A FEASIBILITY ANALYSIS OF THE PEANUT GRADING SYSTEM AND POTENTIAL IMPACT OF ALTERNATIVE SYSTEMS ON THE U.S.

PEANUT INDUSTRY

The U.S. peanut industry is undergoing many changes. Government programs to manage supply have been discontinued, and global competition is on the rise. Increasing bottlenecks at peanut buying points which are key players in moving peanuts from the farm to peanut shellers are due to several developments including an increase in peanut acreage, especially in Georgia and the Southeast; a shortening of the peanut harvest period due to improved technology; and disease management practices which have shortened the time window for planting, harvesting and delivering peanuts to buying points. Compared to other crops, peanut grading is costly and has remained virtually unchanged while the value of peanuts is approximately half of what it was under the old marketing quota program. The peanut grading system dates from the 1960s and it is archaic, expensive, and time-consuming, and it fails to provide an adequate measure of quality at each stage--from grower to buying point to sheller, through to manufacturer. The goal of this project is to develop an accurate, effective and efficient grading system for U.S. farmer stock peanuts.

Final Report

Contact: Dr. Stanley M. Fletcher Professor and Director National Center for Peanut Competitiveness The University of Georgia 770-228-7231 x127 smf@uga.edu

2 A Feasibility Analysis of the Peanut Grading System and Potential Impact of Alternative Systems on the U.S. Peanut Industry (USDA-Federal State Marketing Improvement Program) FINAL REPORT December 12, 2008 BACKGROUND AND JUSTIFICATION Peanut industry is undergoing many changes due to the passage of the 2002 Farm Security and Rural Investment (FSRI) Act. The 2002 FSRI Act symbolizes that the U.S. peanut industry has entered an era characterized as no supply management and increasing global competition. Overlaying this change in the U.S. peanut industry, there has been an increasing bottleneck at the peanut buying point in trying to move peanuts from the farm to the peanut shellers. This bottleneck has been created by several factors. One factor is the increase in peanut acreage, especially in Georgia and the Southeast, due to the change in the peanut program. A second key factor is the shortening of the peanut harvesting period due to improved peanut harvesting technology such as the switch from two row peanut combines to six row peanut combines. Furthermore, current disease management has shortened the time window for planting which leads to a shortening time window for harvesting and delivering the peanuts to the buying points. For example, TSWV (tomato spotted wilt virus) has significantly shortened the planting window for peanuts in the Southeast and is starting to impact the other peanut producing regions in the U.S. To properly manage this disease, the planting window has narrowed significantly. Peanut farmers in the Southeast, prior to TSWV, were able to plant peanuts during April and May. Now, the planting window has been narrowed to the last 3 weeks of May. Consequently, this has compacted the harvesting system time. While the current system may have been satisfactory with the longer harvesting system, it is definitely not satisfactory with the compacted time period. The current grading and handling procedures and equipment date back to the 1960s when it was established for the old marketing quota peanut program which included a supply management component. Relative to other crops, grading is costly for peanuts. It cost approximately $5.25 per FSP ton to grade a sample of peanuts which is approximately 1.5% of the value of the ton of peanuts to the farmer. This grading cost as a percentage of value is significantly higher relative to cotton, corn, wheat and soybeans which range for these latter crops from .3% to .7%. Grading cost has remained virtually unchanged while the value of peanuts is approximately one half of what it was under the old marketing quota program. The grading system is archaic, expensive, time-consuming, and fails to provide the adequate measure of quality needed from the grower to buying point to sheller to manufacturer. Figure 1 provides a flow chart of the current process followed in the grading of Runner and Spanish type peanuts. The process for Virginia and Valencia type peanuts is very similar. While it is difficult to read the flow diagram in Figure 1, the intent is to convey the complexity of the grading process. Thus, one can readily understand why this system needs to be enhanced if the U.S. peanut industry is to maintain competitiveness.

To further elaborate, at buying points peanut trailers are sampled by probing with a pneumatic probe. Probing alone may take 15-20 minutes depending upon the size of the trailer. This current probe method does damage to the peanut, create LSKs (loose shell kernels) and damaged kernels. In addition, it will miss or under represent some types of foreign material. While the current grading system does adjust the grade to take this damage into account, yet it uses an average value. Other systems may be available that would provide a better sampling method plus reduce the time to take the sample. A 1500 to 2000 gram sample is then collected from this probe. Foreign materials (FM) and loose shelled kernels (LSK) are removed and weighed. From the remaining peanuts, a predetermined sub-sample is pre-sized, shelled and kernels are sized through slotted screens. Those damage-free kernels that ride the slotted screen are referred to as sound mature kernels (SMK) and these along with damage-free split kernels (SS) comprise the basis of the farmers stock grade and value of the lot. Growers are penalized for excessive amounts of damage, splits, and foreign materials. Damage, split, and LSK are further examined visually for the presence of a. flavus spores. After shelling, kernels are sized and sold according the manufacturers specification, typically as Jumbo, Medium, or No. 1 kernels that meet both a size and count per pound specification. Technology is available to develop a superior grading system than currently in use. Quality analyses are available today such as near infrared measurements (NIR) for sugar, oil, and protein. Machine vision, ultrasound, and magnetic resonance imaging are used in grading other commodities and may have applications for peanuts. Research has been conducted in automating the pneumatic probe used in sampling, as well as mechanizing the shelling and sizing components in the grading process, but to date none have been adopted. Another issue with the current grading system deals with labor. Adequate labor supply and quality of labor has become a critical issue. Historically, farmer wives served as a major labor pool for this seasonal work. Due to the economic situation on the farms, most wives now work full time in the town or surrounding areas. Thus, this qualified labor pool has shrunk significantly. To further illustrate the significant shrinking qualified labor pool, in a recent peanut harvest season one state had to use prison inmates for peanut grading. The grade of a peanut lot is highly dependent on the quality of the labor. Thus, alternative grading systems need to be examined that would reduce the need for labor. For U.S. peanut producers to maintain and enhance their competitiveness, cost factors of getting the peanuts from the farm to the consumer must be examined. Grading is one key component of this cost. Obtaining a more time efficient and cost effective grading system will enhance the competitiveness of U.S. peanuts relative to our competitors. In view of these issues, the Georgia Department of Agriculture, Georgia Agricultural Commodity Commission for Peanuts, the American Peanut Shellers Association, the National Peanut Buying Points Association, the Georgia Federal-State Shipping Point Inspection Service, Inc. (a nonprofit corporation), and The University of Georgia College of Agriculture and Environmental Science had requested that a comprehensive study be made of the current peanut grading and inspection system. Thus, this study helped the industry to identify the key factors and potential solutions.

Figure 1 - Current Incoming Farmers Stock Grading Procedures for Runners and Spanish Type Peanuts

5 GOAL(S) and OBJECTIVES The grading of farmer stock peanuts should be conducted in a manner that determines an accurate grade. The procedure should be both time efficient and cost efficient. Specific objectives for this research were: 1. 2. 3. Complete a literature search on both published and nonpublished research conducted in prior years on the grading system. Evaluate the current grading system and potential modifications. Analyze and evaluate the automated grader developed by United States Department of Agriculture-Agriculture Research Service-National Peanut Research Laboratory in the 1990s. Analyze and evaluate other alternative grading systems and emerging grading technologies. and Evaluate the potential impacts of alternative systems on the U.S. peanut industry. WORK PLAN The appropriate procedures were followed to adequately address each objective. The appropriate scientists like agricultural engineers, agricultural economists, food scientists and agencies (e.g., United States Department of Agriculture-Agriculture Research Service-National Peanut Research Laboratory, United States Department of Agriculture-Agriculture Marketing Service-Federal State Inspection Service) were involved with this project. Joint meetings were held among the scientists and the appropriate peanut industry personnel. The Georgia Department of Agriculture was the lead agency in this endeavor. Dr. Stanley Fletcher, Director of the National Center for Peanut Competitiveness with the University of Georgia, was the primary collaborator with the Georgia Department of Agriculture on this study. Dr. Fletcher was the project coordinator. The plan of work addressed both the short and long term goals for this project. Among the short term goals, one of the first items was the completion of the literature search on both published and nonpublished research conducted in prior years on the peanut grading system, as well as on the methodology/systems used in grading other commodities. This search included time and cost information collected by various organizations. Methodology and equipment for grading other commodities may be adaptable to peanuts to increase the precision, accuracy, and efficiency of the grading system. A core group composed of a representative from the growers, buying points, shellers, and Federal-State Inspection Service (FSIS) were organized to guide and provide on-going evaluations for this project. In the mid-1990s, the United States Department of Agriculture-Agriculture Research ServiceNational Peanut Research Laboratory developed an automated machine for grading that incorporated components of the Federal State Inspection Service foreign material machine,

4. 5.

