STORAGE OF B100

  This project was designed to determine the extent of deterioration of Rapeseed B100 in storage. The study involved triplicate samples of B100 RME and B100 REE stored in both glass and steel containers at room temperature (inside) and at the local ambient outdoor temperatures (outside). The study was conducted for 24 months. At the beginning of the study and at 3 month intervals, samples were taken for measurement of peroxide value, acid value, density, viscosity, and heat of combustion. At the conclusion of the study, engine performance tests were conducted with the two year stored B100 REE and B100 RME, new B100 REE and B100 RME, and low sulphur diesel reference fuel.

LITERATURE REVIEW
Previous work on the storage stability of vegetable oil fuels includes a three year study reported by Klopfenstein and Walker (1984), Klopfenstein and Walker (1985), Klopfenstein and Walker (1986). In that study, samples of soybean B100 were stored in three different environments (indoor, outside and underground), two container types (steel and plastic) and two formulations (with and without antioxidants). Samples were analyzed initially and at four month intervals for: fatty acid composition, peroxide value, density and viscosity. Results favored underground storage, plastic-lined containers and the use of 0.5% butylated hydroxy toluene as an antioxidant.

A study performed at the University of Idaho by Korus (1983) and continued by Jo (1984) involved neat vegetable oils as fuels. Oil deterioration was measured as a function of storage conditions (aerobic and anaerobic at room temperatures) and vegetable oil composition (fatty acid saturation vs. unsaturation). Parameters measured at six month intervals over a two year period were peroxide values and fatty acid profiles. Engine testing was also performed using the stored oils to test injector coking as a function of fuel deterioration. Fuels tested were 50/50 blends of winter rapeseed, linoleic safflower and oleic safflower with diesel fuel and 100% 2-D as a reference. All vegetable oil fuel blends gave a statistically significant (a<0.05) increase in carbon deposits relative to diesel fuel with linoleic safflower having an injector coking area of 7.57 cm² relative to diesel fuel, 50% oleic safflower 5.01 cm² and 50% winter rapeseed 3.93 cm². Results indicated that deterioration was reduced by anaerobic storage and by high levels of saturated fatty acids in the oil.

OBJECTIVE

1. Produce batches of B100 RME and B100 REE and characterise their properties according to ASAE EP552.

2. Store triplicate sets of vented glass and steel containers of B100 in flammable solvent storage cabinets at both inside and outside ambient temperatures.

3. Analyze fuel properties (peroxide value, acid value, density, viscosity, and heat of combustion) from each storage container every 3 months for a 24 month period and compare with the initial values.

4. Compare two year stored B100 in short term engine performance tests with new B100.

MATERIALS AND METHODS
Fuel Preparation
132 litres each of B100 Rapeseed Methyl Esters and B100 Rapeseed Ethyl Esters were made at the beginning of the study. Samples of the were taken for triplicate determinations in all of the 5 test procedures at the beginning of this test and at each 3 month interval thereafter.

The B100 fuels were processed in a batch type reactor.

Storage Containers
Fuel containers for the study were: 4-litre brown glass reagent bottles and 8-litre steel pails with crimp on lids. All containers were filled 3/4 full and vented to the atmosphere to simulate actual fuel storage. The study used triplicate samples of B100. Each fuel was arranged in four different configurations: glass and steel containers stored at room temperature (inside) and at the local outside ambient temperature (outside) for a total of 24 samples.

Fuel Analysis
After each three month storage period samples were analyzed according to the following procedures (AOCS, 1987; ASTM, 1991): Peroxide Value, AOCS Cd 8b-90; Acid Value, ASTM Test D974; Density, ASTM Test D1298; Viscosity, ASTM Test D445; and Heat of Combustion, ASTM Test D240.

