gas
increasing the temporal resolution of global horizontal solar irradiance series. turbines operating under transient conditions in hybrid power plants. J Eng Gas
Renew Energy Feb. 2016;86:375–83. https://doi.org/10.1016/j. Turbines Power Oct. 2020;142(10):101002. https://doi.org/10.1115/1.4048340.
renene.2015.08.004. [106] Turie SE. “Gas turbine plant modeling for dynamic simulation. Master’s Thesis.
[77] Zhang W, Kleiber W, Florita AR, Hodge B-M, Mather B. Modeling and simulation Stockholm, Sweden: KTH School of Industrial Engineering and Management;
of high-frequency solar irradiance. IEEE J Photovolt Jan. 2019;9(1):124–31. 2012.
https://doi.org/10.1109/JPHOTOV.2018.2879756. [107] Pérez AAG. Modelling of a gas turbine with ModelicaTM. Master’s Thesis. Lund,
[78] Ahmed A, Khalid M. A review on the selected applications of forecasting models Sweden: Lund Institute of Technology; 2001.
in renewable power systems. Renew Sustain Energy Rev Feb. 2019;100:9–21. [108] Konečná E, Teng SY, Máša V. New insights into the potential of the gas
https://doi.org/10.1016/j.rser.2018.09.046. microturbine in microgrids and industrial applications. Renew Sustain Energy
[79] Cirés E, Marcos J, de la Parra I, García M, Marroyo L. The potential of forecasting Rev Dec. 2020;134:110078. https://doi.org/10.1016/j.rser.2020.110078.
in reducing the LCOE in PV plants under ramp-rate restrictions. Energy Dec. [109] Blondeau J, Mertens J. Impact of intermittent renewable energy production on
2019;188:116053. https://doi.org/10.1016/j.energy.2019.116053. specific CO2 and NO emissions from large scale gas-fired combined cycles.
[80] Antonanzas J, Osorio N, Escobar R, Urraca R, Martinez-de-Pison FJ, Antonanzas- J Clean Prod Jun. 2019;221:261–70. https://doi.org/10.1016/j.
Torres F. Review of photovoltaic power forecasting. Sol Energy Oct. 2016;136: jclepro.2019.02.182.
78–111. https://doi.org/10.1016/j.solener.2016.06.069. [110] Sujesh G, Ramesh S. Modeling and control of diesel engines: a systematic review.
[81] Vallance L. Synergie des mesures pyranométriques et des images hémisphériques Alex Eng J Dec. 2018;57(4):4033–48. https://doi.org/10.1016/j.
in-situ avec des images satellites météorologiques pour la prévision aej.2018.02.011.
photovoltaïque. phdthesis. PSL Research University; 2018 [Online]. Available: [111] Mahon LLJ. Diesel generator handbook. Oxford ; Boston: Butterworth-
https://pastel.archives-ouvertes.fr/tel-02097021. [Accessed 26 March 2020]. Heinemann; 1992.
[82] Lappalainen K, Wang GC, Kleissl J. Estimation of the largest expected [112] Heywood JB. Internal combustion engine fundamentals. New York: McGraw-Hill;
photovoltaic power ramp rates. Appl Energy 2020;278:115636. https://doi.org/ 1988.
10.1016/j.apenergy.2020.115636. Nov. [113] Finland Wartsila. Wärtsilä 32 Product guide. 2018. p. 284.
[83] SoDa Pro. HelioClim-3 real time and forecast. www.soda-pro.com. [Accessed 27 [114] L. Guzzella and C. H. Onder, “Introduction to modeling and control of internal
January 2020]. http://www.soda-pro.com/fr/web-services/radiation/helio combustion engine systems,” p. 362.
clim-3-real-time-and-forecast. [115] Rakopoulos, Constantine Giakoumis, Evangelos. Diesel Engine Transient
[84] Abuella M, Chowdhury B. Forecasting of solar power ramp events: a post- Operation : principles of operation and simulation analysis. Springer-Verlag
processing approach. Renew Energy Apr. 2019;133:1380–92. https://doi.org/ London; 2009.
10.1016/j.renene.2018.09.005. [116] Bastida H, Ugalde-Loo CE, Abeysekera M. Dynamic modelling and control of a
[85] Mathiesen P, Rife D, Collier C. Forecasting solar irradiance variability using the reciprocating engine. Energy Procedia Dec. 2017;142:1282–7. https://doi.org/
analog method. In: 2016 IEEE 43rd photovoltaic specialists conference (PVSC); 10.1016/j.egypro.2017.12.486.
Jun. 2016. p. 1207–11. https://doi.org/10.1109/PVSC.2016.7749806. [117] Kuznetsov A, Kharitonov S, Ryzhov V. Experimental and calculation study of
[86] Huang J, Davy RJ. Predicting intra-hour variability of solar irradiance using diesel-generator performance in transient conditions. J Eng Gas Turbines Power
