XXVIII Congreso Interamericano de Ingeniería Sanitaria y Ambiental Cancún, México, 27 al 31 de octubre, 2002

ANAEROBIC PILOT PLANT STUDIES FOR DAIRY PLANT WASTEWATERS

Reinaldo Antonio González Quevedo Burns & McDonnell Engineering Company Reinaldo González is a Chemical Engineer from the Universidad del Zulia, Venezuela, 1978. He holds MS and Ph.D. Degrees in Environmental Engineering from Oklahoma State University, Stillwater, Oklahoma. He worked for the Sistema Triestadal Torondoy Water Treatment Plant and Petroquímica de Venezuela, S.A in Venezuela. In the United States he has worked for Stover & Associates, Inc. in Oklahoma and since 1995 has been working as a Project Manager and Senior Process Engineer for Burns & McDonnell Engineering Company in Kansas City, Missouri. 9400 Ward Parkway – Kansas City – Missouri – 64114 – USA – Phone: (816) 822-3185 – Fax: (816) 822-3414 – e-mail: [email protected]. SUMMARY As a result of the diversity of the nature of environmental factors that could possibly be creating a negative influence on the successful operation of the anaerobic digester at the dairy facility pretreatment system, the decision was made to operate a number of pilot bench-scale anaerobic digesters, under controlled environmental conditions, in order to determine problem areas, and define proper environmental conditions for successful, reliable operations. A pilot study program was developed in conjunction with monitoring of the full-scale system in order to determine the impacts of environmental factors on the performance of the biological system and sludge settleability. The investigative program was segregated into two phases (three weeks each phase). The primary factors that were evaluated in the experimental pilot reactors included the following: Hydraulic loading rates and associated reactor retention times; different organic loading rates; pH and alkalinity control; supplemental micronutrients (Ni, Co, Cu, and Mo) addition; cultured bacteria addition; and addition of ultrafiltration permeate as feed stock in lieu of whey. The results of the investigative program indicated the following: The biological treatment kinetics indicated excellent anaerobic treatment characteristics; micronutrients addition significantly enhanced the methane gas production rate; gas quality remained approximately the same in all the reactors irrespective of environmental conditions, micronutrients addition, or cultured bacteria addition; a high level of dispersed solids was observed in all reactor effluents; adequate mixing in digester was a concern; degasification tank does not appear to be accomplishing effective degassing of effluent from digester; and significant amounts of flocculant sludge were carried out of the Lamella by high gas production and associated gas flotation. Key Words:

Dairy, anaerobic digester, micronutrients, loading,

INTRODUCTION The dairy plant process wastewater and a portion of the washwater generated at the dairy facility subject of this investigation are treated biologically in an anaerobic pretreatment system prior to discharge to the municipal system. The primary process flow schematic consists of whole whey storage, anaerobic suspended growth reactor (digester), degasification tank, inclined plate clarifier (Lamella), sludge handling facilities, and digester gas collection and handling system. The digester gas is compressed and used in the plant boilers for steam and heat generation. Caustic is used when required for pH adjustment and control in the digester. A major problem existed at the treatment facility since its startup relative to sludge settling and handling characteristics. The microorganisms grown in the digester never exhibited adequate settling characteristics in the Lamella. Polymer was used with some success in improved solids removal; however, excessive amounts of solids loss in the Lamella persisted. Solids concentration in the Lamella effluent consistently remained in the 4,000 mg/L to 6,000 mg/L range with occasional lower values with polymer usage. There are a number of factors that can contribute to poor sludge settling characteristics in anaerobic treatment systems. Proper environmental conditions must be maintained in an anaerobic digester in order to create conditions conducive to both good treatment removal performance and good sludge settling. Equipment problems can also contribute to poor sludge settling characteristics since the solids separation process is simply a physical