6 presizer, flat-bed sheller, and vibratory screens. Because of the machines complexity and cost, it was not adopted by the industry at that time. This project re-evaluated this machine in view of the current program, as well as investigated small rotary shellers that may provide better correlation of split kernels to that of commercial shellers. Further the potential for mechanization of each individual machine that FSIS uses in the grading procedure was evaluated. Each component of the existing grading system was evaluated for its current need. For example, split kernels were critical because of the price differential under the marketing quota program. While less desirable than whole kernels, split kernels are similar in price to medium runner kernels, and therefore not as critical in todays grading criteria but is a function of the market. For the longer term, the project started analyzing and evaluating other alternative grading systems and emerging grading technologies for their applicability to peanuts. Finally, a centralized grading concept was investigated similar to what is done in the cotton grading. OUTCOMES and PROJECT EVALUATION The primary outcome of this project is the development of a system that accurately determines the grade of a ton of farmer stock peanuts (FSP) in a less expensive and time efficient manner than is currently being utilized. This grading system may be totally new or may be a refined and enhanced current system. The outcomes and project evaluation are viewed specific to each of the short term and long term goals and objectives. Prior to the start of this grant, several industry meetings were held with the primary purpose being the grading issue. Also, meetings with the state supervisors for peanut inspections to explain the project were held. In order to help this project meet its objectives and maintain communications with the industry, an industry committee was formed with representation from each of the segments of the peanut industry. There are representatives from the grower sector, buying points sector, shellers, and Federal State Inspection Service. This group met several times during the time of this project and has provided valuable insight and direction to the research. Presentations and discussions have also occurred at the U.S. State Peanut Supervisors annual meetings in 2007 and 2008, at 2007 and 2008 Peanut Industry meetings and at the 2007 USA Peanut Congress convention. In depth discussions occurred at each of these meetings as to the results and what was going to be the next step in the project. The industry has been very supportive of this project and recognizes the complexity in solving the grading issues. Objectives: Literature Search: The literature search on prior years grading research summarized available technologies that could be either considered for implementation or re-evaluated under the current peanut policy and market structure. Several of the previous research activities demonstrated the potential for improvements in the grading process. However, due to various reasons such as time constraints, costs/benefits, etcetera were not commercially adopted and implemented. It provides a background of the current peanut

7 grading system and issues associated with it. The literature search is provided in Appendix A. Given that change is critical, the latter part of the literature review focused on a new book published by IFT in summer 2007. The title of the book is Nondestructive Testing of Food Quality by Joseph Irudayara and Christoph Reh. (Irudayaraj, J. and Reh, C. (2008). Nondestructive testing of food quality, Wiley-Blackwell, Ames, IA). Within this book, ten various methods for food quality/safety were investigated either for future use or continued use within the food industry. The main objective of this review was accomplished, in that readers were exposed to available technologies; and that someway it will have an impact on the use and further development of nondestructive testing methods. Of the ten methods discussed, ultrasound, online imaging, and NMR seem to be the most readily applicable to the peanut grading process. These techniques could possibly aid in RMD detection, kernel sizing, and detection and extraction of foreign material. However, the benefit of these nondestructive methods greatly exceeds peanut grading. The implementation of these methods could revolutionize food quality and food safety as we know it. Evaluate Current Grading System and Modifications: The cost and time to grade a peanut sample has remained relatively the same for many years. However, the value of farmer stock peanuts has been reduced and the window for harvest has been drastically narrowed due to changes in planting times due to Tomato Spotted Wilt Virus and the investment in larger, higher capacity harvesting equipment. The demand to reduce cost and time has greatly increased. Prior to the 2007 peanut harvesting season, a better understanding of the peanut buying point structure was needed. Based on the 2007 peanut season, data was collected as to the number of buying points by state and the tonnage handled. This data is reflected in Figure 1 for the distribution of buying points by state and Table 1 for the distribution of buying points by size. In 2007 there were 356 buying points. Two-thirds of these buying points handled less than 5,000 tons for farmer stock peanuts. Yet, those same buying points in the aggregate handled less than 29% of the 2007 total tonnage. In contrast, the recently built buying points typically handle at least 15,000 tons or more. Based on the 2007 data, buying points handling more that 15,000 tons of farmer stock peanuts in their operation represented less than 6% of the total number of buying points. But, these larger buying points handled more than 25% of the total tonnage. This snapshot of the buying point sector of the peanut industry points out the challenges in determining an alternative system that accurately determines the grade of a ton of farmer stock peanuts (FSP) in a less expensive and time efficient manner than is currently being utilized. Some of the alternatives mentioned in the literature review will be cost prohibitive for the smaller buying points.

Figure 1. 2007 U. S. Buying Point Distribution by State

MS, 0.28% VA, 7.58% OK, 3.09% NM, 1.40% SC, 2.25%

FL, 6.18% GA, 41.29%

NC, 14.33%

AL, 13.20% TX, 10.39%

9 Table 1. Distribution of Buying Points by Size, 2007 Crop Year, United States Size of Buying Points (tons) LT 1,000 LT 3,000 LT 5,000 LT 7,000 LT 10,000 LT 12,000 LT 15,000 LT 20,000 GE 20,000 <1,000 1,000<=x<3,000 3,000<=x<5,000 5,000<=x<7,000 7,000<=x<10,000 10,000<=x<12,000 12,000<=x<15,000 15,000<=x<20,000 20,000<=x % of Buying Points 17.13 46.35 66.29 76.97 87.08 92.42 94.38 97.75 2.25 17.13 29.22 19.94 10.68 10.11 5.34 1.96 3.37 2.25 % of Tonnage Handled 1.96 13.26 28.92 41.38 57.77 69.05 74.32 86.32 13.68 1.96 11.30 15.66 12.46 16.39 11.28 5.27 12.00 13.68

Given the size distribution of the buying points and the potential cost of alternative grading systems, a centralized grading concept was examined. This concept is currently being utilized in the cotton industry. The idea is that several buying points would send their peanut samples to a central location for grading. Instead of being graded within a few hours after the sample is drawn, the sample would be sent to a central site and graded within a couple of days. This concept will require additional changes in how the peanuts are handled at a peanut buying point. Currently, the semi-trailers and wagons need to be turn around in a timely fashion so that peanut farmers can use them again to bring their peanut crop in. This implies that the peanuts will need to be stored temporarily after the sample is pulled. Nosales and regrades would become almost nonexistent which peanut farmers use occasionally. With these conditions, the peanuts in the wagon need to be evaluated for moisture, aflatoxin, foreign material and freeze damage before a sample is pulled. While these points are challenging in of themselves, the centralized grading concept can still be tested for potential use in the future. The concept was initially tested during the 2006 peanut crop year in cooperation of the testing of the United States Department of Agriculture-Agriculture Research ServiceNational Peanut Research Laboratorys proposed grading equipment (which will be reported in the next section). The results are provided in Appendix B Tables 1-5. The samples from the three peanut growing regions were sent to Georgia. The samples were divided into two equal subsamples with one subsample being used to test the United States

10 Department of Agriculture-Agriculture Research Service-National Peanut Research Laboratorys proposed grading equipment and the other subsample was run through the current grading procedure. Georgia Federal State Inspection Service personnel oversaw the test utilizing trained peanut graders. When the samples were shipped to Georgia, the samples official FV-95 form was included. The FV-95 form is the official grade form for that load of peanuts represented by the sample. When comparing the mean values of the FV-95s from the buying point and the FV-95s from the central grading facility, there was statistical difference in the means. However, when one looks at the numerical values, the values are not really different. The major differences in values were found in peanut samples other than the runner type peanuts. This could be explained in that peanut graders from Georgia do not really come into contact and grade peanuts of the other type of peanuts. Another concern for centralized grading dealt with the moisture level of the sample after being shipped to a central site. It was felt that the moisture level would drop significantly which would influence the grade factors. Interestingly, the moisture level did drop some but not as much as expected. This may have been due to using heavy sealed plastic buckets to ship the samples. The conclusion from this first years results was that a potential existed with a centralized grading model. In 2007, the centralized grading concept was tested on a larger scale. In addressing the potential issue from the 2006 test, a regional centralized grading test was conducted. Since there are three distinct peanut growing regions, a centralized grading test was conducted for each region. Each peanut producing region except for the Southwest region produces a particular type of peanut. The Virginia-Carolina region grows mainly Virginia type peanuts. The Southeast region grows mainly runner type peanuts. The Southwest region grows those two types of peanuts plus Spanish and Valencia type peanuts. Extensive communication with the each of the states peanut grading supervisors was conducted. Every peanut states peanut grading supervisor was supportive of the test. The supervisor from Texas was in charge of the Southwest peanut growing region composed of Texas, Oklahoma and New Mexico. The North Carolina supervisor was in charge of the Virginia, North Carolina and South Carolina peanut growing region (Virginia-Carolina region). The Georgia supervisor was in charge of the Alabama, Florida and Georgia peanut growing region (Southeast region). After several communications with a statistician, a sample size of 500 peanut samples for each region was determine to be appropriate. Based on a states peanut production share in the region, the sample size out of 500 was distributed to each states peanut grading supervisor. Each supervisor collected the samples such that it was distributed across the states producing area and across the harvest season. The results for each growing region are provided in Appendix B Tables 6-8. The FV-95s were collected with the sample at the buying point and compared to the FV-95s from the respective regional centralized grading facility. As was found with the 2006 test, there was statistical difference in most of the means. However, when one looks at the numerical values, the values are not really different. For the second year, the moisture level did not change much numerically. In discussion with various state peanut grading supervisors, the observed differences in grade items could be partially explained by the equipment used in the grading as well as the human factor. One additional item was examined and reported in the tables. Peanut farmers want to know the value of their load of peanuts as represented by

