0% found this document useful (0 votes)
5 views88 pages

Document CT

Download as pdf or txt
Download as pdf or txt
Download as pdf or txt
You are on page 1/ 88

www.automationforum.

co

CONTROL SYSTEM INSTRUMENTATION


www.automationforum.co

Control system instrumentation


Having considered PID controllers, we now consider the
other components of the feedback control loop. Consider
the stirred-tank heating system:
www.automationforum.co

Control system instrumentation

A thermocouple measures the liquid temperature and converts it


to millivolt-level electrical signal.

This signal is then amplified to a voltage level and transmitted to


the electronic controller.

The feedback controller performs the control calculations and


sends the calculations and sends the calculated value as an output
signal to the final control element, an electrical heater that adjusts
the rate of heat transfer to the liquid.
www.automationforum.co

This example illustrates three important functions of a


feedback control loop:

1)measurement of the controlled variable;


2)adjustment of the manipulated variable;
3)signal transmission between components.
The interconnection www.automationforum.co
between the
process and the controller can be
considered to be an interface (analog
or digital). The interconnection is
required for a single controller or for a
number of controllers in a computer
control system. In each case, the
interface consists of all measurement,
manipulation, and transmission
elements.
www.automationforum.co

SENSORS,TRANSMITTERS,ANDTRANSDUCERS
www.automationforum.co

INTRODUCTION
The operation of complex industrial plants would be difficult, if not
impossible, without the measurement and control of critical process
variables.
Large plants typically have hundreds or thousands of process
variables that are repetitively measured online every few seconds or
minutes.
In addition, important product properties are measured in quality
control labs less frequently – e.g. once per hour, once an eight hour
shift, or daily.
Consequently, the design and maintenance of accurate, reliable
measurement system is a critical aspect of process control.

The lack of a reliable, cost-effective on-line sensor can be a key


limitation on the effectiveness of a process control system.
www.automationforum.co

Sensors and transmitters


A physical variable is measured by a sensor which produces a physical
response (e.g. electrical or mechanical) that is related to the value of
the process variable.

For example in the stirred tank heating system, the thermocouple


generates a millivolt electrical signal that increases as the
temperature increases. However for this temperature measurement
to be used in the control calculations, the millivolt-level signal must be
converted to an appropriate voltage or current signal in a standard
input range for the controller.

This conversion is done by a transmitter.

In process control literature, the terms sensor, transmitter, and sensor-


transmitter, are used more or less interchangeably.
www.automationforum.co

tRANSDUCERS
It is often necessary to convert an instrumentation signal from one
form to another. A device that performs this conversion is referred to
as a transducer.

One common application is when the controller output signal is a


current signal and the final control element is a pneumatic control
valve. The required conversion is performed by a current-to-pressure
(I/P) transducer. Voltage to pressure (E/P) are also quite common.
www.automationforum.co

Standard instrumentation signal


levels
Before 1960, instrumentation in the process industries utilized
pneumatic (air pressure) signals to transmit measurement and
control information almost exclusively. These devices make use of
mechanical force-balance elements to generate signals in the
range of 3 to 15 psig, an industry standard.

Since about 1960, electronic instrumentation has become


predominant. The standard signal ranges for analog instruments
are 4 to 20mA and 1 to 5 V, direct current (VDC).
www.automationforum.co

SENSORS – SELECTION
CRITERIA
The selection of a measurement device should consider the
following factors:
1.Measurement range (span). The required measurement for the
process variable must lie entirely with the instrument’s range of
performance.

2.Performance. Depending on the application, accuracy,


repeatability, or some other measure of performance is
appropriate. For closed-loop control, speed of response is also
important.
www.automationforum.co

3.Reliability. Manufacturers provide baseline conditions. Previous


experience with the measurement device is very important.
4.Materials of Construction. The instrument may need to
withstand high temperatures, high pressures, and corrosive and
abrasive environments. For some applications, seals or purges
may be necessary.
5.Prior Use. For the first installation of a specific measurement
device at a site, training of maintenance personnel and purchases
of spare parts might be necessary.
www.automationforum.co

6.Potential for releasing process materials to the environment.