1

means of removing the sludge from the water. Inadequate degasification prior to the solids separation step, high biological activity in the Lamella, too long or too short detention times in the Lamella, etc. can all contribute to excessive solids loss in the effluent. As a result of the diversity of the nature of environmental factors that could possibly be creating a negative influence on the successful operation of the anaerobic digester, the decision was made to operate a number of pilot benchscale anaerobic digesters under controlled environmental conditions to determine problem areas and define proper environmental conditions for successful, reliable operations. A pilot study program was developed in conjunction with monitoring of the full-scale system in order to determine the impacts of environmental factors on the performance of the biological system and sludge settleability. The specific environmental factors that were investigated under controlled conditions in the pilot reactors were as follows: • Caustic addition and pH control • Organic (COD) loading rate • Addition of micronutrients nickel (Ni), cobalt (Co), copper (Cu), and molybdenum (Mo) • Addition of cultured bacteria • Addition of ultrafiltration permeate as primary feed stock in lieu of whey • Increased washwater flow rate through the treatment facilities to simulate conditions of total flow A control pilot digester was also operated in a manner representative of the full-scale digester for direct comparisons with the test reactors and the full-scale system. The specifics of the test program, test results, conclusions, and recommendations developed for the investigation are presented in the remainder of this paper. INVESTIGATIVE PROGRAM Four anaerobic pilot-scale reactors were initially used in the treatability studies. These systems were complete-mix, continuous flow, anaerobic suspended growth activated sludge systems. They were constructed of Plexiglas with anaerobic reactor volume of 7.5 liters and settling compartment volumes of 3.5 liters. The reactors were seeded with sludge from the full-scale digester. A wastewater collection and compositing program was developed for feeding a representative combination of whey and washwater to the pilot reactors. This combined composite sample was then pumped in the pilot reactors under controlled conditions to provide the desired hydraulic retention times and organic loading rates to the reactors. The investigative program was segregated into two time periods or phases. Each phase consisted of a three week time period for a total of six weeks on-site operations of the pilot equipment. The first week of each phase consisted of acclimation and stabilization to the environmental conditions to be evaluated and the next two weeks then consisted of an extensive data collection program. All the pilot reactors were operated in parallel utilizing one reactor as a control unit simulating full-scale operations. The remaining pilot reactors were subjected to a variety of different environmental conditions. The environmental factors evaluated during Phases I and II are summarized in Table 1 along with their appropriate reactor number designations, condition variable and objective. TABLE 1: Treatability Study Schedule – Phases I and II PHASE I Reactor Number

Condition Variable

Objective

1

Control Reactor (No Variable)

Simulate Full-Scale Reactor except at pH = 6.8

2

COD Loading

50% of Current Loading. Determine how loading effects removals and setlleability

3

Micronutrients

Determine impact of addition of Ni, Co, Cu, and Mo

4

pH = 7.2

Determine impact of caustic addition and high pH

5

Full-Scale Unit (No Variable)

Compare against treatability control PHASE II

1

Microorganisms

Determine impact of addition of cultured bacteria

2

Feedstock

Determine impact of ultrafiltration permeate addition in lieu of whey

3

Micronutrients

Determine impact of addition of Ni, Co, Cu, and Mo

4

COD Loading

Run unit at full-scale design flow and loading

5

Full-Scale Unit (No Variable)

Compare against treatability control

6

Control Reactor (No Variable)

Simulate Full-Scale Reactor with pH = 6.8

2

TEST RESULTS Phase I Program A summary of the Phase I test results is presented in Table 2 for each of the pilot test reactors. The average influent flow and COD loading characteristics in terms of both concentration and food/microorganism (F/M) ratio as lbs COD/lb mixed liquor volatile suspended solids (COD/MLVSS) are presented at the top of Table 2. Mixed liquor operating conditions such as pH, temperature, MLSS, and MLVSS; effluent characteristics and treatment efficiency in terms of COD; and gas quality and production rates are also presented in Table 2. TABLE 2: Summary of Phase I Test Results Reactor Number 1 Control