hourly local weather forecasts. Sol Energy Dec. 2016;139:633–9. https://doi.org/ Oct. 2018;140(12):121009. https://doi.org/10.1115/1.4040690.
10.1016/j.solener.2016.10.036. [118] Semshchikov E, Hamilton J, Wu L, Negnevitsky M, Wang X, Lyden S. Frequency
[87] Chu Y, Pedro HTC, Li M, Coimbra CFM. Real-time forecasting of solar irradiance control within high renewable penetration hybrid systems adopting low load
ramps with smart image processing. Sol Energy Apr. 2015;114:91–104. https:// diesel methodologies. Energy Procedia Feb. 2019;160:483–90. https://doi.org/
doi.org/10.1016/j.solener.2015.01.024. 10.1016/j.egypro.2019.02.196.
[88] Gauchet C, Blanc P, Espinar B, Charbonnier B, Demengel D. Surface solar [119] Stavrakakis GS, Kariniotakis GN. A general simulation algorithm for the accurate
irradiance estimation with low-cost fish-eye camera. 2012. p. 5. assessment of isolated diesel-wind turbines systems interaction. I. A general
[89] Lonij VPA, Brooks AE, Cronin AD, Leuthold M, Koch K. Intra-hour forecasts of multimachine power system model. IEEE Trans Energy Convers Sep. 1995;10(3):
solar power production using measurements from a network of irradiance 577–83. https://doi.org/10.1109/60.464885.
sensors. Sol Energy Nov. 2013;97:58–66. https://doi.org/10.1016/j. [120] Yu M, Tang X, Lin Y, Wang X. Diesel engine modeling based on recurrent neural
solener.2013.08.002. networks for a hardware-in-the-loop simulation system of diesel generator sets.
[90] Bosch JL, Zheng Y, Kleissl J. Deriving cloud velocity from an array of solar Neurocomputing Mar. 2018;283:9–19. https://doi.org/10.1016/j.
radiation measurements. Sol Energy Jan. 2013;87:196–203. https://doi.org/ neucom.2017.12.054.
10.1016/j.solener.2012.10.020. [121] Andric J, Schimmel D, Sediako AD, Sjoblom J, Faghani E. Development and
[91] ETN Global. ETN R&D recommendation report. ETN Global; Oct. 2018 [Online]. calibration of one dimensional engine model for hardware-in-the-loop
Available: https://etn.global/research-innovation/. applications. 2018. p. 8.
[92] Kubik ML, Coker PJ, Barlow JF. Increasing thermal plant flexibility in a high [122] Philippe Beaulac. Jumelage éolien-diesel avec stockage d’air comprimé. 2014.
renewables power system. Appl Energy Sep. 2015;154:102–11. https://doi.org/ [123] Nikolic D, Negnevitsky M, de Groot M. Effect of the diesel engine delay on
10.1016/j.apenergy.2015.04.063. stability of isolated power systems with high levels of renewable energy
[93] Warstila. Combustion engine vs gas turbine: ramp rate. 2019 [Online]. Available: penetration. In: 2015 international Symposium on smart electric distribution Systems
https://www.wartsila.com/energy/learn-more/technical-comparisons/combusti and technologies (EDST), vienna, Austria; Sep. 2015. p. 70–3. https://doi.org/
on-engine-vs-gas-turbine-ramp-rate. 10.1109/SEDST.2015.7315185.
[94] Saravanamuttoo HIH, Cohen H, Rogers GFC. Gas turbine theory. Harlow: Prentice [124] Hamilton J, Negnevitsky M, Wang X. The potential of variable speed diesel
Hall; 2001. application in increasing renewable energy source penetration. Energy Procedia
[95] Razak AMY. Industrial gas turbines: performance and operability. Boca Raton, Feb. 2019;160:558–65. https://doi.org/10.1016/j.egypro.2019.02.206.
Fla: CRC Press [u.a.; 2007. [125] Hamilton J, Negnevitsky M, Wang X, Lyden S. High penetration renewable
[96] Jansohn P. Modern gas turbine systems: high efficiency, low emission, fuel generation within Australian isolated and remote power systems. Energy Feb.
flexible power generation. 2016. 2019;168:684–92. https://doi.org/10.1016/j.energy.2018.11.118.
[97] Thermoflow. Thermoflow, inc. (home). https://www.thermoflow.com/. [126] Hamilton J, Tavakoli A, Negnevitsky M, Wang X. Investigation of no load diesel
[Accessed 23 April 2019]. technology in isolated power systems. In: 2016 IEEE power and energy society
[98] E. A. EA Internacional. EcosimPro | PROOSIS modelling and simulation software. general meeting (PESGM), boston, MA, USA; Jul. 2016. p. 1–5. https://doi.org/