The fuels were characterized initially and after the two year storage study by evaluating the parameters in ASAE EP552. The tests for specific gravity, viscosity, cloud point, pour point, flash point, heat of combustion, total acid value, peroxide value, catalyst, and fatty acid composition were performed at the Analytical Lab, Department of Biological and Agricultural Engineering, University of Idaho. The boiling point, water and sediment, carbon residue, ash, sulfur, cetane number, copper corrosion, Karl Fischer water, particulate matter, iodine number, and the elemental analysis were performed by Phoenix Chemical Labs, Chicago, Illinois.

Short Term Engine Tests
Each of the 24, 4 litre samples from the storage study were for use in the short term engine testing. New batches of B100 were produced for comparative purposes. Phillips 66 low sulfur number two diesel fuel (2-D) was used in this study as a reference. The test engine was a John Deere (4239T) direct-injection, turbocharged diesel. This is a four cylinder engine with a bore of 106 mm, a stroke of 110 mm, a displacement of 3.917 L and a compression ratio 16.2:1. It has a high RPM of 2650, with 61 kW (82 HP) at 2500 RPM and 290 N m of torque at 1500 RPM. Attached to the engine was a General Electric 119 KW (159 HP) cradle type dynamometer.

Two short term engine performance test procedures were performed. The first was a rapid engine test to measure injector fouling in diesel engines using vegetable oil fuels (Korus, 1985). For this test the engine was operated at maximum power and 2500, 2300, 2100, 1900, 1700, and 1500 revolutions pa minute (RPM) for 10 minutes at each RPM. Readings of ambient air, opacity, exhaust, fuel, lube oil and intake air temperatures and opacity were acquired every 30 seconds. After each fuel test the injectors were removed and the carbonaceous tips were measured using machine vision. There was only enough fuel for one replication of this test.

The second test was a SAE torque test (SAE J1349, 1990). This test was performed under full throttle and full load conditions at 100 RPM intervals from 2600 to 1300 RPM. This test was replicated once also. Both test procedures were set up in random order. For more information on engine test procedures and equipment refer to (Hammond, 1996; Perkins et al., 1991).

Statistical Analysis
The storage study was set up as a randomized complete block with replication. Blocks were inside and outside storage locations. Treatments were fuels and containers (glass and steel). Each fuel and container type were replicated 3 times for a total of 12 containers inside and 12 containers outside. Statistical analyses were performed with SAS (SAS Institute, Cary, NC). This consisted of a statistical analysis (ANOVA) followed by a Tukey multiple range procedure to separate the means for each of the parameters, peroxide value, acid value, density, viscosity, and heat of combustion.

In addition a regression model was formulated for each parameter by using SAS. Initially a multiple regression was run on each parameter against the variables container, location, time, location and time, time squared and fuel. The variables in the models were coded as shown in Table 2.

Table 1. Storage Variables

RESULTS AND DISCUSSION
In the following discussion the nomenclature used is as follows: G=Glass, M=Metal, ME=B100 Methyl Ester, EE=B100 Ethyl Ester, I=Inside and O=Outside. The storage study was started in November, 1993 with analysis occurring in February, May, August, November, May, August, and November. Many of the trends observed in the results may be attributed to the time of year the samples were analyzed.

Peroxide Value
Peroxide values, measured in milliequivalents of peroxide per kilogram of sample for each sampling period, are shown in Figure 1. There was a consistent increase in peroxides over time with an acceleration of that increase from the 6th through the 18th month.

Figure 1. Peroxide value versus container type-fuel type location for two years.

Peroxide values for either fuel were not significantly affected by the type of container. Statistical analysis of the interactions between fuel and location, based on the overall means, indicated B100 REE was significantly higher between inside and outside storage for peroxide values. B100 RME showed no significant difference for interactions between fuel and location for peroxide values. There was a significant increase for peroxide from initial readings with B100 RME after six months. B100 REE showed a significant increase after three months.

Figure 2.Peroxide value versus two year storage study for B100 RME increased at a faster rate and B100 REE.