2 One-Half Load

3 MicroNutrient

0.515 0.515 36,750 0.575

0.515 0.515 18,375 0.297

0.525 0.525 36,750 0.569

0.525 0.525 36,750 0.572

6.7-7.2 35-39 12,740 8,946 0.063

6.8-7.4 35-40 11,787 8,757 0.069

6.7-7.2 36-40 13,266 9,285 0.050

7.1-7.4 36-41 13,105 9,181 0.051

EFFLUENT VFA, mg/L as CH3COOH Alkalinity, mg/L as CaCO3 VFA/Alkalinity TSS, mg/L sCOD, mg/L % COD Removed

527 2,219 0.24 2,361 1,696 95.4

392 1,950 0.20 956 1,086 94.1

425 2,213 0.19 1,649 1,711 95.3

468 3,375 0.14 1,707 1,750 95.2

GAS Gas Prod., L/day % CO2 % CH4 3 ft CH4/lb COD Removed

12.0 44.0 54.0 2.9

6.0 41.0 57.0 3.1

21.0 43.0 55.0 5.0

16.0 42.0 56.0 3.0

Parameter INFLUENT Whey Flow Rate, L/day Washwater Flow Rate, L/day COD, mg/L F/M, lbs COD/lb MLVSS MIXED LIQUOR PH Temp., C MLSS, mg/L MLVSS, mg/L Observed Yield, lbs SS/lb COD Removed

4 Higher pH

The gas quality was about the same in all the reactors; however, there were significant differences in gas production rates with the different environmental conditions. The reactors with micronutrient additions and higher mixed liquor pH consistently produced more methane gas per pound of COD removed. At the same loading rates and treatment efficiencies, the methane gas production rate was greater with micronutrient additions, followed by the higher mixed liquor pH reactor (pH>7.0) and the lowest gas production was at lower mixed liquor pH (pH 6.7 to 7.2 range). Phase II Program A summary of the Phase II test results is presented in Table 3 for each of the pilot test reactors. The average test results for influent, effluent, mixed liquor, and gas production are presented in a similar manner, for ease of direct comparisons, as Table 2 for the Phase I test results. TABLA 3: Summary of Phase II Test Results Reactor Number Parameter INFLUENT Whey Flow Rate, L/day Washwater Flow Rate, L/day COD, mg/L F/M, lbs COD/lb MLVSS MIXED LIQUOR PH Temp., C MLSS, mg/L MLVSS, mg/L

1 Cul. Bact.

2 U.F. Perm.

0.460 0.460 39,654 0.470

0.445 0.445 34,500 0.421

0.455 0.455 39,654 0.467

0.333 2.327 13,442 0.565

0.440 0.440 39,654 0.466

6.7-7.1 34-37 15,496 10,507

6.6-7.0 34-37 13,682 9,664

6.7-7.1 34-37 15,382 10,454

6.8-7.0 34-37 11,764 8,289

6.9-7.2 34-38 13,585 9,892

3 MicroNutrient

4 Full Flow

5 Control

3

0.106

0.082

0.086

0.125

0.053

EFFLUENT VFA, mg/L as CH3COOH Alkalinity, mg/L as CaCO3 VFA/Alkalinity TSS, mg/L SCOD, mg/L % COD Removed

Observed Yield, lbs SS/lb COD Removed

882 2,369 0.37 2,450 1,827 95.0

621 1,925 0.32 1,623 1,331 96.1

880 2,500 0.35 2,226 2,108 94.3

239 1,956 0.12 1,446 623 95.4

344 2,146 0.16 1,443 1,992 94.9

GAS Gas Prod., L/day % CO2 % CH4 3 ft CH4/lb COD Removed

12.0 42.0 56.0 3.1

10.0 42.0 56.0 3.0

20.0 39.0 59.0 5.5

12(19) * 36.0 60.0 3.5(5.8) *

10.0 40.0 58.0 2.0

* Numbers in parenthesis represent last three days of micronutrient addition study

