Top Navigation Bar

ASAE Conference Proceeding

This is not a peer-reviewed article.

Comparison of Pilot-scale Batch and Semi-continuous Aerobic Thermophilic Swine Waste Reactor Energy Production

Z. Wang and J. W. Blackburn

Pp. 172-181 in the Animal, Agricultural and Food Processing Wastes, Proceedings of the Ninth International Symposium, 11-14 October 2003 (Raleigh, North Carolina, USA), ed. Robert Burns. ,11 October 2003 . ASAE Pub #701P1203

Abstract

Over the past five years, an advanced aerobic thermophilic process for treatment of livestock waste, especially swine waste, has been developed at Southern Illinois University at Carbondale. Like aerobic thermophilic systems of the past, it successfully bio-oxidizes the waste, including odor-causing compounds, and produces enough energy to meet the thermophilic needs of the process (operation at 55-60 ° C). As an improvement on the past, extra energy is produced for farm use in drying and heating. This extra energy increases the profitability of the process.

This paper will focus on the ongoing activity of conversion of the batch system into a semi-continuous system design more suitable for a full-scale demonstration under progress. Three data sets are now available, one from batch operating on gestation building waste, a second from semi-continuous operation of a gestation waste at system residence times of around 6 days and a third, running semi-continuous on finishing waste at system residence times of over 13 days. Biological heat production is calculated from COD removals and all cases are similar with total biological heat production calculated in the range of 7-9 kW 100 hogs-1.

KEYWORDS. Swine, livestock, wastes, manure, aerobic treatment, aerobic thermophilic, heat, energy

Introduction

While a few studies on aerobic thermophilic treatment in the United States have been conducted in the past thirty years, and U.S patents have been filed on the topic, this technology has lagged the application of anaerobic or ananerobic-thermophilic biogas production in this country. Aerobic-thermophilic systems were operated as a cost center, with energy production expected only for autothermal operation. Investigators in Scotland had forseen, modeled and operated aerobic-thermophilic units with the goal of heat generation (Baines, et al., 1986; Svoboda and Evans, 1987; Svoboda and Fallowfield, 1989). Among their accomplishments were developing a model to project designs into the 50 ° C range and operating a farm-scale unit where the reactor operated typically in the 30-40 ° C range. That temperature was then boosted to more useful temperatures by heat pump. Their approach to aeration was very important. They used a surface-mounted agitator in a closed tank. Air input was regulated by a control system sensing low levels of oxygen. In practice, these investigators employed an air recycle system where the air recycle was internal to the reactor. Each volume of air had multiple exposure to the liquid, thus increasing the oxidant availability for the bio-oxidation reactor with lower additions of fresh air. They report a transfer efficiency of 1-2 kg O2/kWh of energy input. This value will be compared with our work and others later in this report.

Another notable large-scale effort to use aerobic thermophilic treatment, which is still in use, is the work by Oechsner and Doll (2000). In this case, only the pathogen-removing ability of the technology was exploited at full-scale. For high levels of pathogen removal, a semi-continuous operation was used with one reactor operating for 3 days residence time and a second in series with the same residence time. Successful pathogen removal has been reported by Oechsner and Doll. In terms of oxygen transfer efficiency, two 3 kW-nominal, completely submerged-motor aerators with fresh air suction were used—one in each vessel. Power consumption was reported for each of the motors as 1-2 kW and each vessel was 15 m3 in volume. We may estimate the power utilized per kg of COD removed from this work even though the authors did not report COD. With a removal of volatile solids (VS) of about 22%, a feed of VS of 5.3% and a total reactor volume of 30 m3 and residence time of 6 days, 58.3 kg/day of VS were being removed. For estimating purposes and using a ratio of COD/VS in swine (ASAE, 2000), 59 kg/day of COD were being removed. Assuming a minimal 2 kW total power usage in both of the submerged motors together, 1.2 kg of COD (O2 uptake) / kWh were estimated.

This study’s objectives are to determine if the biological heat production in batch aerobic thermophilic reactors are similar to heat production in semi-continuous systems. This is an important consideration since scale-up from the pilot data to full-scale designs are underway.