EcosimPro | PROOSIS, https://www.ecosimpro.com/. [Accessed 23 April 2019]. 10.1109/PESGM.2016.7741324.
[99] Alobaid F, Mertens N, Starkloff R, Lanz T, Heinze C, Epple B. Progress in dynamic [127] Benato A, Bracco S, Stoppato A, Mirandola A. LTE: a procedure to predict power
simulation of thermal power plants. Prog Energy Combust Sci Mar. 2017;59: plants dynamic behaviour and components lifetime reduction during transient
79–162. https://doi.org/10.1016/j.pecs.2016.11.001. operation. Appl Energy Jan. 2016;162:880–91. https://doi.org/10.1016/j.
[100] Mehrpanahi A, Hamidavi A, Ghorbanifar A. A novel dynamic modeling of an apenergy.2015.10.162.
industrial gas turbine using condition monitoring data. Appl Therm Eng Oct. [128] Faisal M, Hannan MA, Ker PJ, Hussain A, Mansor MB, Blaabjerg F. Review of
2018;143:507–20. https://doi.org/10.1016/j.applthermaleng.2018.07.081. energy storage system technologies in microgrid applications: issues and
[101] Yee SK, Milanovic JV, Hughes FM. Overview and comparative analysis of gas challenges. IEEE Access 2018;6:35143–64. https://doi.org/10.1109/
turbine models for system stability studies. IEEE Trans Power Syst Feb. 2008;23 ACCESS.2018.2841407.
(1):108–18. https://doi.org/10.1109/TPWRS.2007.907384. [129] Krishan O, Suhag S. An updated review of energy storage systems: Classification
[102] Tsoutsanis E, Meskin N. Dynamic performance simulation and control of gas and applications in distributed generation power systems incorporating
turbines used for hybrid gas/wind energy applications. Appl Therm Eng Jan. renewable energy resources. Int J Energy Res 2019;43(12):6171–210. https://doi.
2019;147:122–42. https://doi.org/10.1016/j.applthermaleng.2018.09.031. org/10.1002/er.4285.
[103] Chaibakhsh A, Amirkhani S. A simulation model for transient behaviour of heavy- [130] Lund PD, Lindgren J, Mikkola J, Salpakari J. Review of energy system flexibility
duty gas turbines. Appl Therm Eng Mar. 2018;132:115–27. https://doi.org/ measures to enable high levels of variable renewable electricity. Renew Sustain
10.1016/j.applthermaleng.2017.12.077. Energy Rev May 2015;45:785–807. https://doi.org/10.1016/j.rser.2015.01.057.
[104] Tsoutsanis E, Meskin N, Benammar M, Khorasani K. Dynamic performance [131] IRENA. Electricity storage : costs and market to 2030. Oct. 2017.
simulation of an aeroderivative gas turbine using the matlab simulink [132] Neoen Australia. HPR fully installed and energised ! – hornsdale Power Reserve.
environment. In: Volume 4A: dynamics, vibration and control, san diego, Nov. 28. 2019. https://hornsdalepowerreserve.com.au/hpr-fully-installed-an
d-energised/. [Accessed 28 April 2020].
17
[133] Hesse H, Schimpe M, Kucevic D, Jossen A. Lithium-ion battery storage for the [158] Dreidy M, Mokhlis H, Mekhilef S. Inertia response and frequency control
grid—a review of stationary battery storage system design tailored for techniques for renewable energy sources: a review. Renew Sustain Energy Rev
applications in modern power grids. Energies Dec. 2017;10(12):2107. https:// Mar. 2017;69:144–55. https://doi.org/10.1016/j.rser.2016.11.170.
doi.org/10.3390/en10122107. [159] Hoke A, Maksimović D. Active power control of photovoltaic power systems. In:
[134] Arani AAK, Karami H, Gharehpetian GB, Hejazi MSA. Review of Flywheel Energy 2013 1st IEEE conference on technologies for sustainability (SusTech); Aug. 2013.
Storage Systems structures and applications in power systems and microgrids. p. 70–7. https://doi.org/10.1109/SusTech.2013.6617300.
Renew Sustain Energy Rev Mar. 2017;69:9–18. https://doi.org/10.1016/j. [160] G. Dickeson, L. McLeod, A. Dobb, L. Frearson, B. Herteleer, and D. Scheltus,
rser.2016.11.166. “Ramp rate control for PV plant integration: experience from karratha AIRPORT’S
[135] Ramli MAM, Hiendro A, Twaha S. Economic analysis of PV/diesel hybrid system hybrid power station,” p. 6.
with flywheel energy storage. Renew Energy Jun. 2015;78:398–405. https://doi. [161] Matevosyan J, et al. Grid-forming inverters: are they the key for high renewable