Preventing exposure to fugitive emissions for maintenance
personnel is important when the process fluid is corrosive or toxic.
Sterility in bioprocesses must be maintained.
*If the sensor is not inherently compatible with possible exposure
to hazards, suitable enclosures must be purchased and included in
the installation costs.
7.Electrical classification
8.Invasive or non-invasive. The insertion of a probe (invasive) can
cause fouling, which leads to inaccurate measurements. Probe
location must be selected carefully to ensure measurement
accuracy and minimize fouling.
www.automationforum.co

Typical sensors
www.automationforum.co
www.automationforum.co
www.automationforum.co

Sensors – static and dynamic


characteristics
The output signal from a sensor-transmitter (or
transmitter) must be compatible with the input range of
the controller that receives the signal. Transmitters are
generally designed to be direct-acting; that is, the output
signal increases as the measured variable increases. In
addition, most commercial transmitters have an
adjustable input range.
www.automationforum.co

For example, a temperature transmitter might be adjusted so that the


input range of a platinum resistance element (the sensor) is 50-150 Oc.
In this case, the following correspondence is obtained:
Input Output
50 OC 4mA
150 OC 20mA

This instrument has a lower limit, or zero, of 50 C and range, or span,


of 150-50 = 100 C.

Note that the transmitter is designed for a specific type of sensor;


hence the zero and span of the overall sensor/transmitter are
adjustable.
www.automationforum.co

In this example, the relation


between temperature and
the transmitted (measured)
signal is linear.

If the sensor power fails, the transmitter output signal has a value of OmA , which would move
the controller output to and final control element to their minimum or maximum values. If this
action could lead to an unsafe condition, the transmitter output signal could be inverted to give
the highest value in operating range.
www.automationforum.co

For this temperature transmitter, the relation between the output


and input is

The gain of the measurement element Km is 0.16mA/OC. For any


linear instrument

Km= output range / span


www.automationforum.co
1. Pressure Transmitter:
• Consider a pressure that varies
from 100 to 1000 kPa g.
• The output signal (current signal)
varies from 4 to 20 mA.
• The range of the transmitter = 100
to 1000 kPa g
www.automationforum.co

• The span of the transmitter =


1000 – 100 = 900 kPa g
• The zero of the transmitter =
low range = 100 kPa g

• The gain of the transmitter

20 mA – 4 mA = 16 mA .
1000 kPa – 100 kPa 900 kPa
www.automationforum.co
• The transmitter has 2 knobs which can be
adjusted to change its span or its zero.

• Consider shifting the zero from 100 to 200


kPa g without changing its span.

• The range of the transmitter =


200 – 1100 kPa
• The span = 900 kPa
www.automationforum.co
2. Temperature Transmitter:
• Consider a temperature transmitter
accepting a thermocouple input
signals.

• Its current output is 4 to 20 mA as the


process temperature changes from 50
to 250 oF.
www.automationforum.co

• The range of the temperature transmitter =


50 to 250 oF
• The span of the transmitter = 200 oF
• The zero = 50 oF
• The gain of the temperature transmitter
= 20 – 4 mA = 16 mA
250 – 50 oF 200 oF
www.automationforum.co

3. Fluid Flow with a P transmitter


• orifice plate as a flow transmitter
• Consider P over the orifice plate =
100 in H2O at a process flowrate of
2000 kg/hr
www.automationforum.co

• P transmitter converts
inches H2O into mA

• The gain of the transmitter


= 16 mA .

100 in H2O
www.automationforum.co

Dynamic characteristics of
sensor-transmitters.
Many sensor-transmitters respond quickly to and have measurement
dynamics that are negligible compared to slower process dynamics.

For other applications where measurement dynamics are not


negligible, significant dynamic errors can occur (large differences in
true values and measured values)
www.automationforum.co

Many sensor-transmitters have overdamped dynamics and


exhibit monotonic responses to a step change in the
variable being measured. Thus it is reasonable to model
this type of measurement dynamics as a first order TF
between the actual value y and the measured value ym:
where Km is the gain

and τm is theYmeasurement
m (s) K m time constant.
=
Y (s)  ms +1
www.automationforum.co

Significant measurement dynamics can occur due to poor sensor


location or a long sampling line.

Example, if a pH sensor for a continuous neutralization process is


located in the exit line, a long distance from the process vessel, a
significant delay can arise due to the distance-velocity lag.