Southern Illinois University Batch Aerobic Thermophilic Reactor

Equipment

Seventeen batch experiments were carried out in the reactor covered with 1 cm of foil-faced polyethylene insulation (Figures 1 and 2). The agitator was a fixed-mounted mixer manufactured by Lightnin ™ with a maximum rating of about 0.75 hp at 300 rpm, two turbine impellers were on the shaft.

The reactor used in the experiments was a 1000-gallon (3.8 m3) stainless steel tank with a flat top and a conical bottom. The diameter of the tank is 1.65 m, and height is 1.96 m. Four legs raise the tank up about 40 cm from the ground.

At the top of the reactor, there is an access for mechanical devices and heat transfer fluids’ piping connections: a fixed mount mixer was equipped at the center, hot water inlet and outlet, cooling water inlet and outlet, air inlet, recycle air outlet and off gas outlet. At the bottom of the reactor, a 3-inch PVC gate valve was used for reactor product.

Inside of the tank, the air pipe reaches to the bottom supporting rack (15 cm above the bottom) along the inner wall like an L, and a membrane air diffuser is used to break the air stream into fine air bubbles and distribute them evenly to the waste. At the center of the reactor there are two coils of 1/2” TFE tubes used as heat exchangers for both hot water, when heating up the reactor and for cooling water, when taking out the excessive heat.

The blower is a type “S” Roots blower. A twin-fan compact-cross-finned heat exchanger is used for removing heat from the reactor. When the reactor temperature condition reaches the set points, cooling water removes heat from the reactor by circulating coolant through the coil in the reactor and back to the cooling water tank. The reactor temperature is controlled by a Honeywell controller.

The data acquired are fresh air and recycle air flow-rates, batch air (fresh and recycle) flow-rate, hot water and cooling water flow-rates, and hot water and cooling water inlet and outlet temperatures. The data acquisition program is LabView 5.0 and acquires data each minute.

Off gas flows from reactor through the recycle line to the absorber. The absorber is a CPVC tower, 1.3 m in height, 34 cm in diameter and about 18 cm above the ground. Off gas feeds to the bottom and passes up through the packing which is a 1 cm plastic ring. At the top of the absorber, there is an acid solution distributor. Sulfuric acid is first pumped from an acid tank into the agitated recirculation tank, where there is a pH controller keeping the solution in the recirculation liquid tank pH at 5. Then, the solution in the recirculation tank is pumped from the tank to the scrubber, absorbing ammonia from the offgas.

Swine waste used in the batch reactor research was from the gestation or finishing buildings of the SIUC Swine Center. A honey wagon delivers the waste to the reactor and fills it to about 90% of its content. Then, the water heater begins to heat the waste with the agitator at about 285 rpm. In the wintertime, it takes 3 days to heat the reactor temperature up to 55 ° C. In summer, it takes two days. When the reactor temperature reaches thermophilic temperature, fresh air with or without recycle air (depends on the running pattern) was started. When recycle is employed, the off gas treatment system will be turned on at the time air is started. The swine waste would be batch treated for at least 6 days after air was turned on. After treatment, the reactor product is unloaded by the same honey-wagon from the reactor and applied to the fields at the SIUC Agriculture Center.

Figure 1. Diagram of Batch Pilot System

172-181isaafpw_files/image1.gif

From the day the reactor was loaded to the last day when it was emptied, 50-500 ml samples were taken and stored in the refrigerator for COD and total solids measurements.

Methods

The COD test procedures follow Hach Standard Method 8000, a dichromate reactor digestion method. Two ml of the 1:100 deionized water diluted sample were transferred from a volumetric flask to a prepared Hach COD digestion reagent vial with a wide mouth pipet. A blank (filled with two ml deionized water) was prepared for calibration. The vial was inverted several times to mix the contents. Then the sample vial was preheated for two hours at about 150ºC, inverted again when it was still warm. The sample then was analyzed by colorimetric measurement on a spectrophotometer at a wavelength of 420 nm. Total solids were measured daily. The samples were thoroughly mixed and then 5-10 grams were transferred to a tared 15 ml scintillation vial. Weight of the vial plus swine waste was recorded and the vial was then placed in an oven with temperature at around 110ºC. After up to 24 hours, the swine waste was taken out of the oven for weighing. The percent total solids were calculated from the weight difference of the swine waste before and after drying. Both COD and total solids measurements were performed in triplicates.

