Optimization Program – Microbial Performance Based Training (PBT)

The TSC optimization team, in collaboration with the West Virginia Bureau of Public Health, conducted the sixth and final session of microbial PBT on April 30, 2013 at the Summersville, WV National Guard Conference Center. WV Drinking water treatment plant operators reported on their efforts to improve filtered water quality using the techniques learned and applied throughout the training project. Participant feedback credited the training for the documented improvements in treated water quality and also for the increased levels of skill and confidence exhibited by the operators. Commitment was unanimous for a follow-up session to be held in 9 – 12 months. The purpose of this session is to evaluate the progress as operators continue to apply the skills and techniques to optimize process control at each of their treatment plants. (Ouro Koumai and Rick Lieberman US EPA, TSC)




Microbial Optimization Goals – Then and Now

By Larry DeMers and Bob Hegg – Process Applications, Inc.

The microbial optimization goals provide the basis for the national Area Wide Optimization Program.  These goals were introduced to the AWOP states at the beginning of the program and were used to encourage a different level of performance from surface water treatment plants than the regulatory requirements at the time.  However, the initial development of these goals goes back even further in time.  The first Composite Correction Program handbook published by EPA in 1991 for water treatment optimization does not specifically list performance goals.  Performance is discussed under the “Conducting Performance Assessment” section of the CPE methodology and reference is made to achieving 2 NTU from sedimentation basins and 0.1 NTU from filters.  Included in a discussion in filter performance after a backwash, acceptable performance is described as a turbidity increase of 0.2 to 0.3 NTU for less than 10 minutes after a backwash.  Moving on to the 1998 edition of the Composite Correction Handbook, Chapter 2 is devoted to protection of public health from microbial pathogens, and the research basis for the optimization goals and specific performance goals are described within the chapter.  The turbidity performance goals, which all AWOP participants are familiar with, are summarized below.

Individual Sedimentation Basin Performance

  • Settled water turbidity less than 1 NTU 95 percent of the time when annual average raw water turbidity is less than or equal to 10 NTU.  This goal increases to 2 NTU when the annual average raw water turbidity is greater than 10 NTU.

Individual Filter Effluent Performance (IFE)

  • Filtered water turbidity ≤ 0.1 NTU 95 % of the time (excluding 15 minute period following backwashes) based on maximum values recorded during 4-hour time increments.
  • Maximum filtered water turbidity of 0.3 NTU.
  • Initiate backwash after turbidity breakthrough has occurred and before turbidity exceeds 0.1 NTU.
  • Maximum filtered water turbidity following backwash of < 0.3 NTU.
  • Maximum backwash recovery period of 15 minutes (i.e., return to < 0.1 NTU).

Although specific goals were not established for combined filter effluent turbidity (CFE) in the 1998 edition of the handbook, the same goals as with IFE are applied to CFE when interpreting performance at this location.  This point is demonstrated in the handbook in the CPE methodology section on assessing plant performance.

Further discussion on interpreting the filter backwash recovery performance goals occurs in Chapter 4 under supplemental data collection.  This section states that the same goals (i.e., limit turbidity spike to < 0.3 NTU and recover to ≤ 0.1 NTU within 15 minutes) should be used to assess filters with filter-to-waste capability.  The 15 minute recovery period starts when the filter begins filtering during the filter-to-waste period.  The rationale for this approach is that the filter-to-waste period is a key indicator of a plant’s process control, and monitoring of performance should start immediately after the filter is placed back in service.  This description is important to remember, since NOLT has made changes to this interpretation that are currently being applied during Targeted Performance Improvement (TPI) activities (e.g., CPEs, PBT), and they will be discussed later in this article.