Figure 2 shows that after 6 months the peroxides in the B100 RME than in the B100 REE. Peroxide values at the outside location had a slower rate of increase than did the values at the inside location for the first six months. This was most likely due to an overall lower sample temperature during that time. Peroxide values at the outside location had a slower rate of increase than did the values at the inside location for the first six months. A steady increase of peroxide value over time was observed for B100 REE. Peroxide value for the B100 RME leveled off after eighteen months. At 24 months the peroxide value was 14.5 times higher for B100 RME and 13.7 times higher for B100 REE compared to the beginning value. Fuel stored outside had a peroxide value 14.7 times higher while fuel stored inside was 13.5 times higher compared to the beginning value.

The best fit regression model for the change in Peroxide value were as follows:

P_perox = -30.906*loc - 33.41*fuel + 35.19*time + 3.95*loc*time + 61.624
The R(2) value for this equation is 0.94 using the values in Table 2.

This equation, over the 24 months, shows that fuel type affected Peroxide by a change of 33.41, each quarter of time increased peroxide by 35.19, outside location decreased peroxide value by 30.906, and there was a small location by time interaction.

Table 3 represents the results of ANOVA analysis on test data over time with groupings according to sampling period. The Tukey multiple range procedure was used to find differences among group means.

Table 2. Results of ANOVA analysis for peroxide values on test data with groupings according to sampling period.

Acid Value
The acid values (Figure 3), measured in milligrams of KOH per gram of sample, increased with time as did the peroxide values. Since both of these values are related to autoxidation, the acid values naturally increase with an increase in peroxides because the esters first oxidize to form peroxides which then undergo complex reactions including a split into more reactive aldehydes which further oxidize into acids. Acids can also be formed when traces of water cause hydrolysis of the esters into alcohol and acids (Formo, 1979).

Figure 3. Acid value versus container type, fuel type, and fuel were not location for two years.

Acid values for either fuel were not significantly affected by the type of container. Statistical analysis of the interactions between fuel and location indicated B100 REE and B100 RME had a significant difference between inside and outside storage for acid values. The effects of time on B100 REE and B100 RME show that after nine months, there was a significant increase in acid values from initial readings and the increase accelerated toward the end of the 24 month period.

Figure 4. Acid value versus container type, fuel type, and location for two years.

Figure 94 shows acid value versus time for both B100 RME and B100 REE, and inside and outside locations. The acid values were fairly constant for the first 6 months then took a significant upward trend. The B100 RME acid values increased at a faster rate than the B100 REE values after 6 months as was the case with peroxides. The acid value of the outside samples lagged behind the inside samples. At 24 months the acid value was 10.3 times higher for B100 RME and 9.2 times higher for B100 REE compared to the beginning values. Fuel stored outside had acid values 9.0 times higher while fuels stored inside had acid values 10.5 times higher compared to the beginning value. The best fit regression model for the change in acid levels was as follows:

P_acid = -0.064*time - 0.172*fuel - 0.0187*loc*time + 0.0209*time2 + 0.473
The R2 value for this equation is 0.93 using the values in Table 9-2.

This model shows a significant time to location interaction and a quadratic relationship for time. The acid value was ­0.172 higher for fuel type (the negative sign indicates that B100 REE was lower than the B100 RME) and the outside samples had a 0.06 lower acid value than the inside samples.

Table 3 represents the results of ANOVA analysis on test data over time with groupings according to sampling period. The Tukey multiple range procedure was used to find differences among group means. A steady increase in acid values over the 24 month storage time is noted.

Table 3. Results of ANOVA analysis for acid values on test data with groupings according to sampling period.

Density
It was found that the density of the esters increased over time; however, outside samples tended downward during the first half of the study before increasing (Fig 9-5). Density values for either fuel were not significantly affected by the type of container. Statistical analysis of the interactions between fuel and location indicated B100 REE had a significant increase between inside and outside storage for density values. B100 RME showed no significant difference for interactions between fuel and location for density values. The effects of time on B100 RME show that after three months, there was a significant difference for density from initial readings. B100 REE showed a significant difference after six months.

Figure 5. Density versus container type, fuel type, and location for two years.