Figure 2. Reactor Detail

172-181isaafpw_files/image2.gif

Conversion to a Semi-continuous System

Recently, funding for design and construction of a semi-continuous farm-scale system has been received. We believe that a semi-continuous system may have various advantages including a more even heat production rate. We have constructed a building at our SIUC Swine Center to house and test a semi-continuous pilot system. Two semi-continuous data sets are presented in this paper and are compared to a batch pilot data set.

Equipment and Operation

The same reactor was used as from the batch experiments. Upgrades include improved temperature and flow measurements, the design of the scrubber section and the blower system. In addition, a 1500 gal feed tank and effluent tank were added. The feed tank is filled with waste from the pits of the gestation or grower/finisher building. The effluent tank is empties when filled. On a preset time interval, a programmable logic controller will activate the transfer of a set amount of reactor liquid from the reactor to the effluent tank. Then it replaces fresh waste in the reactor. By controlling the frequency of this operation, the residence time for the reactor is set. Temperature control is established as was discussed earlier. Figure 3 presents a diagram of the semi-continuous system.

The operation of the system includes initial startup with fresh waste in the reactor and a heat-up and aeration stage similar to the batch operation. Then after about six days, the PLC controller is started and semi-continuous operation begins. Samples are taken from the fresh waste feed tank and from the reactor. In this way material balance and energy balances may be calculated on a daily average basis.

Heat Balance Calculations

For the batch and semi-continuous reactors, the Daily Biological Heat Production Rate, B, was calculated using the reduction in COD. It was developed for this work and is consistent with other work by the author and has been corroborated with thermal measurements.

172-181isaafpw_files/image3.gif

Where: B, the total biological heat production rate is in kW, F is the fresh waste feed rate in kg/day, CODf is the COD fractional concentration (wt. percent /100), CODr is the fractional daily reduction of COD, and ? hcod is the specific energy released from the oxidation of COD (kJ/kg COD) and was found to be 13.9 MJ/kg of COD in earlier work (Cheng and Blackburn 2003).

We normalized the energy release rate to the dry solids in the feed and therefore the number of hogs producing the feed to the reactor. If a finishing hog in its growth process is 65 kg and we use the waste production factor given in ASAE, 2000, an amount of solids is generated for each animal irrespective of the water dilution of the waste and any subsequent solids degradation. This works out to be 0.715 kg day-1 hog-1.

172-181isaafpw_files/image4.gif

Where, n is calculated as the equivalent number of finishing hogs feeding waste to the system each day, and TSfeed is the weight fraction of solids in the feed. The rate of energy production of the system based on these assumptions in kW/100 hogs is: B • 100/n. This equation is unreferenced, however has been used in prior papers of the author as a way to compare heat production from system-to-system and scale-to-scale. While some disagreement may arise from the use of ASAE data when Midwest Plan Survey Data are available, it is used consistently throughout this author’s work for easy comparisons.

The calculation was similar for both batch and semi-continuous designs except the time interval included the batch run from when air was started through the 6 or more days of batch residence time. B • 100/n for the batch runs included the lag time for the biological reaction to begin, while the time interval for the semi-continuous runs was calculated daily for each feed tank load. Within a single feed tank batch, the power output was more constant than the daily variation for a batch run.

Results

Heat Production

Total biological heat production is calculated from COD removal and a COD heat value of 13.9 MJ/kg of COD removed (Cheng and Blackburn, 2003). Results from the batch runs on fresh gestation building waste are presented in Table 1. Data reported include the reactor air recycle rate, the fresh air feed rate, the initial total solids and COD concentrations, the COD and TS percent removals (over the six-day period), and the Biological heat production rates by estimation of energy produced through COD removal per 100 (65 kg) animals, each producing a constant per-capita 0.715 kg TS day-1 (ASAE, 2000). Residence times were six days and fresh air flow averages for the 40-60% recycle case were 7.4 cfm. In this last column, energy production independent of feed dilution may be estimated and compared. Over a feed concentration span of 1.0 to 3.7 % TS, the 40 and 60% reactor air recycle cases averaged 8.9 kW per 100 hogs and had an average COD removal of 58%. No published data are conducted where heat production was studied in a reactor with a known air recycle rate.