Time delays can also result when an on-line composition


measurement requires a long sample line because of an expensive
analyzer in a protected environment is located a long distance from
the sample location near the process unit. This common situation
can produce a significant distance velocity-lag.
www.automationforum.co

FINAL CONTROL ELEMENTS


www.automationforum.co

Final control elements


Every process control loop contains a final control element (or
actuator), the device that enables a process variable to be
manipulated. For most chemical and petroleum processes, the final
control elements (usually control valves) adjust the flow rates of
materials –solid, liquid, and gas feeds and products – and indirectly,
the rates of energy transfer to and from the process.
www.automationforum.co

Control valves
There are many different ways to manipulate the flows of
material and energy into or out of a process; for example,
the speed of a pump drive, screw conveyor, or blower can
be varied.

However, a simple and widely used method of


accomplishing this result with fluids is the use of a control
valve.
www.automationforum.co

Control valves
The control valve components include the valve body, trim, seat, and
actuator.

Valve body - contains an orifice that allows for the flow of liquids and/or gases;

Trim/Plug - modulates the flow rate and can be a plug, ball, disk, or gate;

Seat - consists of protective material (typically metal or soft polymer) inserted


around the orifice to provide a tight shutoff and to increase the life of the valve
when corrosive or solid materials pass through it;

Actuator - provides the force for opening and closing the valve;
www.automationforum.co

Other key components of a valve


Bonnet – is the top portion of the valve body and connects the valve body to
the actuator and can be removed to allow entry into the valve body; usually
contains the packing box that provides a seal around the sliding stem
mechanism that connects the actuator to the valve plug.

Spring – provides the energy to move the valve in the opposite direction of the
diaphragm loading motion; this provision is made so that the valve can be
opened and closed proportionally with the instrument signal; the spring
provides the energy to return the valve back to its fail-safe condition.

Handwheel – is an actuator accessory that is used to manually override the


actuator or to limit its motion; may be located on the top of the actuator or on
its side.
www.automationforum.co

Control valve key


components
www.automationforum.co

Control valves
Control valves are either linear (rising stem /
reciprocating) or rotary in design.
Linear valves – usually globe valves that open and close the valve
by moving a plug vertically away from the orifice and the seat; this
movement changes cross sectional area available for fluid flow.
Rotary valves – are closed by a 90o turn of the closing element
(also called quarter-turn valves); these valves are used both for on
off and flow modulating control valves.
www.automationforum.co

Control valves
Control valves typically utilize some type of mechanical driver to
move the valve plug into and out of its seat, it can either be:
A DC motor or a stepping motor that screws the valve stem in and out in much
the same way as a hand valve would be operated.
A pneumatically operated diaphragm device that moves the stem vertically
against the opposing force of a fixed spring, called a rising stem valve.
www.automationforum.co

Control valves

- a pneumatic control valve (air to open)


www.automationforum.co

Control valve failure conditions


Despite the growing use of motor-driven valves, most control
applications utilize pneumatically driven control valves of the rising
stem type.
Air-to-Open (A-O). As the pneumatic controller output signal
increases, increased pressure on the diaphragm compresses the
spring, thus pulling the stem out and opening the valve further. (fail-
close)
Air-to-Close (A-C). By reversing either the plug/seat or the spring/air
inlet orientation, the valve becomes air-to-close. (fail-open)
www.automationforum.co

Control valve failure conditions


Normally the choice of A-O or A-C valve is based on safety
considerations.
The way the valve should operate is chosen based on the desired
response in an emergency situation (full flow or no flow).
Hence A-C and A-O valves are often referred to as fail-open (FO)
and fail-close(FC) respectively.
www.automationforum.co

Types of control valves - Linear


Globe valves are a family of valves characterized by a closure
member that travels in a line perpendicular to the valve seat.
They are used primarily for throttling purposes and general flow
control.
They are the most common type of valve used in the processing
industry.
www.automationforum.co

Single seated
Two general areas: the
actuator and the body
The actuator is the part of the
valve that converts the energy
input to the valve mechanical
motion to increase or decrease
the flow restriction.
Body: single-seated and
double-seated valves, split
body valve
www.automationforum.co

Double seated
Double-seated valves can handle high
process pressure with a standard actuator.
However, when tight shut-off is required,
single-seated valves are generally used.
Double-seated valves tend to have greater
leakage when closed than single-seated
valves.
www.automationforum.co