Similar observations in comparing the test results from each of the reactors during the Phase II study program can be made as observed during the Phase I program relative to effluent quality and gas production. For example, the micronutrient reactor consistently produced more methane gas per pound of COD removed. The effluent sCOD, VFAs, and TSS concentrations in reactor number four were significantly lower than all the other reactors. At first glance one would assume that this reactor was significantly out performing the other reactors relative to treatment performance. However, when one compares the influent flow rate and COD concentrations, it is readily apparent that this system was operating at similar organic loading rates. The treatment efficiency was about the same as that achieved in the other reactors at similar loading rates. In other words at the same mass (pounds) COD loading rates, the mass (pounds) of effluent COD were the same. At higher flow rates and lower COD concentrations in the influent, lower COD concentrations will be observed in the effluent, but the same pounds of COD will be discharged in the effluent. METHANE GAS PRODUCTION The impacts on methane gas production of both loading rates and addition of micronutrients to the test reactors are presented graphically in Figure 1. In the top half of Figure 1 the total methane gas production from the control, the low load reactor, and the micronutrient reactors from the Phase I studies are presented for direct comparisons.

Figure 1: Comparison of Gas Production Rates The actual methane gas production rates in the control and low load reactor were almost identical at 2.9 and 3.1 ft3 CH4/lb COD removed, respectively. However, in the micronutrient reactor was 5.0 ft3 /CH4/lb COD removed. The methane production per pound of COD removed in the control and low load reactor was only around 60% of that produced in the micronutrient reactor. In the bottom half of Figure 1 the total methane gas production from the control, the full flow and load,

4

and the micronutrient reactors from the Phase II studies are presented for direct comparisons. Again, the micronutrient reactor produced significantly more methane gas at 5.5 ft3 CH4/lb COD removed when compared with the other reactors at around 3.0 ft3 CH4/lb COD removed. All the reactors without micronutrients addition only produced around 60% of the methane gas produced by the reactor receiving the micronutrients. Micronutrients of nickel, cobalt, copper, and molybdenum were added at a dosage of 0.1 mg/L of each based on the forward flow rate fed to the reactor. Micronutrients at 0.1 mg/L were also added to the full flow and load reactor on day 10 of the time period represented in Figure 1. The impact of this was an immediate increase in the gas production rate, as observed in the bottom part of Figure 1. By the last three days of this time period the gas production rate in this reactor (full flow and load) was about the same as that observed in the micronutrient reactor at 5.8 ft3 CH4/lb COD removed compared to 5.5 ft3 CH4/lb COD removed. The gas quality (%methane) was approximately the same in all the reactors during both Phases I and II. The methane content of the gas produced in all the reactors averaged around 55 to 60% methane. The micronutrients additions significantly increased the total gas production rate, but the % methane remained about the same. MICRONUTRIENT STUDIES As previously indicated the significant impact of adding nickel, cobalt, copper, and molybdenum at 0.1 mg/L each, based on forward flow, was enhanced methane gas production rate. Zinc was found to be present in sufficient amounts in the wastewater, and therefore, zinc was not added to the wastewater feeds during these pilot studies. A summary of the metals analysis of the raw wastewater and the feed to Reactor 3 after micronutrient additions is presented in Table 4. The concentrations of the heavy metals observed in the raw wastewater prior to micronutrient addition should not create toxic conditions under normal operating conditions in the reactors. Addition of these micronutrients had no apparent impact on either biological kinetics or sludge settling characteristics. TABLE 4: Influent Metals in Raw Wastewater and Reactor 3 1*

Raw Wastewater 2*

Avg.

1*

2*

Feed to Reactor 3**

4**

Avg.

Co

0.01

0.12

0.065

0.09

0.20

0.06

0.09

Cu

0.02

0.01

0.015

0.04

0.09

0.09

0.10

0.08

Fe

0.75

2.80

1.77

0.71

2.71

---

---

1.71

Mo

0.01