Since the last update of the turbidity performance goals in 1998, further research using higher resolution turbidimeters and particle counters have confirmed the validity of these goals and provided the basis for refinements.  One specific refinement is the use of two significant figures when referring to the filtration goals (i.e., 0.10 in place of 0.1 NTU, 0.30 in place of 0.3 NTU).  One specific research project that is included in the “Why Optimize” presentation during PBT references work conducted by Emelko (Water Quality Technology Conference, 2000) on Cryposporidium removal during filtration.  The researchers were able to demonstrate 5 to 6 log removal during optimized, stable filter operation, and the measured turbidity during this period of operation was approximately 0.04 NTU.  At the end of the filter run the log removal decreased to 2 to 3 log with a corresponding turbidity increase to approximately 0.10 NTU.  Advances in turbidimeter resolution also support the refinement in turbidity readings.  A common low range process turbidimeter currently used in water treatment plants (i.e., Hach 1720E) has an accuracy of ± 2 % of the reading or ± 0.015 NTU from 0 to 40 NTU.

A recent refinement to the sedimentation and filtration optimization goals was reported in the August 2009 edition of AWOP News.  This article described a recommended approach for establishing the frequency of data collection for continuous reading turbidimeters when pursuing process optimization.  For sedimentation basins a frequency of at least 15 minutes is recommended.  For individual filters and combined filter effluent at least a 1 minute frequency is recommended.  Additional information on this refinement can be found in the article.

For the last microbial goal refinement discussed in this article, revisions to the filter-to-waste performance goal will be reviewed.  The performance goals for plants without filter-to-waste capability remain the same, other than the change to using two significant figures.  These plants should strive to limit their turbidity spike following backwash to < 0.30 NTU and should achieve ≤ 0.10 NTU within 15 minutes of return to service.  Recent experience from PBT plants implementing special studies has shown that filter backwash spikes can be reduced substantially through use of practices such as filter rest periods and the extended terminal subfluidization wash (ETSW).  In many cases filter spikes can be reduced to < 0.10 NTU.  For filters with filter-to-waste capability the performance goals have changed to the following:

  • Minimize the turbidity spike during the filter-to-waste period (i.e., record the highest turbidity and direct optimization efforts at minimizing this value).
  • Return the filter to service at ≤ 0.10 NTU.

This refinement does not establish a maximum turbidity value or length of time to the filter-to-waste period.  The primary reason for this change is that, under the revised goal, filters are not returned to service until the turbidity is ≤ 0.10 NTU thus limiting the need to establish specific maximum turbidity and duration goals for water that is going to waste.  The revised goal recommends that plant operators monitor performance during filter-to-waste and minimize the turbidity spike during this period, a practice that has not been pursued by plants in the past but is a key activity during PBT.  The magnitude of the spike during the filter-to-waste period can be used as a relative indicator of filter conditioning prior to the filter going into service.  Most plant operators have a desire to keep filter-to-waste periods short, since they want to minimize wasting production water.  Consequently, operators are motivated to achieve ≤ 0.10 NTU as quickly as possible during filter-to-waste.

An example post filter backwash assessment from a recent Oregon PBT session is shown in the chart below.  These turbidity data describe the performance of a filter during the filter-to-waste period.  During the initial part of the period the turbidity reflects the quality of the backwash water exiting the filter.  Once this water is removed, settled water passes through the filter, and the maximum turbidity of 0.13 NTU occurs at about 12 minutes into the period.  A turbidity of 0.10 NTU is reached between 17 to 18 minutes.  At this time the plant operator changed from filter-to-waste to filter-to-clearwell operation.

It is important to understand that the initial lag in the turbidity response can be impacted by the size of the filter, including the underdrain volume, as well as the filter-to-waste rate.  Ideally, the filter-to-waste rate is similar to the filtration rate; however, this is not always the case for plants with different size waste piping.  It is also important to understand that the potential exists to minimize the turbidity spike and duration of the filter-to-waste period described by the performance in the chart.  As described previously, many PBT operators have been able to utilize the special study approach to maintain the turbidity spike during filter-to-waste at ≤ 0.10 NTU.