SPLIT BODY
This type of body is
frequently used in process
lines where frequent changes
of plug and seat are required
because of corrosion.
www.automationforum.co

Three way
Can either be diverging or
converging
Consequently, they can either split
one stream into two other streams
or blend two streams into only one.
They are sometimes used for
control purposes.
www.automationforum.co

CAGE
have hollow plugs with
internal passages
www.automationforum.co

Other types of linear valves


Y-style valve - used in molten metal or cryogenic services
Pinch valves or diaphragm valves – commonly used for
highly corrosive fluids, slurries and high-viscosity liquids, as
well as in food processing operations (beer and wine)
Gate valve – used mainly as a block valve, for fully open or
fully closed services. These are not used as automatic
valves in throttling services.
www.automationforum.co

TYPES OF CONTROL VALVES -


Rotary
Butterfly valve – consist of a disk rotating about a shaft. This type of
valve requires minimum space for installation and provide high-
capacity flow at low cost.

Ball valve – commonly used to handle slurries or fibrous materials.


They have low leakage tendency, are small in size and can also
provide high-capacity flow at low cost.

** There are a great number of valves available to meet requirements


for specialized services as well as safety and other types of regulation.
www.automationforum.co

BUTTERFLY
consist of a disk rotating about a shaft.
This type of valve requires minimum
space for installation and provide high-
capacity flow at low cost.
www.automationforum.co

BALL
commonly used to handle slurries or
fibrous materials. They have low
leakage tendency, are small in size
and can also provide high-capacity
flow at low cost.
www.automationforum.co

Types of actuators
Pneumatically Operated Diaphragm Actuators – these are the most
common actuators in the process industries. These actuators consist
of a flexible diaphragm placed between two casings.
One of the chambers resulting from this arrangement must be made
pressure-tight.
The force generated within the actuator is opposed by a “range”
spring.
The controller air pressure produces a force that is used to overcome
the force of the actuator’s range spring and the forces within the valve
body.
The action of the valve (FC or FO, is determined by the actuator.
Some valves can also have the action set at the body so that the stem
always moves down. In these, cases the valve is either FC or FO when
the stem moves down.
www.automationforum.co

Types of actuators
The size of the actuator depends on the process pressure against w/c
it must move the stem and on the air pressure available.
The most common air pressure range is 3 to 15 psig but ranges of 6 –
30 psig and 3 – 27 psig are sometimes also used.
These diaphragm actuators are simple in construction and also
dependable and economical.
www.automationforum.co
www.automationforum.co

Types of actuators
Piston Actuators
Piston actuators are normally used when
maximum thrust output is required along
with fast response.
This usually occurs when working against
high process pressure.
These actuators operate using high air
pressure supply, up to 150 psig.
The best designs are double-acting to give
maximum thrust in both directions.
www.automationforum.co

Types of actuators
Electrohydraulic and Electromechanical Actuators
These are becoming more common with the use of electrical control
signals.
They require electric power to the motor and an electric signal from
the controller.
The most common in this family of actuator is probably the solenoid
actuator.
Solenoid valve can be used to actuate a double-acting piston
actuator.
By making or breaking an electric current signal, the solenoid
switches the output of a connected hydraulic pump to either above
or below the actuator piston.
Accurate control of valve position can be obtained with this unit.
www.automationforum.co

Manual-Handwheel Actuators:
Used where automatic control is not
required. They are available for
reciprocating stem and rotary stem
valves.
www.automationforum.co

Control valve accessories


Positioner – a device that acts very much like a proportional controller
w/ very high gain. Its job is to compare the signal from the controller
with the valve stem position.
If the stem is not where the controller wants it be positioned, then the
positioner adds or exhaust air from the valve until the correct valve
position is obtained. i.e., when it is important to position the valve’s
stem accurately, a positioner is normally used.
Use of positioner tends to minimize the effects of :
Lag in large-capacity actuators
Stem friction due to tight stuffing boxes.
Friction due to viscous or gummy fluids
Changes in process line pressure
Hysteresis
Some control loops for w/c positioners are common are temperature,
liquid level, concentration, and gas flow loops.
Control valve www.automationforum.co
accessories
Valve with a positioner
www.automationforum.co