Figure 6 shows density versus time for both density of the B100 RME at the beginning of the study was higher than that of B100 REE and increased at a faster rate after 6 months. Significant differences were found among means from each sampling period with the exception of the density values of B100 REE between O and 3 months. Location was not a major influence in changing density values over time, however, the inside samples had higher density values than the outside samples. A 1.046% increase of density with time was measured. B100 RME density increased 1.22% and B100 REE density increased 0.88%. Fuel stored outside increased 1.048% and fuel stored inside 1.037%.

Figure 6. Density versus two year storage study for B100 RME and B100 REE.

The best fit regression model for the change in density values was as follows:

P_den = -0.4815*loc - 4.61*fuel + 0.114*time2 + 885.28
The R2 value for this equation is 0.966 using the values in Table 1.

The model shows a quadratic relationship for time. The density increment for fuel type was -4.61 (the negative sign indicates a lower density for B100 REE than B100 RME) and a -0.48 reduction in density for the outside stored samples.

Table 5 represents the results of ANOVA analysis on test data over time with groupings according to sampling period. The Tukey multiple range procedure was used to find differences among group means.

Table 4. Results of ANOVA analysis for density on test data with groupings

Viscosity
Viscosity tended to increase over time (Fig 7); however, the outside samples tended downward during the before going up. This could be due in part to lower average temperatures of the outside samples during the early period. Viscosity values for either fuel were not significantly affected by the type of container. The viscosity of B100 REE (Fig 8) started out about 10% higher than that of B100 RME but after year of storage was only 5% higher. B100 RME experienced a significant rise in viscosity after 6 months of storage and continued to rise significantly over the 9 to 24 month periods. B100 REE stayed fairly constant for the first 12 months and then increased at about the same rate as B100 RME for the final 12 months. The viscosity of B100 RME increased 23.1% and viscosity of B100 REE 16.8% over the 24 months.Inside and outside samples had nearly identical 19+% increases.

Figure 7. Viscosity at 40°C versus container type, fuel type, and 5% higher B100 RME.

Figure 8. Viscosity at 40°C for the two year period for B100 RME, B100 REE, and inside and outside locations.

The best fit regression model for the change in viscosity was as follows:

P_visc = -0.0807*loc + 0.0394*time + 0.5006*fuel + 0.0198*time2 + 5.46
The R2 value for this equation is 0.91 using the values in Table 9-2.

The model shows viscosity was 0.08 cSt lower for outside stored samples, the effect of fuel type changed viscosity by 0.5 cSt (B100 REE higher than B100 RME) and a quadratic effect of time.

Table 5 represents the results of ANOVA analysis on test data over time with groupings according to sampling period. The Tukey multiple range procedure was used to find differences among group means.

Table 5. Results of ANOVA analysis for density on test data with groupings according to sampling period.

Heat of Combustion
The values for heat of combustion (Fig 9) tended to decrease over time. This was most likely due to the breakdown of the fuel by oxidation which was verified by the increase in the peroxide and acid values. Heat of combustion values for either fuel were not significantly affected by the type of container. Statistical analysis of the interactions between fuel and location indicated B100 REE had a significant difference and B100 RME no significant difference between inside and outside storage for heat of combustion values. Figure 9-10 shows a decline in the heat of combustion of the two fuels with the exception of B100 REE between month 6 and 9. The heat values for B100 RME 12 months were significantly lower than values at both 0 and 3 months. The B100 REE heat of combustion values on the other hand showed a significant drop at 6 months which corresponds with a rise in the other four parameters at 6 months, however it then increased at 9 months, although not significantly. The final reading at 24 months was significantly lower than the beginning value for heat of combustion.

Figure 9. Gross heat of combustion versus container type, fuel type, and location for two years.

Figure 10. Gross heat of combustion for the two year storage period for B100 RME and B100 REE, and inside and outside locations.

Heat of combustion declined about 1.4% over the 24 months of storage. B100 RME declined 1.50% and B100 REE 1.27%. Inside and outside samples declined at about the same 1.4% rate.