Table 2 provides daily sequential data from a semi-continuous run where the batch feed tank held fresh gestation pit building waste at 2.92% TS and a concentration of 75.5 g l-1 of COD. Two days of results were presented as a preliminary indication of semi-continuous reactor performance. Further data were not available. Residence times were 6.6 and 7.3 days respectively and fresh air flow averaged 4.5 cfm. COD removals were higher than the batch 60% recycle runs (average 58%) with an average of 64% for the two days reported.

Table 3 presents operating data and results from a semi-continuous run on finishing waste. The feed concentration was relatively strong at 3.25% TS and 105 gl-1 COD. Fresh air flow averaged 9.5 cfm while the air recycle ratio was about 53%. COD removal averaged about 65% and TS removal averaged 24.2%. Average biological heat production was 6.8 kW per 100 standardized size hogs.

Semi-continuous runs were nearly as high in biological heat production as was the batch series of runs. This helps in the system full-scale design process as semi-continuous results may be expected to be similar to batch and semi-continuous pilot results.

A further comparison may be made between our very preliminary semi-continuous runs and the large scale runs reported from Scotland and Germany (These have been the only large runs found in the literature for hog wastes and will be our basis for comparision). A key efficiency parameter in this system is the oxygen transferred (or COD removed) per kWh of electrical power applied. As noted earlier, in Scotland, a range from 1-2 kg oxygen transferred per kWh were reported and from Germany a value of 1.2 kg COD removed per kWh was calculated from reported data.

The blower power input for the semi-continuous system modeled to systems on the order of the European cases indicate a connected power usage of 0.3 kW of blower power and an estimated 0.15 kW of mixer power applied totaling 0.45kW power applied in air supply and mixing. From 1.2 to 2 kg oxygen was consumed per kWh of electricity applied. This value favorably compares to the other large-scale cases.

Figure 3. Diagram of the Southern Illinois University Advanced, Semi-continuous Aerobic Thermophilic System

172-181isaafpw_files/image5.gif

Table 1. Batch System Operation Conditions and Performance at the End of Six Days—Gestation Building Waste

Run

No

Air Feed

(% recycle reactor offgas)

Ave fresh air (cfm)

Initial

Total Solid

(%)

Initial

COD

(g/l)

COD removal

(%)

Solid removal

(%)

B • 100/n: Ave. biological heat production rate by COD consumption

(kW/100 hogs)

7

0

12.09

4.267

51.4

47.3

43.7

6.5

8

0

16

1.583

23

52.7

36.6

8.9

9

0

16

2.697

33.9

61.1

35.7

8.0

10

0

10

4.587

64

33.8

23.0

5.5

11

40

10

1.037

11.5

78.6

19.7

10.1

12

40

9.47

3.460

46

52.0

37.6

7.9

13

60

7.6

3.703

61.3

50.0

12.3

10.3

14

40

9.66

3.313

27

69.7

28.5

6.5

15

60

7.57

2.517

41.7

43.0

10.2

8.2

16

60

6.21

1.537

22

60.0

17.9

8.6

17

40

10.57

1.400

23.2

51.6

16.8

9.8

Table 2. Preliminary Results and Performance Data from the Final Two Days from the Semi-continuous System Operation of Waste from the Gestation Building Pit

Sample day

Air Feed

(% recycle reactor offgas)

Ave fresh air (cfm)

Nominal

Residence

Time

(days)

Total Solid

(%)

COD

(g/l)

COD removal

(%)

Solid removal

(%)

B • 100/n: Ave. biological heat production rate by COD consumption

(kW/100 hogs)

12/03/2002

Feed tank

2.92

75.5

12/9/2002

61

4.7

6.6

1.57

31.5

58.3

46.2

7.4

12/11/2002

51

4.4

7.3

1.72

22.4

70.3

41.1

7.3

Averages of Data

46

4.6

7.0

1.65

27.0

64.3

43.7

7.3

Table 3. Preliminary Results and Performance Data for Thirteen Final Days from the Semi-Continuous System Operating on Finishing Waste

Sample day

Air Feed

(% recycle reactor offgas)

Ave fresh air (cfm)

Nominal

Residence

Time

(days)

Total Solid

(%)

COD

(g/l)

COD removal

(%)

Solid removal

(%)

B • 100/n: Ave. biological heat production rate by COD consumption

(kW/100 hogs)