11 that sample graded. The mean grade items were entered into the loan schedule provided by United States Department of Agriculture/Farm Service Agency to determine the average grade value of all of the samples graded. One will note that the grade value was higher by $5-$6 per ton FSP for the centralized grading relative to the buying point for the VirginiaCarolina region and the Southwest region. In contrast, for the Southeast region the buying point mean grade value was about $2.00 per ton FSP higher than the centralized grading. While the two years of test indicate the potential feasibility of centralized grading, the cost savings have not been projected yet. Additionally, further tests were not planned. There has been some reluctance in the centralized grading concept. Some industry personnel interpret it to mean that smaller buying points will be closed which is not the case. The limiting factor in this concept are the issues of determining moisture, aflatoxin, foreign material and freeze damage before a sample is pulled and the load of peanuts are assimilated in a temporary storage facility. Research is critically needed for these issues. When looking at the grading data for a state, a key observation of the data was the significant amount of no-sales for peanut load. There can be up to 10% of the loads being no-sale. Many times, the no-sale may occur after 90+% of the grading is done. The charge for a no-sale is minimal. One of the major reasons for the no-sale is the sample fails the moisture test. In discussions with the industry grading coalition committee, they felt that if an accurate moisture test could be done prior to drawing a sample, this would significantly improve the current grading process. This would also address one of the key limiting factors for centralized grading. Thus, a project was started to address this issue. The Principal Investigator of this project was Dr. Stuart Nelson (United States Department of Agriculture-Agriculture Research Service, Athens, GA). In the latter part of 2007, Dr. Nelson retired and Dr. Samir Trabelsi assumed the leadership role. Even though Dr. Nelson retired he still worked on the project to ensure its success. They have an experimental prototype that is being tested in the lab with the charge to have field testing done on the 2008 peanut crop. The industry has been very supportive of this work. Five major shellers have shipped 50 pounds of shelled peanuts and 50 pounds of Farmer Stock peanuts to Dr. Trabelsi to aid them in their research. The shellers are providing these peanuts as their inkind contribution to the project. The samples have been Virginia peanuts, Hi O/L runner peanut, regular runner peanut, Valencia peanuts and Spanish peanuts. Thus, this moisture machine research was conducted on all types of peanuts. In addition, each lot of peanuts was provided by a different sheller, indicating the strong support across the peanut shelling sector. Based on the shipments received, there is a very high correlation in measuring moisture and is independent of the type of peanut. This latter finding is very important given that the current moisture testing machines require one to input the type of peanut plus it must be shelled. This new prototype will be able to measure the moisture of the peanut as a farmer stock peanut as it is being brought to the buying point by the peanut farmer. Dr. Trabelsi is preparing a proto-type machine to be tested in the field during the 2008 peanut harvest. A more detail report by Dr. Trabelsi on the moisture machine is provided in Appendix C. Automated Grader by United States Department of Agriculture-Agriculture Research Service-National Peanut Research Laboratory: This objective was addressed during the

12 2006 peanut harvest. United States Department of Agriculture-Agriculture Research Service-National Peanut Research Laboratory developed an automated peanut grading device which demonstrated that peanut grading could be done with less labor requirements while still yielding the same grade factors and dollar value. However, the cost of investing in an automated grader and the fact that temporary labor availability was not a significant problem at that time made the automated grader not feasible. Under todays environment the automated peanut grader was reevaluated in side by side comparisons with existing grading procedures. While the initial plans was to have this machine built for testing in each of the three peanut regions, United States Department of Agriculture-Agriculture Research Service-National Peanut Research Laboratory did not feel comfortable with this approach and thus only one machine was built. The test was conducted at the United States Department of Agriculture-Agriculture Research Service-National Peanut Research Laboratory facility. The machine actually was a new foreign material separator and a new shelling machine. The experiment was set up such that there were two peanut grading lines. One line used the new/modified equipment and the second line was a duplicate of the current grading system. Virginia, North Carolina, Georgia, Florida, Alabama, Texas, Oklahoma and New Mexico provided samples for this test. Approximately 420 peanut samples were collected and run through the test. Each state provided a sample for the new equipment grading line and a sample for the current grading system. These samples came from the same peanut load. When the samples were sent to the central test site, the associated FV-95 form was included. This allowed us to know the original sample grade which will allow us to compare it against the grades done at the central site. Besides the grades being collected for each sample, the time it took to do the grading was collected. A large database was formed matching the samples up to their respective peanut load. This allowed comparison across methods of grading. Simple statistics like means were calculated and reported in Appendix D Tables 1-8. The simple statistics indicate that the new equipment did not perform significantly better. While mean times reported in Appendix D Table 1 indicate that the new equipment was faster than the current grading equipment overall, that was only true for the runner type peanuts. Why those results for the runner peanut occurred are not clear given the results presented in Appendix D Tables 2 and 3. These latter results indicate that the new foreign material separator and sheller were either slower or no different in times with the current grading equipment. The results in Appendix D Tables 4-8 indicate that there were statistical differences in the means of the grading factors. Upon closer examination of the tables, one will note that the total sound mature kernels plus sound splits (SMKSS) is significantly lower for the automated grader relative to the current grading method. The lower total SMKSS translates into lower grade dollar value. In addition, sound splits for the automated grader was statistically higher than with the current grading equipment. This also translates into a lower grade dollar value. This analysis of the data clearly indicates that the automated grader in its present form did not perform significantly better than the current grading equipment. Thus, the automated grader was not further considered. This view was also held by the industry peanut grading coalition. However, individual components like the foreign material separator is being further examined where it could replace the current foreign material separator currently being used. This work is being done by the United States Department of AgricultureAgriculture Research Service-National Peanut Research Laboratory.

13 Alternative grading systems: Some of the industry peanut grading coalition committee members have attended some trade shows to gain a better understanding of potential equipment. The committee realized that the peanut industry needs to decide what constitutes quality and eventually grade. Thus, each segment had discussions to determine the parameters needed to be graded. Once these are determined, we can approach some of the companies and researchers for potential equipment. The shellers reported that they wanted to keep the current grading standards but increase the grading speed and reduce cost. The producers have not decided due to the Farm Bill period which had kept them preoccupied. However, upon further investigation, it was discovered that the peanut industry has a committee addressing what factors that it wants the peanut breeders to strive for in their peanut breeding programs. Some of the seed index factors being considered for the peanut breeders are relative maturity, fat content, sugar content, protein content, shelflife indicators, roast variability, seed size, flavor, etc. Thus, these factors have been presented to the peanut industry at the various meetings as the factors that should be the primary ones used to determine the final grade. In discussions with the peanut breeders, they liked that approach. At the present time, no consensus has occurred as to the grading factors. In order to test an alternative grading system for the 2007 peanut crop harvest system, an XRay machine was determine to be a potential alternative. This machine would be testing some of the current grading factors. The company that is providing this test equipment provides similar equipment for some of the shellers shelling plant. The company also provided technicians and computer programmers to develop the software. Some very small samples had been tested from the 2006 crop season. Based on the results, it was determined to try it on a larger test scale. The United States Department of Agriculture-Agriculture Research Service-National Peanut Research Laboratory did the testing of this equipment on the 2007 crop harvest. The samples that were run through this machine were also sent through the current grading equipment so that the results can be compared. The results looked very favorable. Thus, a more comprehensive test during the 2008 crop harvest will be done involving all three peanut growing regions. A report by the United States Department of Agriculture-Agriculture Research Service-National Peanut Research Laboratory on this X-Ray machine is presented in Appendix E. While this X-Ray machine shows promise, the machine can not address two key grading elements damage and aflatoxin. Thus, alternative methods are being discussed. Evaluate potential impacts on the industry: In preparation 2007 data has been collected for each state as to the number of buying points and the respective tons handle by each of those buying points. Prior data could not be obtained. However, successive years data will be collected in order to compute averages which would eliminate swings in sizes due to changes in production in particular areas by year. This data will help in looking at the total structure and the potential changes and costs associated with the changes. Given the complexity and timing of this study, this objective was not completed. Contact person: Dr. Stanley M. Fletcher, 770-228-7231 x127, smf@uga.edu .