The best fit regression model for the change in heat of combustion was as follows:

P_HoC = 19.94*loc +101.231*fuel - 2.89* time2 + 17254
The R2 value for this equation is 0.747 using the values in Table 9-2.

The model shows a quadratic change in heat of combustion with time, 101.2 Btu/lb less energy for B100 RME than for B100 REE and 19.94 Btu/lb more energy for samples stored outside compared the inside stored samples.

Table 6 represents the results of ANOVA analysis on heat of combustion test data over time with groupings according to sampling period. The Tukey multiple range procedure was used to find differences among group means.

Table 6. Results of ANOVA analysis for density on test data with groupings according to sampling period.

Fuel Characterization
Table 7 is the fuel characteristics for the initial and two year stored fuel. The SREE and SRME are the B100 REE and B100 RME fuels, respectively, stored for two years. The cetane number increased more than 12%, viscosity increased more than 16%, and particulate matter all increased during storage as did the peroxide value.

Table 7. Fuel Characterization Data

Figure 11. Power and torque plotted against engine speed for five fuels.

Short Term Engine Tests
At maximum power output, the fuel stored for two years increased 1.9% for B100 REE and 1.5% for B100 RME compared to newly produced fuel (Fig. 11). The stored fuels also increased maximum torque by 0.9%. The new fuels had 3.4% less power than 2D at maximum power and 1.4% less torque than 2-D at peak torque RPM.

Smoke density for the Biodiesel fuels were 18% to 52% that of 2-D. The smoke density for the stored B100 REE was 1.175 times more than that of the new B100 REE and the stored B100 RME produced 3% less smoke than the new B100 RME (Fig. 9-12)

Figure 12. Smoke density and fuel economy for five fuels at five different engine RPM's.

Fuel economy was compared at each RPM level (Fig. 12). For example, at 1500 RPM, the stored B100 RME used 1.13 percent more fuel than did the new B100 RME and the stored B100 REE used 1.2 percent more fuel than the new B100 REE.

The injector tip coking index for the 2-D, new B100 RME and B100 REE, and stored B100 RME and B100 REE fuels were found by dividing each of the fuel's coking area by the diesel coking area. The coking on the injector tips for the stored fuels was 7.8% more for B100 RME and 2.8% more for B100 REE. The coking observed for B100 RME in this test was extremely low (1.016, for SRME (1.102), for B100 REE (1.474), and 1.517 for SREE. This data was at or below other tests for Biodiesel reported by the authors recently. The effect due to storage was extremely small in both cases.

CONCLUSIONS

• Peroxide values were increased 14.5 times for B100 RME and 13.7 times for B100 REE after 24 months.

• Acid values were increased 10.3 times for the B100 RME and 9.2 times for B100 REE after 24 months.

• Density increased 1.22% for B100 RME and 0.88% for B100 REE after 24 months.

• Viscosity increased 23.1% for B100 RME and 16.87% for B100 REE after 24 months.

• Heat of Combustion declined 1.50% for B100 RME and 1.27% for B100 REE after 24 months.

• The cetane number increased more than 12% and particulate matter increased during storage.

• Regression models showed second order time effects for acid value, density, and viscosity.

• In general, outside samples stored slightly better than indoor samples. It was noted that during the autumn and winter, the change in samples stored outdoors was less than those inside. In the summer when outside temperatures were higher the changes were similar.

• The short term engine tests, including torque, power, and fuel economy, and coking tests showed very small amounts of change between the stored fuels and the new fuels. The high peroxide and acid values of the stored fuels had little effect on the performance of the engine in short term tests. Long term durability tests would be the next step needed to determine if these changes affect engine life. The small amount of fuel stored in this study precluded such testing.

REFERENCES
ASAE. 1995. Reporting of Fuel Properties When Testing Diesel Engines with Alternative Fuels Derived From Biological Materials. ASAE EP552. ASAE, St Joseph, MI 49085-9659.

ASTM. 1991. Standard Test Method for Kinematic Viscosity of Transparent and Opaque Liquids (and the Calculation of Dynamic Viscosity). Method D 445. Section 5, Vol.05.01. ASTM, 1916 Race St., Philadelphia, PA 19103-1178.