Almost 20 years after the publication of the Composite Correction Program handbook for surface water treatment plants, the basis for the turbidity goals remains in place.  The refinements described in this article have been made based on industry research, changes in instrumentation capability, and the considerable amount of experience gained through implementing AWOP activities.

Update on Status of Microbial Performance Based Training

By Bob A. Hegg and Larry D. DeMers – Process Applications, Inc.  

BACKGROUND:  The Microbial Performance Based Training approach is a key Targeted Performance Improvement tool used by the AWOP network to allow operators and managers at water utilities to achieve the full capability and performance potential of their existing facilities.   This protocol, originally piloted in 1999 has been demonstrated on a broad scale (an estimated 22 states and 200 water utilities) and has documented performance data available as well as documented state and operator skill improvement available to show overall impact.  From a microbial perspective operational skills for participating facilities have been enhanced to allow utilities to aggressively pursue optimized performance goals for filters of <0.10 NTU.  The protocol has impacted the way training is done for water utilities in many states.  In addition the protocol is being developed to apply to disinfection by-product and groundwater optimization efforts.  The purpose of this article is to provide an overview of the significant enhancements to the protocol that have occurred over the last ten plus years of PBT experience.  

CURRENT STATUS:  The current PBT approach has evolved and been enhanced with almost each training series that has been conducted, however, the basic protocol that was developed over 10 years ago has remained the same.  The protocol is applied at multiple treatment plants using a multi-faceted long-term training process.  Five strategic centralized training sessions are used to introduce key optimization concepts and skills to representatives from each of the participating plants and to facilitators.  The sessions are conducted over a 12 to15 month period.  The training emphasis of each session is:  Session 1 – Adopt Goals and Assess Data, Session 2 – Developing Priority Setting and Problem Solving Skills, Session 3 – Coagulation Control Tool Development, Session 4 – Assessing Current Plant Performance/Applying Skills and Tools, Session 5 – Reporting on Success.  

Between the training sessions, training facilitators work with the individual plants through limited site visits and phone contact to encourage implementation of the concepts and skills presented during the training sessions.  The impact of project activities on plant performance is measured by comparing turbidity data for the one-year periods before, during, and after the projects.    

ENHANCEMENTS:  Enhancements made to the training protocol have been based on an expanded experience base in implementing the protocol.  Some of the significant enhancements will be described in more detail in this article.  

  • Facilitator Training Session:  It was recognized after several initial PBT efforts that facilitator training was just as important during PBT as training for the actual utility personnel.  It was also recognized that facilitation skills vary widely.  Currently a formal one and ½ day facilitator training session has been developed and implemented.  This session focuses on providing an overview of PBT, teaching facilitation skills such as avoiding troubleshooting, teaching the facilitators how to obtain initial jar calibration settings for a utility, recognizing sampling challenges to supporting PBT data collection, and identifying on site activities.  Additionally this training is supplemented by phone consultation and routine facilitator/trainer conference calls throughout the PBT effort.  An associated major change has been that the jar test calibration spreadsheet is no longer taught or introduced in the training sessions.  Alternatively, facilitators are trained on using the spreadsheet, and they collect the necessary data and determine the initial settings.  The facilitators provide this information to their utility participants during or shortly after Session 3.  The utility staff is trained to “tweak” these settings using special studies to make the jar test tool specific to their plant.  As a component of the facilitator training session, a plant is visited and the initial jar calibration parameters are calculated using the jar test spreadsheet.  The facilitators that have been through this training have commented positively on the support this provides for their efforts.
  • Filter Backwash Performance Trending Spreadsheet:  This spreadsheet has been added to PBT and is used to collect data from participating PBT facilities on post filter backwash performance.  The spreadsheet was previously described in the August 2009 issue of AWOP News.  An example graph from the spreadsheet is shown in Figure 1

Figure 1: Example Spreadsheet Chart Showing Multiple Days of Backwash Recovery Data