14 APPENDIX A

Peanut Grading Literature Review

By Micah Lewis Under the direction of Dr. Bill Tollner Department of Biological and Agricultural Engineering College of Agricultural and Environmental Science University of Georgia

15 Table of Contents 1. INTRODUCTION 1.1. Peanut History 1.2. Emergence in North America 1.3. Types of Peanuts 2. PLANTING & HARVESTING 2.1. Traditional Techniques 2.2. Technological Advances Increase Productivity 2.3. External Circumstances Compact Planting and Harvest Time 3. PEANUT GRADING 3.1. Initial Grading 3.2. Cleaning 3.3. Shelling 3.4. Blanching 4. PROBLEMS WITH CURRENT GRADING SYSTEM 4.1. Time Consuming 4.2. Costly 4.3. Crop Destructive 4.4. Inaccurate Measure of Quality 5. POTENTIAL NEW GRADING TECHNIQUES 6. BIBLIOGRAPHY Page 15

16

18

20

21 22

16

CHAPTER 1 INTRODUCTION
1.1 Peanut History Despite the nonexistence of fossil records, it is believed that the peanut plant originated in South America, possibly Brazil or Peru, approximately 3500 years ago. Recovered pottery artifacts include jars shaped like peanuts and decorated with peanuts (American Peanut Council, 2007). Graves of the ancient Inca civilization, found along the western coast of South America, often contained jars filled with peanuts. By the time the Spanish conquistadors arrived in America, the growth of the peanut crop had spread as far north as Mexico. The Spanish explorers took peanuts back to Spain; and through further trade and exploration, peanuts were spread to Africa and Asia. 1.2 Emergence in North America The peanut crop was introduced to North America as a result of the slave trade from Africa around the 1700s (University of Georgia, 2006). Peanuts were initially grown only in the southern United States and used as food for pigs. Around 1800, the commercial growth of peanuts began in South Carolina for food, oil and a substitute for cocoa (American Peanut Council, 2007). In 1860, the peanut consumption in the United States had a substantial increase due to the Civil War. Peanuts served as a vital source of food for soldiers. However, outside of the soldier consumption of peanuts, the demand remained low because harvesting was done by hand. This resulted in trashy, low quality peanuts. It wasnt until 1900 that proper equipment was invented for planting, harvesting, and grading peanuts. After this point, peanut production increased dramatically (Putnam, 2000). The peanut crop was put into heavy rotation following a recommendation given to the southern farmers by Dr. George Washington Carver. The repetitive planting of cotton crops had exhausted the nutrients needed for good yields. Dr. Carver proposed that the peanut crop be put into rotation to restore nitrogen and protein to the low quality soils. Implementing this crop rotation resulted in improved cotton yields, another food source, and an alternative cash crop (Putnam, 2000). Today, peanut production is very high in the United States. The growth of peanuts is no longer limited to the southeastern states. It is spread throughout the country, providing different varieties for different uses. 1.3 Types of Peanuts There are many different varieties of peanuts grown. However, for market, there are four basic types; Runner, Virginia, Spanish and Valencia. Varieties of peanuts can be distinguished by their growth location, branching habit, and branch length (Virginia-Carolina Peanut Promotions, 2007). Runners are known for their attractive, uniform kernel size and good flavor. They also have better roasting characteristics and higher yields than Spanish type peanuts. Nearly 54% of

17 all runners grown are used for peanut butter. They are typically grown in Georgia, Alabama, Florida, and South Carolina. Virginia type peanuts have the largest kernels and account for the majority of the supply of roasted and in-shell peanuts. They are mainly grown in southeastern Virginia, northeastern North Carolina, South Carolina and Tennessee. Spanish type peanuts are known for their high oil content. Their pods contain small kernels covered in a reddish-brown skin. They are used mostly in peanut candies. Before 1940, 90% of the peanuts grown in Georgia were Spanish type. However, now they are grown in Oklahoma and Texas. Valencia type peanuts are known for the sweetness and are usually roasted and sold inshell. Their outstanding flavor also makes them the main candidate for boiled peanuts. Pods usually contain three or more small kernels covered with a bright red skin. The Valencia type peanuts are primarily grown in eastern New Mexico.

CHAPTER 2 PLANTING & HARVESTING

2.1 Traditional Techniques Peanuts are planted traditionally after the last frost, which is usually April or May. This time is preferred because the soil temperature is usually between 65 to 70F. The seeds are planted at a depth of one to two inches and spaced apart by three to four inches. Favorable growing conditions includes a climate with 160 to 200 frost free days, warm weather, adequate moisture, and rich fertile soil (Virginia-Carolina Peanut Promotions, 2007). When the majority of the peanuts have reached maturity, harvesting begins. Harvesting occurs in two phases. In the first phase, diggers are driven up and down the rows of peanuts. The digger has a long blade that runs four to six inches underground. While passing through the rows, it loosens the peanut plant and cuts the taproot. Located directly behind the blade, the shaker lifts the plant from the soil, shakes dirt from the peanuts and lays the plant down inverted to expose the peanuts. At the time of initial exposure, peanuts range in moisture content from 25 to 50%. Therefore, they are allowed to remain in the field for two or three days to allow the moisture to drop. The ideal moisture for storage is less than 10%. The second phase of harvesting is combining. After the peanuts have had a reasonable chance to dry in the field, they are collected using combines, also known as threshers. The combine lifts the entire plant and separates the peanuts from the vine. The peanuts are then blown into a hopper located on top of the combine. The vines are laid back in the field to aid in soil fertility and increase organic matter. When harvesting is complete, the peanuts are transferred to wagons with perforated floors. Here they are cured to low moisture by forcing warm air through the floors of the wagons.

18 2.2 Technological Advances Increase Productivity Even though the consumption of the peanuts in the United States increased during the Civil War, production was still limited because peanuts were mainly planted and harvested by hand. After 1900, equipment was invented to mechanize the processes of planting and harvesting. This resulted in a boom in the production of peanuts and the demand for them. Even throughout the 20th century, and now into the 21st, the efficiency of machinery increased substantially, furthering the increase in production of peanuts at the same time. Two types of peanut planters exist, the plate planter and the air planter (Roberson, 2007). Plate planters use either a horizontal or inclined plate to meter seeds accurately. Air planters use a disk mounted vertically in the metering chamber to meter seeds. Both have been modified to handle the seeds gently while planting. However, air planters are superior in efficiency because they handle a wider variety of seed sizes and maintain metering accuracy at higher ground speeds than plate planters (Roberson, 2007). For the digging phase of harvesting, blades have been designed to optimize the yield obtained. This also requires great skill of the operator. Blade depth and tractor ground speed are key factors in this optimization of yield. Newer tractors have several gears to ensure the preferred ground speed can be obtained. While yield loss is inevitable, under good circumstances it can be kept to 5% or less. Combining has also seen mechanical advances. The advancement from two row to six row peanut combines has really diminished the time needed for combining. Also, research has been done to advise farmers on the optimal settings for components of the combine. One major component is air velocity. The velocity in the air conveyor should be just enough to lift peanuts into the bin. Likewise, the air velocity in the cleaning screens should be high enough to blow out marketable peanuts (Roberson, 2007). 2.3 External Circumstances Compact Planting and Harvesting Time As noted earlier, peanuts are usually planted when the soil temperature is between 65 70F. They also need 160 200 frost free days during maturation. Adhering to these conditions usually meant a planting window from April to May. However, with the change in climate, the normal window has been reduced to the last three weeks of May. Another contributor to the altering of the planting window is the tomato spotted wilt virus. This disease is so detrimental because it can not be controlled once a plant is infected, and early symptoms are difficult to diagnose (Kansas Statue University, 2004). Research has shown that altering the planting window was the main way of managing this disease. These limitations on the planting window directly affect the harvesting window. With planting no longer being spread over two months and with increased acreage, the time allotted for harvesting is not sufficient for getting all the peanuts grown to shellers.