ASTM. 1991. Standard Practice for Density, Relative Density (Specific Gravity), or API Gravity of Crude Petroleum and Liquid Petroleum Products by Hydrometer Method. Method D 1298. Section 5, Vol.05.01. ASTM, 1916 Race St., Philadelphia, PA 19103-1178.

ASTM. 1991. Standard Test Method for Neutralization Number by Color-Indicator Titration. Method D 974-80. Section 5, Vol.OS.01. ASTM, 1916 Race St., Philadelphia, PA 191031178.

ASTM. 1991. Standard Test Method for Heat of Combustion of Liquid Hydrocarbon Fuels By Bomb Calorimeta. Method D 240. Section 5, Vol.05.01. ASTM, 1916 Race St., Philadelphia, PA 19103-1178.

AOCS. 1987. Peroxide Value (Acetic Acid - Iso-Octane Method). Method Cd 8b-90. Sampling and Analysis of Commercial Fats and Oils. Blackwell Scientific Publications Inc., PO Box 50009, Palo Alto, CA 94303.

Formo, Marvin W., E. Jungermann, F. Norris and N. Sonntag. 1979. Bailey's Industrial Oil and Fat Products. Volume 1 Chapter 2. John Wiley and Sons, NY.

Hammond, B.L. 1995. Performance and Durability Testing of Diesel Engines Using Ethyl and Methyl Ester Fuels. M.S. Thesis. Ag Engr Dept, University of Idaho, Moscow, ID 83843.

Jo, J. 1984. Improving Storage and Use of Vegetable Oil for Fuel. M.S. Thesis, University of Idaho, Chemical Engineering Department. Moscow, ID 83843

Klopfenstein, W. E., and H. S. Walker. 1984. Effects of a Year Storage on Soybean Oil Esters. USDA Southern Agricultural Energy Center, Solar and Biomass Energy Workshop (4th Annual Meeting). Atlanta, Georgia.

Klopfenstein, W. E., and H. S. Walker. 1985. Effects of Two Years of Storage on Soybean Oil Esters. USDA Southern Agricultural Energy Center, Solar and Biomass Energy Workshop (5th Annual Meeting). Atlanta, Georgia.

Klopfenstein, W. E., and H. S. Walker. 1986. Effects of Three Years of Storage on Fatty Acid Esters From Soybean Oil and Engine Performance on the Stored Fuel. USDA Southern Agricultural Energy Center, Solar and Biomass Energy Workshop (6th Annual Meeting). Atlanta, Georgia.

Korus, R A. 1983. Vegetable Oil Storage Stability. USDA Southern Agricultural Energy Center, Solar and Biomass Energy Workshop (3rd Annual Meeting). Atlanta, Georgia.

Perkins, L.A, C.L. Peterson and D.L. Auld. 1991. Durability Testing of Transesterified Winter Rapeseed Oil (Brassicanapus L.) As a Fuel in Small Bore, Multi Cylinder, DL CI Engines. SAE Technical Paper No. 911764. SAE, Warrendale, PA 15096-0001. 16pp.

Peterson, C.L., M. Feldman, R. Korus and D.L. Auld. 1991. Batch Type Transesterification of Ethyl and Methyl Esters of Winter Rape Oil. ASAE Paper No. 89-6569. ASAE, St. Joseph, MI 49085-9659.

Peterson, C.L., D.L. Reece, B.J. Hammond, J.C. Thompson and S.M. Beck. 1994. Processing, Characterization & Performance of Eight Fuels From Lipids. ASAE Paper No. 946531, ASAE, St. Joseph, MI 49085.

Reece, D.L. 1995. On-Road Testing of Rapeseed Biodiesel. M.S. Thesis. Ag Engr Dept, University of Idaho, Moscow, ID 83843.

SAE J1349. 1990. Engine Power Test Code - Spark and Compression Ignition - Net Power Rating. SAE, 400 Commonwealth Drive, Warrendale, PA 15096.