org/10.1016/j.renene.2015.01.026. penetration? IEEE Power Energy Mag Nov. 2019;17(6):89–98. https://doi.org/
[136] Habib HF, Mohamed AAS, El Hariri M, Mohammed OA. Utilizing supercapacitors 10.1109/MPE.2019.2933072.
for resiliency enhancements and adaptive microgrid protection against [162] Sasmono S, Chandra JA, Wijaya A, Hariyanto N. Transient stability method to
communication failures. Elec Power Syst Res Apr. 2017;145:223–33. https://doi. determine maximum intermittent power plant penetration to small island isolated
org/10.1016/j.epsr.2016.12.027. system. In: 2018 IEEE PES asia-pacific power and energy engineering conference
[137] Akram U, Khalid M, Shafiq S. An innovative hybrid wind-solar and battery- (APPEEC), kota kinabalu; Oct. 2018. p. 498–501. https://doi.org/10.1109/
supercapacitor microgrid system—development and optimization. IEEE Access APPEEC.2018.8566468.
2017;5:25897–912. https://doi.org/10.1109/ACCESS.2017.2767618. [163] Rossi R, Masoum MAS. Application of swing equation and equal area criteria to
[138] Mohammadi E, Rasoulinezhad R, Moschopoulos G. Using a supercapacitor to determine maximum penetration of wind generation in an islanded network. In:
mitigate battery microcycles due to wind shear and tower shadow effects in wind- 2013 australasian universities power engineering conference (AUPEC), hobart,
diesel microgrids. IEEE Trans. Smart Grid Sep. 2020;11(5):3677–89. https://doi. Australia; Sep. 2013. p. 1–8. https://doi.org/10.1109/AUPEC.2013.6725387.
org/10.1109/TSG.2020.2979140. [164] Magro MC, Giannettoni M, Pinceti P, Vanti M. Real time simulator for microgrids.
[139] Meyer B, Jerosolimski M, Stubbe M. Outils de simulation dynamique des réseaux Elec Power Syst Res Jul. 2018;160:381–96. https://doi.org/10.1016/j.
électriques. Tech. Ing. Réseaux Électr. Transp. Répart., vol. base documentaire : epsr.2018.03.018.
TIB263DUO., no. ref. article : d4120. 1998 [Online]. Available: https://www.tech [165] Polleux L, Sandoval-Moreno J, Guerassimoff G, Marmorat J-P. Impact of thermal
niques-ingenieur.fr/base-documentaire/energies-th4/reseaux-electriques-de generation flexibility on power quality and LCOE of industrial off-grid power
-transport-et-de-repartition-42263210/outils-de-simulation-dynamique-des-rese plants. In: Proceedings of 11th international Conference on applied energy, Part
aux-electriques-d4120/. 2. vol. 3. Sweden: Västeraas; Aug. 2019.
[140] Kundur PS. Power system stability and control. Indian edition. Mc Graw Hill [166] Mohammed A, Pasupuleti J, Khatib T, Elmenreich W. A review of process and
Education (India) Private Limited; 1994. operational system control of hybrid photovoltaic/diesel generator systems.
[141] The shift project, “power systems 2050 – power system operation V1.1. Feb. 2020. Renew Sustain Energy Rev Apr. 2015;44:436–46. https://doi.org/10.1016/j.
[142] IRENA. Planning for the renewable future : long-term modelling and tools to rser.2014.12.035.
expand variable renewable power in emerging economies. Abu Dhabi: [167] Kanchev H, Lu D, Colas F, Lazarov V, Francois B. Energy management and
International Renewable Energy Agency (IRENA); 2017 [Online]. Available: http: operational planning of a microgrid with a PV-based active generator for smart
//www.irena.org/DocumentDownloads/Publications/IRENA_REthinking_Energy grid applications. IEEE Trans Ind Electron Oct. 2011;58(10):4583–92. https://
_2017.pdf. [Accessed 11 May 2020]. doi.org/10.1109/TIE.2011.2119451.
[143] Fraisse J-L, Karsenti L. Raccordement de la production décentralisée aux réseaux [168] Chaouachi A, Kamel RM, Andoulsi R, Nagasaka K. Multiobjective intelligent
de distribution - conditions d’intégration. 2014. p. 41. energy management for a microgrid. IEEE Trans Ind Electron Apr. 2013;60(4):
[144] Uriarte FM, Smith C, VanBroekhoven S, Hebner RE. Microgrid ramp rates and the 1688–99. https://doi.org/10.1109/TIE.2012.2188873.
inertial stability margin. IEEE Trans Power Syst Nov. 2015;30(6):3209–16. [169] Wang T, He X, Deng T. Neural networks for power management optimal strategy