19

CHAPTER 3 PEANUT GRADING

3.1 Initial Grading After the digging phase of harvesting is complete, peanuts are left in the windrows to dry. The moisture at which the peanuts are combined can have an overall effect on the grade. Researchers have determined that peanuts combined when their moisture content is between 25 and 35% have less hull damage and fewer loose shelled kernels. However, when peanuts are combined with lower moisture contents, the curing time and costs are reduced. Farmers usually develop a routine that creates a balance between the two. After the peanuts have completed curing (drying), they are inspected and graded before being shelled, cleaned or crushed. This inspection provides a determination of the overall quality and on-farm value of the peanuts for commercial sales. The grading and inspecting is facilitated by the Agricultural Marketing Service (AMS) and the FederalState Inspection Service (FSIS) at buying stations or shelling plants near the peanut fields (American Peanut Council, 2007). A representative sample is drawn from each wagon of peanuts and analyzed by a pneumatic sampler. With this analysis, the meat content, size of pods, amount of damaged kernels, amount of foreign material, and kernel moisture content are determined. These parameters are used to establish the grade, which is used to determine the commercial value. The first parameter evaluated in the representative sample is the amount of foreign material. Foreign materials include, but are not limited to, rocks, dirt, vines, stalks, sticks, insect parts and shriveled peanuts (Chapin, 2004). Shriveled peanuts are referred to as raisins because of their appearance. These are pods that are immature and incapable of being shelled by machine. Farmers are not penalized for foreign material below 5% of the sample. However, for 5% foreign material, the penalty is $1 per ton; and for 10% foreign material, the penalty is 6$ per ton (Chapin, 2004). The second parameter evaluated is the amount of LSKs (loose shelled kernels). LSKs are kernels or parts of kernels that became free from their hulls during harvesting. They have a negative effect on the overall grade of the sample because they spoil quicker and are more susceptible to aflatoxin contamination. On average, each percent of LSKs costs the farmer about $2.20 per ton (Chapin, 2004). After foreign material and LSKs have been removed from the representative sample, the peanuts are shelled for further observation. After being shelled, the peanuts are observed for SMKs (sound mature kernels). SMKs are kernels that remained intact and are undamaged. However, if they are slightly broken, they are still classified as SMKs if more than of the kernel still remains. SMKs have a large effect on the overall grade. For every percent increase in SMKs, the overall value increases about $5.00 per ton. Broken kernels are grouped into SS (sound splits) and OK (other kernels). Sound splits are undamaged, split kernels that are between and of the size of the original, whole kernel. Other kernels are pieces of split kernels that are smaller than the size of the whole kernel and small, immature kernels (Lamb, 2005). There is no main penalty for other kernels; however, they are worth less than SMKs. For sound splits totals less than 4%, there is no penalty. For each percent above 4, farmers are penalized $0.80 per ton (Chapin, 2004).

20 The next parameter evaluated is (DK) damaged kernels. These kernels are classified mechanically and by human examination. Kernels are considered to be damaged if they contain decay, mold, insect damage, skin discoloration greater than 25%, or freeze damage. Kernels with freeze damage usually have a hard, discolored, translucent flesh (Chapin, 2004). Some kernels are checked for concealed damage. Damage as such is discovered when kernels are split only to reveal a rancid, moldy or decayed interior. Farmers face strong penalties for damaged kernels. For a level of 2% damaged kernels, the penalty is $3.40 per ton. At 5% and 10%, the penalties are 25$ and $100 per ton, respectively. Next, the sample is inspected for aflatoxin contamination. This testing is mostly done chemically. Only the LSKs, OKs, and DKs are examined because research has shown that aflatoxin contamination is associated with small kernels and discolored kernels (Whitaker, 2005). Although the Food and Drug Administration (FDA) has set an acceptable limit for aflatoxin of 20 ppb in all food products, the USDA has set a stricter limit of 15 ppb in peanuts used in food for humans (FAO, 2004). When testing for aflatoxin, three 22 kg samples are taken from each lot, and the aflatoxin concentration of the three samples must average 15 ppb or less for these peanuts to be shipped to a manufacturer. If the average exceeds 15 ppb there are three options; the lot can be crushed for oil (oil processing eliminates aflatoxin), the lot can be reprocessed, or the lot can be blanched (skins removed from kernels and damaged kernels detected with electronic color sorters and removed) (Adams, 2004). The overall grade of peanuts from each lot is a function of many different parameters. However, the main influence is not just from the quality of the crop. The efficiency of the machinery and the workers used in the harvest also plays a crucial role in the final grade of the peanuts. The benefits of an excellent crop can be spoiled by poor workmanship. Each process must be handled meticulously to ensure that the farmer receives full value for the crop. 3.2 Cleaning When the grading of the representative sample is complete, the peanuts from that wagon are cleaned. During the cleaning process, foreign materials are removed from the quality peanuts. The standard peanut cleaner is a system that consists of a vibrator, air-flow sections of various velocities, and a conveyer (Monitor, 1920). As peanuts are fed to the cleaner, they first travel along a vibrating plane that loosens dirt and other small debris from the peanuts. Then they fall through a chamber with a low-velocity, lateral air flow that removes light trash from the peanuts. The peanuts fall through another chamber with a lateral air flow of higher velocity to separate the peanuts from rocks and other heavier debris. At the completion of cleaning, the peanuts contain minimal to no foreign materials.

21 3.3 Shelling After cleaning is complete, the peanuts are separated into two groups; those that will be shelled and those that will not. Annually, about 10% of the peanut crop is sold in-shell (Virginia-Carolina Peanut Promotions, 2007). These are usually Virginia and Valencia types. These peanuts are passed through a machine that cuts off any remaining stems from the shells. Then the pods are sorted and grouped by using sizing screens. The peanuts that are to be shelled pass through the shelling machine. The shells are cracked as the peanuts collide with each other while passing through a section with spinning, perforated grates. Next, the kernels and hulls pass through a low-intensity vacuum section that separates the hulls from the kernels. The kernels are sorted by sizing screens into their market grades: jumbo, medium, number one, other kernels, sound splits and oil stock (Whitaker, 2005). 3.4 Blanching The last process to occur before peanuts are ready for market use is blanching. When peanuts are initially shelled, there is a reddish skin that covers the kernels. Blanching is used to remove this skin. During this process, the shelled peanuts pass through warm air which loosens the skins (Virginia-Carolina Peanut Promotions, 2007). Then they enter a machine where large rollers rub against the kernels to remove the loosened skins. Kernels are scanned with electronic color sorters to see if any skins remain. If there are kernels that still have skins, the blanching process is repeated until all skins have been removed.

CHAPTER 4 PROBLEMS WITH CURRENT GRADING SYSTEM

4.1 Time Consuming As mentioned in Chapter 2, the window for planting peanuts has narrowed over the years from a two-month time frame to about three weeks. Also, technological advances have increased the production of peanuts. As a result, this has created more peanuts to harvest and less time to do it. In past years, the current grading system was sufficient to maintain a steady flow. However, at present, it is no longer sufficient for the larger supply and shorter harvest period. The grading process, as explained in Chapter 3, is a multifaceted procedure that consumes a lot of time. Initially, up to three days are lost as the peanuts remain in the field to dry after digging (American Peanut Council, 2007). Current standards ensure that peanuts are graded at a moisture content of 10.49% or less. Even after remaining in the field three days, additional curing is usually needed to lower the moisture content. Once the desired moisture is achieved, probing can take up to 20 minutes per trailer. There are five parameters that are used to determine the grade of peanuts from each lot. Each parameter is measured or analyzed individually. Time could be saved if some of the analyses were done in parallel.

22

4.2 Costly The cost to grade peanuts (estimated over $5 per ton) is high when compared to the grading costs of cereal grain and oil seed products (Chapin, 2004). In some cases the cost is double or greater. Unfortunately, the market value of peanuts has decreased while the grading cost has more or less remained the same. This decline in profit has made it harder for farmers to maintain efficient machinery and labor. Efficiency of machinery and labor is important because the overall grade of peanuts is highly dependent on it. In order for farmers to make a profit, they must produce high grade peanuts. 4.3 Crop Destructive One aspect that determines the overall grade of a lot of peanuts is physical quality upon arrival. LSKs result from peanuts that are inadvertently shelled during harvesting. Farmers are penalized for high percentages of LSKs in the representative sample (Chapin, 2004). The pneumatic probe has been observed to cause damage to the peanuts, creating more LSKs. This occurrence is destructive and diminishes the grade of the peanuts. At present, concealed damage is only detected by splitting the kernel to evaluate the interior. The accuracy here is not 100%; so, there are times when kernels are split only to find that the interior is not damaged. A nondestructive method for carrying out this analysis would avoid this problem. 4.4 Inaccurate Measure of Quality The concept of probing is somewhat synonymous to that of random sampling. However, with random sampling there is always the chance of misrepresenting certain parameters. These inaccurate readings can be costly to the farmer and to the buyer.