Control valve accessories


Boosters (also called air relays) – are used on valve
actuators to speed up the response of the valve to
changing signal from a low-output capacity pneumatic
controller or transducer.
The use of boosters may be the proper choice for fast-responding
control loops where the use of positioners are discouraged.
Possible uses
Amplify a pneumatic signal. Some amplification ratios are 1:2 and 1:3.
Reduce a pneumatic signal. Typical ratios are 5:1, 3:1, and 2:1.
Limit Switches – are mounted on the side of the valves
and are triggered by the position of the stem. These
switches are generally used to drive alarms, solenoid
valves, lights or other such devices.
www.automationforum.co

Control valve dynamics


Control valve dynamics tend to be relatively fast compared to the
dynamics of the process itself. However, the overall behavior of
pneumatic control include non linear behavior (dead band, hysteresis,
backlash, stick-slip phenomena). Fortunately, their effects can be
reduced by employing valve positioners.
www.automationforum.co

For purposes of control system analysis using transfer functions, the


dynamic behavior of the control valve (and the positioner) can be
approximated by a first-order transfer function Gv(s) between the
manipulated variable u(t) and the signal to the control valve p(t),

τv << τp and τp is the largest process time constant

U (s) Kv
= Gv( s ) =
P(s)  vs +1
www.automationforum.co

ACCURACY IN
INSTRUMENTATION
accuracy and inaccuracy of control is important
accuracy requirements are inherently related to control system
objectives:
ex: a)cooling water flow errors on the order of 10% (of the
measured flow rate might be acceptable in a control loop
regulating the temperature of a liquid leaving a condenser, as long
as the measurements are simply biased from the true value by this
constant amount
b) On the other hand, errors in the feed flow rate to a process on the
order of 1-2% might be unacceptable if the inventory calculations
must be made with these data
www.automationforum.co

Terms used TO DESCRIBE


INSTRUMENTATION ACCURACY
accuracy designations for control instruments are often
misused
we say that a transmitter has ±X% accuracy when we
should say that it is ±X% inaccurate
confusion also exists among the terms precision,
resolution, accuracy, and repeatability
to make these definitions clear, error should be defined
www.automationforum.co

Error is the difference between a perfect measurement and the


measurement that actually is made
Because an instrument is designed to operate over a particular
input range, error is often expressed as
percentage of full scale (%FS); or,
referred to the input (RTI), that is, a fraction of the input value
www.automationforum.co

The term full scale refers to the span of the instrument. Suppose
that the %FS error of a temperature transmitter is 1% and the zero
and span are adjusted so that the instrument operates over the
range 10-70 C.
Since the span is 70-10 = 60 C, the measurement error is 1% of 60
C, or 0.6 C. Consequently, the relative error (obtained by dividing
the error by the value of the measurement) at 10 C is 0.6/10 = 6%
Thus, when an instrument accuracy is expressed as %FS, the
relative error can be quite large for small values of the measured
variable
www.automationforum.co

Resolution - refers to the smallest interval between two


numerical values that can be distinguished.
For example, if a temperature transmitter has a resolution
of 0.1 C, it is not possible for it to distinguish between
actual temperatures of 21.62 C and 21.67 C.
www.automationforum.co

Precision – refers to the variability of a measurement for specified


conditions and a particular instrument. It is usually expressed as
standard deviation or range.

For example, suppose that a composition sample was carefully


prepared and divided into four parts, and that the composition of
each part was measured using an analyser. If the four
measurements of a key component are 21.3, 22.7, 20.6, and 21.5%,
the analyser precision could be expressed as the range, 22.7-
20.6=2.1%, or as the standard deviation, 0.87%.
www.automationforum.co

Repeatability – is similar to precision but refers to the


variability of replicate measurements in a set of data. The
variability of data is due to random errors from both
known and unknown sources. The variability can be
expressed as a range or standard deviation.
www.automationforum.co

Bias- refers to a constant error in data due to a


deterministic cause rather than random variations. A
thermocouple measurement in a vessel could be
consistently lower than the actual fluid temperature
because of conduction heat losses, in which the
thermocouple is in contact with the vessel wall.
www.automationforum.co

EXAMPLE
Consider a hypothetical experiment with some device
measuring the flow rate of a liquid
Under conditions that should yield a constant flowrate, we
make a number of measurements and record them,
assuming the instrument can be read to the nearest 0.01
flow unit
The following figure might result
www.automationforum.co
www.automationforum.co