https://doi.org/10.1109/TPWRS.2014.2387700. in hybrid microgrid. Neural Comput Appl Jul. 2019;31(7):2635–47. https://doi.
[145] Soni N, Doolla S, Chandorkar MC. Improvement of transient response in org/10.1007/s00521-017-3219-x.
microgrids using virtual inertia. IEEE Trans Power Deliv Jul. 2013;28(3):1830–8. [170] Askarzadeh A. A memory-based genetic algorithm for optimization of power
https://doi.org/10.1109/TPWRD.2013.2264738. generation in a microgrid. IEEE Trans Sustain Energy Jul. 2018;9(3):1081–9.
[146] Katiraei F, Iravani MR. Power management strategies for a microgrid with https://doi.org/10.1109/TSTE.2017.2765483.
multiple distributed generation units. IEEE Trans Power Syst Nov. 2006;21(4): [171] Maulik A, Das D. Optimal operation of microgrid using four different optimization
1821–31. https://doi.org/10.1109/TPWRS.2006.879260. techniques. Sustain Energy Technol Assess Jun. 2017;21:100–20. https://doi.org/
[147] Guerrero JM, Vasquez JC, Matas J, de Vicuna LG, Castilla M. Hierarchical control 10.1016/j.seta.2017.04.005.
of droop-controlled AC and DC microgrids—a general approach toward [172] Marzband M, Azarinejadian F, Savaghebi M, Guerrero JM. An optimal energy
standardization. IEEE Trans Ind Electron Jan. 2011;58(1):158–72. https://doi. management system for islanded microgrids based on multiperiod artificial bee
org/10.1109/TIE.2010.2066534. colony combined with Markov chain. IEEE Syst J Sep. 2017;11(3):1712–22.
[148] Arteaga CC. Optimisation of power system security with high share of variable https://doi.org/10.1109/JSYST.2015.2422253.
renewables: consideration of the primary reserve deployment dynamics on a [173] Ju C, Wang P, Goel L, Xu Y. A two-layer energy management system for
Frequency Constrained Unit Commitment model. Paris, France: Thèse de microgrids with hybrid energy storage considering degradation costs. IEEE Trans.
doctorat, Univertié Paris-Saclay; 2016. Smart Grid Nov. 2018;9(6):6047–57. https://doi.org/10.1109/
[149] Morstyn T, Hredzak B, Agelidis VG. Control strategies for microgrids with TSG.2017.2703126.
distributed energy storage systems: an overview. IEEE Trans. Smart Grid Jul. [174] Lara JD, Olivares DE, Cañizares CA. Robust energy management of isolated
2018;9(4):3652–66. https://doi.org/10.1109/TSG.2016.2637958. microgrids. IEEE Syst J Mar. 2019;13(1):680–91. https://doi.org/10.1109/
[150] Li J, Xiong R, Yang Q, Liang F, Zhang M, Yuan W. Design/test of a hybrid energy JSYST.2018.2828838.
storage system for primary frequency control using a dynamic droop method in an [175] Mohiti M, Mazidi M, Moghaddam AA, Guerrero JM. Microgrid optimal energy
isolated microgrid power system. Appl Energy Sep. 2017;201:257–69. https:// and reserve scheduling considering frequency constraints. In: 2019 international
doi.org/10.1016/j.apenergy.2016.10.066. Conference on smart energy Systems and technologies (SEST); Sep. 2019. p. 1–6.
[151] Serban I, Marinescu C. Control strategy of three-phase battery energy storage https://doi.org/10.1109/SEST.2019.8849072.
systems for frequency support in microgrids and with uninterrupted supply of [176] N’Goran A, Daugrois B, Lotteau M, Demassey S. Optimal engagement and
local loads. IEEE Trans Power Electron Sep. 2014;29(9):5010–20. https://doi. operation of a grid-connected PV/battery system. In: 2019 IEEE PES innovative
org/10.1109/TPEL.2013.2283298. smart grid technologies europe (ISGT-Europe); Sep. 2019. p. 1–5. https://doi.org/
[152] Toge M, Kurita Y, Iwamoto S. Supplementary load frequency control with storage 10.1109/ISGTEurope.2019.8905617.
battery operation considering SOC under large-scale wind power penetration. In: [177] Olivares DE, Canizares CA, Kazerani M. A centralized energy management system
2013 IEEE power & energy society general meeting, vancouver, BC; 2013. p. 1–5. for isolated microgrids. IEEE Trans. Smart Grid Jul. 2014;5(4):1864–75. https://
https://doi.org/10.1109/PESMG.2013.6672323. doi.org/10.1109/TSG.2013.2294187.
[153] Suvire GO, Molina MG, Mercado PE. Improving the integration of wind power [178] Palma-Behnke R, et al. A microgrid energy management system based on the