CHAPTER 5 POTENTIAL NEW GRADING TECHNIQUES

The current peanut grading system has been in place for over 50 years. At the time when it was developed, it was sufficient to handle the current peanut production. However, at present, changes need to be made to remedy its costliness, time consumption, destructiveness, and inaccuracy. Many innovations have been proposed, but none have been adopted. One proposal is to raise the allowable peanut grading moisture content from 10.49% to 18.49% (Lamb, 2005). The benefits of this increase would be a reduction in time delay from the field to the shelling station and a reduction in drying expenses. Approximately three days are lost as peanuts remain in the field to decrease in moisture content. In some cases, additional curing is needed to bring the moisture content to 10%. If it were acceptable for peanuts to be graded at a higher moisture content, newer technologies could be utilized such as continuous flow drying (Lamb, 2005). Drying could be done in parallel with other procedures. The

23 allowable peanut grading moisture content has not been raised because USDA has concern that the efficiency of the sheller will drop with peanuts at higher moisture content. Research has shown that the efficiency of the peanut sheller is not affected by an increase in moisture content. The grading process is a multifaceted procedure, as described in Chapter 3. The fact that each parameter that determines the overall grade is analyzed individually adds to the time and money spent during this procedure. Time consumption and cost could be significantly reduced if some of these parameters were evaluated simultaneously. Many existing technologies could have a beneficial application. X-ray technology is used to detect various metal and non-metallic objects in loose and packed products in the food industry. This same technology could be used in detection of foreign material within the grading sample. It could also be useful in the identification of LSKs. Infrared technology is another option for sample analysis. Besides detection of foreign materials, infrared technology would offer analysis of other parameters such as moisture content, density, hardness and color. This would eliminate human examination needed to identify damaged kernels. Similar to this technology is that of microwaves (frequencies between 2 to 18 GHz). Microwaves are also capable of analyzing multiple parameters with one scan and cause no damage to the sample. The implementation of such technologies could provide a substantial economic benefit to farmers. These technologies are nondestructive and would not damage the sample as the pneumatic probe. The time for probing would also be eliminated. Analyzing various parameters in parallel would also diminish the cost and time usually required for grading.

BIBLIOGRAPHY
Adams, J. and Whitaker, T.B. (2004). Peanuts, aflatoxin, and the U.S. origin Certification program. Meeting the Mycotoxin Menace. Wageningen Academic Publishers. The Netherlands, p. 183-196. American Peanut Council. (2007). About the peanut industry. Retrieved March 12, 2007 from http://www.peanutsusa.com/index.cfm?fuseaction=home.page&pid=12 Chapin, J.W., and Thomas, J.S. (2004). Peanut money-maker production guide 2004. Blackville, SC: Clemson Extension and South Carolina Peanut Board. Food and Agriculture Organization (FAO). (2004). Worldwide regulations for mycotoxins for food and feeds in 2003. FAO Food and Nutrition Paper 81. FAO, Vialle Della Terme di Caracalla 00100. Rome, Italy. Kansas State University. (2004). Tomato Spotted Wilt Virus. Kansas State University Research and Extension. Retrieved March 1, 2007 from http://www.oznet.ksu.edu/dp_hfrr/extensn/problems/tomswv.htm Lamb, M.C., and Blankenship, P.D. (2005). The capacity and efficiency of official grade shellers. Peanut Science. 32: 132-135.

24

Monitor. (1920). Monitor peanut cleaning, shelling and grading machinery for whole and shelled nuts. Silver Creek, NY: Huntley Manufacturing Co. Putnam, D.H., Oplinger, E.S., Teynor, T.M., Oelke, E.A., Kelling, K.A., & Doll, J.D. (2000). Peanut. Alternative field crops manual. St. Paul, MN: University of Minnesota. Roberson, G.T. (2007). Chapter 7: Planting, harvesting and curing peanuts. Peanut Information 2007. Raleigh, NC: North Carolina State University. University of Georgia. (2006). World geography of the peanut. Retrieve February 20, 2007 from http://lanra.anthro.uga.edu/peanut/ Virginia-Carolina Peanut Promotions (VCPP). (2007). Everything about peanuts. Retrieved March 22, 2007 from http://www.aboutpeanuts.com Whitaker, T.B., Dorner, J.W., Lamb, M., and Slate, A.B. (2005). The effect of sorting farmers stock peanuts by size and color on partitioning aflatoxin into various shelled peanut grade sizes. Peanut Science. 32: 103-118.

Additional Annotated Bibliography Davidson, J.I., U.J. Tsai, F.E. Dowell, J.W. Dorner and R.J. Cole. 1990. Comparison of pneumatic and automatic spout samplers to determine grade of farmers stock peanuts. Peanut Sci. 17:76-80. Tattersfield, J.R. 1969. The relationship between screen grading and kernel count in four groundnut varieties. Rhod. Agr. J. 66:6 Reprinted as Bul. 14. Anon. Better ways to test quality. 1960. Agr. Res. 8:11, pp. 8-9. Dickens, J.W. and D.D. Mason. A peanut sheller for grading samples: an application of statistics in design. 1962. Trans. ASAE. 5:42-45. Dickens, J.W. Shelling equipment for samples of peanuts. 1962. USDA-AMS Marketing. Res. Rpt. No. 528. Dickens, J.W. Development of a pneumatic sampler for peanuts. 1964. Trans. ASAE. 7:4, pp. 384-387. Moder, Jr., J.J. Analytical study of the sampling and grading of Farmers stock peanuts. Feb. 1952. Georgia Institute of Technology. Engineering Experiment Station. Special Report No. 1, Project No. 168. 44 p.

25 Moder, Jr., J.J. and N.M. Penny. Industrial Engineering and Economic Studies of peanut marketing. 1954. Georgia Institute of Technology. Engineering Experiment Station. Cooperating with Georgia Experiment Station. Bulletin No. 286. Elliott, T.A. and B.W. Carmichael. Machinery for grading farmersstock peanuts. 1955. Georgia Institute of Technology. Engineering Experiment Station. Annual Progress Report. Project No. 147. 40 p. Penny, N.M., T.A. Elliot, J.J. Moder, Jr., and B.W. Carmichael. Sampling, grading, and cleaning farmers stock peanuts. 1956. Univ. of Ga. Agr. Expt. Stn. & Eng. Expt. Stn., Ga. Inst. Tech. Bul. No. N.S. 32. Gilliland, C.B. and T.B. Smith. 1956. An analysis of the peanut shelling industry, 1950-51 through 1952-53. USDA-AMS Marketing Res. Div., Washington, D.C. Marketing Res. Rpt. No. 134. Moder, Jr., J.J. Comments on analytical aspects of sampling, grading, and quality picking peanuts. 1957. Proc. Peanut Res. Conf., Atlanta, Ga. Penny, N.M. Implications of results of research in cleaning, grading, and sampling farmers stock peanuts, problems in application, and need for additional research. 1957. Proc. Peanut Res. Conf., Atlanta, Ga. Gilliland, C.B. Marketing research by U.S. Department of Agriculture, pertaining to storage, sampling, grading, farmers stock peanuts. 1957. Proc. Peanut Res. Conf., Atlanta, Ga. Elliott, T.A. Results of the Bainbridge tests on cleaning, sampling and grading farmers stock peanuts. 1957. Proc. Peanut Res. Conf., Atlanta, Ga. Kramer, H.A. Spout type automatic sampler for farmers stock peanuts. 1959 USDA-AMS Marketing Res. Div. Marketing Res. Rpt. No. 353. Davidson, Jr., J.I. T.B. Whitaker, and J.W. Dickens. Grading, cleaning, storage, shelling, and marketing of peanuts in the United States. 1982. Ch. 15, Pp. 571-623. In: Peanut Sci. and Technology. Am. Peanut Res. Educ. Soc. Dickens, J.W. Peanut grading and quality evaluations. 1986. Proc. American Peanut Res. Edu. Soc. 18:73. Dowell, F.E. Automatic peanut grading development using machine vision. 1989. Am. Soc. of Agr. Engineers. St. Joseph, MI. Paper No. 89-3006. Ni, B, M.R. Paulsen, and J.F. Reid. Size grading of corn kernels with machine vision. 1998. Trans. ASAE. Vol. 14(5): 567-571.