3/24/2002

Feed tank

3.25

105

3/18/2003

58

10.5

16.7

2.05

34.7

66.9

36.9

7.7

3/19/2003

54

8.6

14.7

2.26

38.3

63.5

25.9

7.3

3/20/2003

54

9.1

24.4

1.84

33.6

68.0

43.3

7.9

3/21/2003

54

8.4

14.2

2.54

39.1

62.8

21.8

7.2

3/22/2003

52

9.7

6.5

2.56

36.4

65.3

21.2

7.5

3/23/2003

53

9.4

24.8

2.74

35.1

66.6

15.7

7.7

3/24/2003

52

10.2

13.2

2.86

41.8

60.1

12.0

6.9

3/25/2003

52

9.9

13.9

1.73

36.6

65.1

46.8

7.5

3/26/2003

56

10.2

12.4

2.88

36.2

65.5

11.4

7.5

3/27/2003

51

9.6

16.8

1.85

33.2

68.3

44.0

7.9

3/28/2003

49

10.0

15.2

2.81

33.6

68.0

13.5

3.4

3/29/2003

52

9.6

8.69

2.8

42.5

59.5

13.8

4.9

3/30/2003

53

8.9

32.2

2.89

34.6

67.0

11.0

4.0

Averages of Data

53

9.5

16.4

2.45

36.6

65.1

24.4

6.8

Conclusion

Batch runs have been completed with the total biological heat release calculated to be 8.8 kw / 100 hogs by COD level and removal (Table 1) in a pilot-scale reactor. Calculation used only runs at 40 and 60% recycle. Pilot-scale (runs of semi-continuous aerobic-thermophilic processing of swine waste) have been completed for waste from a gestation barn pit (Table 2) and waste from an external pit from a finishing scraper barn (Table 3). Biological energy production per 100 standardized hogs as calculated by COD is similar for the first and second semi-continuous runs even though it was run on gestation and finishing wastes at different concentrations. We should be careful to note that the residence times for the second semi-continuous runs were considerably longer than that from the first. The Biological batch and semi-continuous operations agree well on this parameter.

The Semi-continuous system compared well with two other full-scale operations in Scotland and Germany in terms of oxygen transfer (kg COD removal) per unit electrical energy input (kWh). The pilot system achieved from 1.2 to 2.0 kg COD removal/kWh, matching the maximum oxygenating efficiency reported for large-scale aerobic-thermophilic systems running on pork or dairy waste. This indicates that the oxygen transfer efficiency and therefore a normalized heat generation efficiency in Southern Illinois University Semi-continuous system has an energy production and oxygen utilization comparable to batch and other large-scale aerobic thermophilic operations, even though the modes of aeration are very differerent.

Acknowledgements

The gratitude of the authors goes to the primary supporting agency of Illinois Council for Agricultural and Food Research, Swine Odor and Waste Management Strategic Research Initiative. We also thank Southern Illinois University at Carbondale, Department of Mechanical Engineering and Energy Processes, College of Engineering, College of Agriculture, Graduate School, and Swine Center for extensive help and matching support. In a directly-related project, we acknowledge the Illinois Attorney General’s Office for funding to build and demonstrate a full-scale semi-continuous, aerobic-thermophilic system. Finally we thank Mr. Tom Rosenthal, Herdsman, and his coworkers at the Swine Center for helping to access and dispose of the swine materials necessary for these projects.

REFERENCES

ASAE Standards , 47th ed. 2000. D384.1. Manure production and characteristics. St. Joseph, Mich.: ASAE.

Baines S, IF Svoboda, MR Evans, and NJ Martin. 1986. A computer program for calculation of the extractable heat from aerobic treatment of animal wastes. J. Agric. Engng. Res. 34:133-140.

Cheng J and JW Blackburn. 2003. Heat production and kinetics from the laboratory-scale batch aerobic thermophilic processing of high strength swine waste. ASAE Trans. In press.

Oechsner H and L Doll. 2000. Inactivation of pathogens by using the aerobic-thermophilic stabilization process. In Animal, Agricultural and Food Pocessing Wastes . JA Moore, Ed. ASAE, St. Joseph, MI.

Svoboda IF and MR Evans. 1987. Heat from Aeration of Piggery Slurry. J. Agric. Engng. Res. 38:183-192.

Svoboda IF and HJ Fallowfield. 1989. An integrated piggery slurry treatment system with integrated heat recovery and high-rate algal ponds. Nat. Sci. Tech. 21:277-287.