generation into AC microgrids using flywheel energy storage. IEEE Trans. Smart rolling horizon strategy. IEEE Trans. Smart Grid Jun. 2013;4(2):996–1006.
Grid Dec. 2012;3(4):1945–54. https://doi.org/10.1109/TSG.2012.2208769. https://doi.org/10.1109/TSG.2012.2231440.
[154] Abbey C, Joos G. Supercapacitor energy storage for wind energy applications. [179] Rezaei N, Kalantar M. Economic–environmental hierarchical frequency
IEEE Trans Ind Appl May 2007;43(3):769–76. https://doi.org/10.1109/ management of a droop-controlled islanded microgrid. Energy Convers Manag
TIA.2007.895768. Dec. 2014;88:498–515. https://doi.org/10.1016/j.enconman.2014.08.054.
[155] Gu W, Liu W, Shen C, Wu Z. Multi-stage underfrequency load shedding for [180] Cominesi S, Farina M, Giulioni L, Picasso B, Scattolini R. A two-layer stochastic
islanded microgrid with equivalent inertia constant analysis. Int J Electr Power model predictive control scheme for microgrids. IEEE Trans Control Syst Technol
Energy Syst Mar. 2013;46:36–9. https://doi.org/10.1016/j.ijepes.2012.10.003. Jan. 2018;26(1):1–13. https://doi.org/10.1109/TCST.2017.2657606.
[156] Hajiakbari Fini M, Hamedani Golshan ME. Frequency control using loads and [181] Cominesi SR, Bella AL, Farina M, Scattolini R. A multi-layer control scheme for
generators capacity in power systems with a high penetration of renewables. Elec microgrid energy management. IFAC-Pap. 2016;49(27):256–61. https://doi.org/
Power Syst Res Jan. 2019;166:43–51. https://doi.org/10.1016/j. 10.1016/j.ifacol.2016.10.700.
epsr.2018.09.010. [182] Hu J, Shan Y, Guerrero JM, Ioinovici A, Chan KW, Rodriguez J. Model predictive
[157] Arif A, Wang Z, Wang J, Mather B, Bashualdo H, Zhao D. Load modeling—a control of microgrids – an overview. Renew Sustain Energy Rev Feb. 2021;136:
review. IEEE Trans. Smart Grid Nov. 2018;9(6):5986–99. https://doi.org/ 110422. https://doi.org/10.1016/j.rser.2020.110422.
10.1109/TSG.2017.2700436.
18
[183] Mayhorn E, Zhang W, Lu S, Samaan N. Optimal control of distributed energy [188] Rigo-Mariani R, Sareni B, Roboam X. Integrated optimal design of a smart
resources using model predictive control. IEEE Power Energy Soc Gen Meet 2012: microgrid with storage. IEEE Trans. Smart Grid Jul. 2017;8(4):1762–70. https://
8. https://doi.org/10.1109/PESGM.2012.6345596. doi.org/10.1109/TSG.2015.2507131.
[184] Moretti L, Polimeni S, Meraldi L, Raboni P, Leva S, Manzolini G. Assessing the [189] Schittekatte T, Stadler M, Cardoso G, Mashayekh S, Sankar N. The impact of
impact of a two-layer predictive dispatch algorithm on design and operation of short-term stochastic variability in solar irradiance on optimal microgrid design.
off-grid hybrid microgrids. Renew Energy Dec. 2019;143:1439–53. https://doi. IEEE Trans. Smart Grid May 2018;9(3):1647–56. https://doi.org/10.1109/
org/10.1016/j.renene.2019.05.060. TSG.2016.2596709.
[185] Wang B. Intelligent control and power flow optimization of microgrid. ” [190] Khodaei A, Bahramirad S, Shahidehpour M. Microgrid planning under
Université de Technologie de Compiègne; 2013 [Online]. Available: https://pdfs. uncertainty. IEEE Trans Power Syst Sep. 2015;30(5):2417–25. https://doi.org/
semanticscholar.org/b2ad/6d61acfceef1969a484f0d895b321f416a0e.pdf. 10.1109/TPWRS.2014.2361094.
[186] Krakowski V. Intégration du renouvelable et stratégie de déploiement du réseau [191] Billionnet A, Costa M-C, Poirion P-L. Robust optimal sizing of a hybrid energy
électrique: réconciliation d’échelles spatio-temporelles dans des exercices stand-alone system. Eur J Oper Res Oct. 2016;254(2):565–75. https://doi.org/
prospectifs de long terme. Paris Sciences et Lettres; 2016. 10.1016/j.ejor.2016.03.013.
[187] Drouineau M, Maïzi N, Mazauric V. Impacts of intermittent sources on the quality [192] Chalil Madathil S, et al. Resilient off-grid microgrids: capacity planning and N-1
of power supply: the key role of reliability indicators. Appl Energy Mar. 2014;116: security. IEEE Trans. Smart Grid Nov. 2018;9(6):6511–21. https://doi.org/
333–43. https://doi.org/10.1016/j.apenergy.2013.11.069. 10.1109/TSG.2017.2715074.
19