26

APPENDIX B

27
2006 Crop Year Results

2006 Crop Year Sample Sizes: Runner 283; Virginia 88; Valencia 27; Spanish 20; Total - 418

Appendix B Table 1. Statistical Test for Difference in Grading Factors Between a Buying Point and a Central Grading Site, All Samples (418), 2006 Crop Year Mean Grading Factors % Foreign Materials % LSK % Moisture % SMKRS % Sound Splits % Total SMK % Other Kernels % Total Damage % Hulls % Total Kernels and Hulls % Freeze Damage % Concealed RMD Buying Point 3.92 2.99 8.24 63.81 5.68 69.62 4.99 0.52 24.64 99.78 0.01 0.03 Central Site 3.74 3.17 7.50 62.26 6.39 68.66 5.06 1.16 24.78 99.65 0.02 0.08 Statistical Difference Yes Yes Yes Yes Yes Yes No Yes Yes Yes No Yes

28

Appendix B Table 2. Statistical Test for Difference in Grading Factors Between a Buying Point and a Central Grading Site, Runner Only (283), 2006 Crop Year Mean Grading Factors % Foreign Materials % LSK % Moisture % SMKRS % Sound Splits % Total SMK % Other Kernels % Total Damage % Hulls % Total Kernels and Hulls % Freeze Damage % Concealed RMD Buying Point 3.75 2.98 8.45 63.67 6.85 70.72 5.34 0.61 23.15 99.82 0.01 0.03 Central Site 3.45 3.14 7.68 62.38 7.52 69.90 5.41 1.14 23.20 99.64 0.01 0.09 Statistical Difference Yes Yes Yes Yes Yes Yes No Yes No Yes No Yes

29

Appendix B Table 3. Statistical Test for Difference in Grading Factors Between a Buying Point and a Central Grading Site, Spanish Only (20), 2006 Crop Year Mean Grading Factors % Foreign Materials % LSK % Moisture % SMKRS % Sound Splits % Total SMK % Other Kernels % Total Damage % Hulls % Total Kernels and Hulls % Freeze Damage % Concealed RMD Buying Point 4.60 1.45 8.10 67.45 3.30 70.75 4.15 0.00 24.80 99.70 0.00 0.00 Central Site 5.65 1.50 7.08 65.90 4.30 70.20 4.15 0.05 25.20 99.60 0.00 0.01 Statistical Difference No No Yes Yes Yes Yes No No Yes No No No

30

Appendix B Table 4. Statistical Test for Difference in Grading Factors Between a Buying Point and a Central Grading Site, Valencia Only (27), 2006 Crop Year Mean Grading Factors % Foreign Materials % LSK % Moisture % SMKRS % Sound Splits % Total SMK % Other Kernels % Total Damage % Hulls % Total Kernels and Hulls % Freeze Damage % Concealed RMD % Broken % Discolor Buying Point 5.52 4.81 7.71 67.85 1.63 69.48 4.33 0.19 25.59 99.59 0.00 0.00 2.44 16.37 Central Site 4.85 5.30 6.75 66.37 2.22 68.59 4.26 0.89 26.00 99.74 0.00 0.01 5.63 10.74 Statistical Difference Yes Yes Yes Yes Yes Yes No Yes Yes No No No Yes Yes

31

Appendix B Table 5. Statistical Test for Difference in Grading Factors Between a Buying Point and a Central Grading Site, Virginia Only (88), 2006 Crop Year Mean Grading Factors % Foreign Materials % LSK % Moisture % SMKRS % Sound Splits % Total SMK % Other Kernels % Total Damage % Hulls % Total Kernels and Hulls % Freeze Damage % Concealed RMD % Fancy % Net ELK Buying Point 3.84 2.78 7.76 62.19 3.69 65.89 4.27 0.45 29.11 99.73 0.03 0.02 79.07 41.18 Central Site 3.91 2.95 7.23 59.80 4.53 64.33 4.39 1.57 29.39 99.67 0.06 0.07 77.61 38.86 Statistical Difference No No Yes Yes Yes Yes No Yes No No No Yes Yes Yes

32
2007 Crop Year Results

2007 Crop Year Sample Sizes: Southeast region 520; Virginia-Carolina region 476; Southwest region 497; Total sample 1,493.

Appendix B Table 6. Statistical Test for Difference in Grading Factors and Grade Value Between a Buying Point and a Central Grading Site, 2007 Crop Year, Southeastern Region Mean Grading Factors Foreign Material Sample (grams) % Foreign Materials % LSK % Moisture % SMKRS % Sound Splits % Total SMK % Other Kernels % Total Damage % Total Kernels % Hulls % Total Kernels and Hulls % Freeze Damage % Concealed RMD Grade Value ($/ton FSP) Buying Point 1727.22 3.00 2.79 8.49 65.12 6.60 71.72 4.21 0.76 76.72 23.08 99.96 0.05 0.05 352.31 Central Site 1825.07 3.11 3.00 8.07 63.32 7.30 70.61 4.21 168 76.51 23.01 99.53 0.14 0.14 350.37 Statistical Difference Yes No Yes Yes Yes Yes Yes No Yes Yes No Yes Yes Yes Yes

33
Appendix B Table 7. Statistical Test for Difference in Grading Factors and Grade Value Between a Buying Point and a Central Grading Site, 2007 Crop Year, Virginia-Carolina Region Mean Grading Factors Foreign Material Sample (grams) % Foreign Materials % LSK % Moisture % SMKRS % Sound Splits % Total SMK % Other Kernels % Total Damage % Total Kernels % Hulls % Total Kernels and Hulls % Freeze Damage % Concealed RMD % Fancy % Net ELK Grade Value ($/ton FSP) Buying Point 1793.44 3.86 2.93 8.25 65.01 3.97 68.77 2.10 0.67 71.75 27.91 99.65 0.00 0.04 80.23 43.40 353.31 Central Site 1856.92 3.82 2.95 7.76 65.22 3.59 68.81 1.99 0.74 71.74 28.20 99.74 0.00 0.09 81.22 43.22 358.86 Statistical Difference Yes No No Yes No Yes Yes Yes Yes Yes Yes Yes No Yes Yes No Yes

34
Appendix B Table 8. Statistical Test for Difference in Grading Factors and Grade Value Between a Buying Point and a Central Grading Site, 2007 Crop Year, Southwestern Region Mean Grading Factors Foreign Material Sample (grams) % Foreign Materials % LSK % Moisture % SMKRS % Sound Splits % Total SMK % Other Kernels % Total Damage % Total Kernels % Hulls % Total Kernels and Hulls % Freeze Damage % Concealed RMD % Fancy % Net ELK Grade Value ($/ton FSP) Buying Point 1696.38 4.43 4.28 8.17 66.81 8.32 75.14 3.01 0.26 78.40 21.59 99.99 0.00 0.00 1.91 0.89 368.18 Central Site 1589.67 4.04 4.55 7.16 65.97 8.80 74.77 3.08 0.17 78.02 21.98 100.00 0.00 0.00 5.09 2.36 374.73 Statistical Difference Yes Yes Yes Yes Yes Yes Yes No Yes Yes Yes Yes No No Yes Yes Yes

35
APPENDIX C

2007 Report

Enhancement of Peanut Grading with a Portable Microwave Sensor Deployable in Static and Dynamic Applications

By

Micah A. Lewis1,2, Samir Trabelsi1,2, Stuart O. Nelson1,2, Ernest W. Tollner2 USDA ARS QSARU Russell Research Center P. O. Box 5677 Athens, GA 30604-5677 2 Biological and Agricultural Engineering University of Georgia Driftmier Engineering Center Athens, GA 30604
1

ABSTRACT With a larger annual supply of peanuts and shorter window for harvesting each year, the optimization of the efficiency of grading practices has become a key issue. Present day grading practices are exhaustive in time, labor and cost. Majority of the properties that make up the grade of the peanuts have to be determined separately. This has led to the investigation of using microwaves to sense dielectric properties in peanuts in the determination of physical properties such as kernel moisture content and bulk density. Current research shows that this methodology is effective in the determination of such properties with the preferred efficiency. The goal here is to package this technology into an economical, portable sensor useful in real-life applications, static or dynamic. The projected cost of the sensor is no more than $2000.

36 INTRODUCTION The objective of this research is to develop a portable, microwave sensor for nondestructive determination of physical properties, such as kernel moisture content and bulk density (), in unshelled peanuts which would be deployable in static and dynamic applications. The technology utilized in this research is permittivity analysis by way of microwave measurements. From the complex permittivity ( = ' - j''), which is comprised of the dielectric constant (') and the dielectric loss factor (''), it is possible to determine the bulk density and the moisture content by identifying correlations between the dielectric properties and the physical properties of interest. The hypothesis is that the efficiency and accuracy observed in permittivity measurements using the sophisticated vector network analyzer (VNA) and horn lens antennas can be replicated economically using an inexpensive prototype consisting of simple circuitry and patch antennas. The experimental plan was designed to fully test this hypothesis introducing no bias or known errors. To ensure a full investigation, measurements were planned for all four types of peanuts. EXPERIMENTATION From August 21 November 20, 2007, permittivity measurements were taken on Georgia Runners, Texas Runners (high oleic content) and Virginias. From each variety, eight samples, each of pods and kernels, were conditioned evenly at moistures from 7% to 21%. At each moisture, measurements were taken at three different densities: loosely packed, partially settled and well settled. Due to temperature fluctuation that is experienced in real-life applications, measurements were performed at six different temperatures: 0C, 17C, 24C, 35C, 45C and 60C. Measurements with the VNA were taken from 2 to 18 GHz; however, due to limitations of components, measurements were taken with the prototype sensor at 5.8 GHz. To ensure integrity of measurements, no changes were made to the samples between measurements on the two systems. Before compressing for a higher density, each sample would be measured identically with the VNA and the prototype. At the conclusion of each measurement sequence, true density and moisture were determined for calibration and reference purposes. Density was calculated using the mass of the sample and the volume of the sample holder. Moisture content was determined using the conventional oven method from the ASAE Standards. The peanuts were dried for 6 h at 130C, and the moisture content was determined by dividing the weight loss from drying by the original weight of the sample. CURRENT RESULTS Preliminary results indicate that measurements taken with the VNA and the prototype compare well. The difference in measurements between the two systems is within a 95% confidence interval, proving a correspondence in performance of both systems. For the Texas runner farmer stock peanuts, both regressions for pods have an r2 value of 0.98.