Referring to the figure:


Precision is limited to ±0.01 flow units because we cannot
read the instrument any more precisely
Precision is related to resolution, which is defined as the
smallest change in input that will result in a significant
change in transducer output
www.automationforum.co

the systematic error or bias gives an average (probable) measured


value that is 0.37-0.25=0.12 flow units (too high at these
conditions)
the maximum error can be as large as 0.14 units, hence, the
accuracy (inaccuracy) is no worse than 0.14 at these conditions
the repeatability of the measured value is ±0.02 flow units at the
constant conditions of this experiment
Note that the precision (resolution) of a transducer can be good
while its accuracy is poor; conversely the accuracy cannot be good
if the precision is poor
www.automationforum.co

Types/sources of instrument
error
1)Nonlinearity - was a significant source of error prior to
the availability of digital instrumentation
2)Hysteresis – characterized by an output result that
depends on the direction of change of the instrument
input, usually from nonideal magnetic or electrical
components
3)Backlash – the mechanical equivalent of hysteresis,
usually from friction effects or gears with play
www.automationforum.co

Deadband – occurs when there is no output change (up to


some threshold value) for a change in input
Drift – characterized by a slowly changing instrument
output when the input is constant; it often results from
faulty or temperature-sensitive electrical components,
particularly operational amplifiers
www.automationforum.co
www.automationforum.co

Calibration of instruments
Any measurement instrument from which a high degree
of accuracy is expected should be calibrated both initially
(before commissioning) and periodically (as it remains in
service)
In recent years, the use of so-called smart sensors has
become more widespread. These devices incorporate a
microcomputer as part of the sensor/transmitter, which
greatly reduce the need for in-service calibration and
checkout
www.automationforum.co

Key features of Smart Sensors


Checks on the internal electronics, such as verifying that the
voltage levels of internal power supplies are within specifications
Checks on environmental conditions with instruments
Compensation of the measured value for conditions such as
temperature and pressure within the instrument
Linearizing the output of the transmitter can be done within the
instrument instead of within the control system
Configuring the transmitter from a remote location, such as
changing the span of the output
Automatic recalibration of the transmitter
www.automationforum.co

-end-
Prepared by Engr. Ronald Ringor, ChE
www.automationforum.co

Sensors - pressure
MECHANICAL
Manometers – (differential)
Pitot -
Bourdon Gauge
ELECTRONIC
Strain Gauge
Capacitance
Potentiometers
Piezoelectric elements
www.automationforum.co

SENSORS-TEMPERATURE
Thermometers
Thermocouple
Thermistors
Pyrometers
Bimetal thermometers
www.automationforum.co

Sensors-level
DIRECT /MECHNICAL MEASUREMENT
Floats
Rotary (paddles)
Vibrating
Dipsticks
ELECTRONIC MEASUREMENT
Ultrasonic (propagation of sound waves)
Capacitance (change in capacitance)
Optical (refraction of light)
Radio Frequency (change in frequency)
www.automationforum.co

Sensors-flow
DIFFERENTIAL PRESSURE (obstruction in flow)
Variable Area
Orifices
Flow Nozzles
Venturi tubes & Flow Tubes
Pitot tubes
Target meters
V-cones
Elbow meters
www.automationforum.co

Sensors-flow
OPEN CHANNEL
Weirs
Parshall flumes
VELOCITY
Turbines & Paddlewheels
Ultrasonic
Electromagnetic
Vortex shedding
www.automationforum.co

Sensors-flow
MASS
Coriolis
Thermal
POSITIVE DISPLACEMENT
Oval Gear & Impeller
Helical Rotors
Rotary Vanes
Reciprocating Pistons
Nutating Disks
www.automationforum.co

composition
ANALYZERS (GAS)
In situ
Extractive
CHROMATOGRAPHY (separation)
GC
HPLC
TLC
SPECTROMETERS (light wavelengths)
Atomic – absorption, emission, fluoresence
Molecular – Spectrophotometers, mass spectrometers, Infrared
spectrometers etc.
www.automationforum.co

Other sensors
TURBIDIMETERS (relative clarity of fluid)
COLORIMETERS (absorbance & transmittance of light)
CONDUCTIVITY METERS
HUMIDITY MEASUREMENTS
pH MEASUREMENT

You might also like