Renewable and Sustainable Energy Reviews

Uploaded by

Copyright:

Available Formats

Renewable and Sustainable Energy Reviews

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Renewable and Sustainable Energy Reviews

Uploaded by

Copyright:

Available Formats

Renewable and Sustainable Energy Reviews 155 (2022) 111955

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

An overview of the challenges of solar power integration in isolated

these applications, microgrids will play a key role.

List of abbreviations IEA International energy agency

Application Mining Offshore oil Industry Industry Water treatment Manufacturing

On grid/off grid Off-grid Off-grid Both Off-grid

Renewable Solar Wind Wind Wind Wind

Storage – – Battery Battery

Fig. 2. Two irradiance profiles captured with 1-s sampling [63].

2.2. Fossil generation

With the penetration of variable and uncertain renewable power,

Fig. 4. Twin-shaft turbine [94].

perform a gas turbine diagnostic under transient operation. Reference

Fig. 8. Principle of ICE power conversion from combustion to electricity supply.

based on thermodynamic expressions, enabling a fairly accurate study of

2.2.2.4. Reduction to a time constant. From a control-science perspec­

Fig. 11. Conventional diesel engine representation [118].

shown that a high penetration rate can be enhanced by low-loading

2.2.3. Aging and maintenance considerations related to PV variability

The performance of a microgrid is closely related to its control

3.1.1. Fundamentals of frequency and voltage regulation

Fig. 14. Categories of storage technology [129].

different droop characteristics participate in the frequency regulation.

3.1.2. Control of fossil engines

3.1.3. Control of storage systems

capacity, minimum value of inertia, etc.).

You might also like

2.2.2.4. Reduction to a time constant. From a control-science perspec