37 For the Texas runner shelled peanuts, the r2 values of the regressions for kernels from the prototype and VNA are 0.96 and 0.98, respectively. Similar results have been observed in the Georgia runners and Virginias. The density prediction is equally as impressive. The regression used to determine density comes from the complex plane, a correlation of the dielectric loss factor ('') plotted versus the dielectric constant ('); both normalized by density. From Texas runners pods samples ranging in bulk density from 0.2944 to 0.3671 g/cm3, the standard error of calibration (SEC) from measurements performed using the VNA was 0.01292 g/cm3, and the SEC from measurements using the prototype was 0.01478 g/cm3. FUTURE WORK Measurements are still pending on Spanish and Valencia peanuts due to lack of arrival. However, upon their arrival, experimentation will be performed just as on the three types currently under investigation. More analysis will be performed to develop ways to derive kernel properties from measurements of the unshelled peanuts. Also, more experimentation is planned to investigate the dielectric properties at different temperatures. This will yield a robust calibration with temperature compensation. The second year (2008) will also involve preliminary field tests with the prototype during the harvest season.

38

APPENDIX D

2006 Crop Year Results

2006 Crop Year Sample Sizes: Runner 283; Virginia 88; Valencia 27; Spanish 20; Total - 418

39
Appendix D Table 1. Statistical Test for Difference in Times (Minutes) to Grading a Sample Between the New and Current Methods, 2006 Crop Year Mean Times Samples Overall Runner Spanish Valencia Virginia New 59 57 59 71 62 Current 66 68 42 51 67 Statistical Difference Yes Yes Yes Yes No

Appendix D Table 2. Statistical Test for Difference in Foreign Material Separator Times (Minutes, Stop Watch) Between the New and Current Equipment, 2006 Crop Year Mean Times Samples Overall Runner Spanish Valencia Virginia New 11 11 9 12 11 Current 11 11 10 11 10 Statistical Difference No No No No No

Appendix D Table 3. Statistical Test for Difference in Sheller Completion Times (Minutes, Stop Watch) Between the New and Current Equipment, 2006 Crop Year Mean Times Samples Overall Runner Spanish Valencia Virginia New 12 13 10 15 9 Current 10 11 8 7 8 Statistical Difference Yes Yes No Yes Yes

40

Appendix D Table 4. Statistical Test for Difference in Grading Factors at a Central Site Between New and Current Grading Methods, All Samples (418), 2006 Crop Year Mean Grading Factors % Foreign Materials % LSK % Moisture % SMKRS % Sound Splits % Total SMKSS % Total Damage % Other Kernels % Hulls % Total Kernels and Hulls % Freeze Damage % Concealed RMD New 3.88 3.12 7.44 50.07 14.41 64.48 0.94 8.51 25.59 99.52 0.01 0.03 Current 3.74 3.17 7.50 62.26 6.39 68.66 1.16 5.06 24.78 99.65 0.02 0.08 Statistical Difference Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

41

Appendix D Table 5. Statistical Test for Difference in Grading Factors at a Central Site Between New and Current Grading Methods, Runner (283), 2006 Crop Year Mean Grading Factors % Foreign Materials % LSK % Moisture % SMKRS % Sound Splits % Total SMKSS % Total Damage % Other Kernels % Hulls % Total Kernels and Hulls % Freeze Damage % Concealed RMD New 3.58 3.11 7.64 48.97 15.84 64.82 9.61 0.92 24.12 99.47 0.00 0.03 Current 3.45 3.14 7.68 62.38 7.52 69.90 5.41 1.14 23.20 99.64 0.01 0.09 Statistical Difference Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

42

Appendix D Table 6. Statistical Test for Difference in Grading Factors at a Central Site Between New and Current Grading Methods, Spanish (20), 2006 Crop Year Mean Grading Factors % Foreign Materials % LSK % Moisture % SMKRS % Sound Splits % Total SMKSS % Total Damage % Other Kernels % Hulls % Total Kernels and Hulls % Freeze Damage % Concealed RMD New 5.20 1.40 6.95 54.10 12.00 66.10 9.95 0.25 26.05 99.35 0.00 0.00 Current 5.65 1.50 7.08 65.90 4.30 70.20 4.15 0.05 25.20 99.60 0.00 0.01 Statistical Difference No No Yes Yes Yes Yes Yes No Yes No No No

43

Appendix D Table 7. Statistical Test for Difference in Grading Factors at a Central Site Between New and Current Grading Methods, Valencia (27), 2006 Crop Year Mean Grading Factors % Foreign Materials % LSK % Moisture % SMKRS % Sound Splits % Total SMKSS % Total Damage % Other Kernels % Hulls % Total Kernels and Hulls % Freeze Damage % Concealed RMD % Broken % Discolor New 4.81 5.19 6.73 62.93 4.96 67.89 4.85 0.70 26.52 99.96 0.00 0.00 6.44 14.11 Current 4.85 5.30 6.75 66.37 2.22 68.59 4.26 0.89 26.00 99.74 0.00 0.01 5.63 10.74 Statistical Difference No No No Yes Yes No Yes No Yes No No No No No

44

Appendix D Table 8. Statistical Test for Difference in Grading Factors at a Central Site Between New and Current Grading Methods, Virginia Only (88), 2006 Crop Year Mean Grading Factors % Foreign Materials % LSK % Moisture % SMKRS % Sound Splits % Total SMKSS % Total Damage % Other Kernels % Hulls % Total Kernels and Hulls % Freeze Damage % Concealed RMD % Fancy % Net ELK New 4.26 2.90 7.15 48.74 13.23 61.97 6.43 1.25 29.92 99.57 0.03 0.04 74.61 31.00 Current 3.91 2.95 7.23 59.80 4.53 64.33 4.39 1.57 29.39 99.67 0.06 0.07 77.61 38.86 Statistical Difference Yes No Yes Yes Yes Yes Yes Yes Yes No No Yes Yes Yes

45 APPENDIX E Evaluation of the X-Ray for Grading Farmer Stock Peanuts: 2007
Hank Sheppard, Marshall Lamb, and Christopher Butts USDA, ARS, National Peanut Research Laboratory, Dawson, GA The current grading system was developed in the 1940s. At that time the peanut harvest season was much longer, 3 to 4 months. With changes in planting dates and higher capacity harvesting equipment, the bulk of the season is now as short as 4 weeks. The current grading equipment can not keep up with this rush of samples and in years where there is a drought this slows the process down further. The current grading system requires numerous, labor intensive steps to complete a grade sample. The grading system speed needs to be improved to match the flow of farmer stock peanut deliveries. For an individual sample, each of the grading steps must be conducted separately during the grading process with measurements taken at the end of each process before the sample can go to the next process. Under current procedures it takes an estimated 11 to 14 minutes and a minimum of 4 employees to complete a farmer stock grade. One option identified by members of the US peanut industry as a potential device to improve farmer stock peanut grading is an X-ray machine. The machine has been implemented in many other agricultural and industrial settings for quality assurance demonstrating excellent results. In a small, preliminary study the X-Ray machine was tested to see if it could accurately determine the meat content of farmer stock peanuts. This small study showed very good results. The data in the graph is preliminary. However, regression analysis on the actual kernel weight versus the kernel weight calculated by the X-Ray machine resulted in a R2=0.9924. Results from the X-Ray machine indicate the ability to categorize differing components of a farmer stock grade sample. The results of this preliminary study warrant that a comprehensive, statistically balanced study should be conducted in the Fall of 2008 to compare the X-Ray machine with the current farmer stock grading system.

190

Calculated weigth of kernels (grams)

180 170 160 150 140 130 120 110 100 90 80 80 90 100 110

y = x - 0.0002
R2 = 0.9924

120

130

140

150

160

170

180

190

Real weigth of kernels (grams)