  • Session 2 Homework:  Session 2 covers development of priority setting and problem solving skills.  The concept of special studies is introduced in this session, and the related homework assignment is to conduct two special studies.  In response to consistent feedback that the participants like visiting other facilities, assignments are now made to pair up the utilities and have them jointly conduct one of their special studies together.  For those ‘partner’ plants that have been able to complete the exchange of special study efforts the feedback has been positive.  The quality of the joint special studies has also improved.  It has been a challenge for some of the partner utilities to find the time to complete the joint efforts.
  • Session 2 and Session 3 Workshops:  The data collection for the workshops associated with Session 2 and Session 3 has been modified to be very prescriptive (e.g., data collection tables are provided in the workshops with the number and time for sampling specified).  This has allowed the data to be compiled during the workshops and graphically displayed at the end to enhance data interpretation during the training sessions, a critical skill needed for problem solving.
  • Session 3 Quality Control Special Study:  As a jar test calibration workshop activity during Session 3, some of the participants are requested to complete a quality control special study.  This study involves collecting equal amounts of raw water in two jars (2L each), adding equal amounts of coagulants at the same time and mixing simultaneously with equal amount of energy.  Once the mixing is complete, the water is drawn out at the same time from both jars at predetermined time intervals (i.e., 0, 1, 2,  4, 6, 8, and 10 min) and measured for turbidity.  Graphs (settling curves) are developed for each jar by plotting turbidity vs. time, and the results are then compared with each other.  If reproducible results cannot be achieved, then jar test techniques need to be developed further before the next steps of jar test calibration can be pursued. 

Variability in jar test results can occur for a variety of reasons such as:   

  • Inconsistency with filling up the jars to 2L mark with raw water
  • Inconsistency with dosing the exact amount of coagulant 
  • Inconsistency with opening the jar valves (taps) for sampling
  • Variability in the optical sensitivity of the round sample cells
  • Turbidity meter variability
  • Inconsistent mixing of samples before measuring turbidity

Despite these variations, proper technique should allow similar curves to be developed.  A spreadsheet-based tool has been developed to allow a comparison of settling curve results.  Two methods have been developed to determine how well settling curves match each other.  The first method assumes that the two settling curves are similar (i.e., they match) if the data fall within a range of +/- 15 percent of the average of the two curves.  Another method uses the development of an “Absolute Difference Ratio” to assess the similarity of the developed curves.  The data collected to date indi­cates that a ratio of less than 1.0 and preferably less than 0.7 indicates good quality control and settling curves that match each other.  This 15 percent difference and the ratio are automatically calcu­lated when the data is entered into the jar test calibration spreadsheet.  The objective is to work on jar testing technique until settling curves that are similar to each other can be developed (i.e., meet the +/- 15 percent criteria or meet the < 0.7 ratio).  Figure 2 shows results demon­strating good jar testing tech­nique.   


FIGURE 2. Example of Settling Curves That Demonstrate



  • PBT Follow-up Meetings:  A PBT follow-up session (i.e., 1 year later) was initiated with the Idaho PBT series in 2006 and a similar session was added to the DBP PBT pilot series in South Carolina.    The sessions were added to provide motivation for the PBT graduates to continue to focus on performance data collection and special studies after Session 5.  Minimal facilitation is provided during this post PBT period.  The agenda for these sessions includes a review of the performance data and short summaries of significant special studies that have been completed.  Feedback from attendees has been positive, and the effort encourages ongoing data collection and sustaining of the operator “network “.
  • State Enhancements:  Several of the AWOP states have made other enhancements to the PBT protocol to suit specific types of plants (KY – developed materials specifically for ActiFlo Plants;  AL – utilized a multi day Session for Session 3; IA inserted a session 2A to better emphasize special studies; and PA modified Session 3 to include running zeta potential on participants water. 


The results of PBT projects indicate that it remains a viable training tool for achieving performance improvements at multiple treatment plants, including small systems.  It also results in enhanced skill development and optimization implementation potential for utility personnel as well as for personnel fulfilling the facilitator role.  After 10 years of successful implementation the PBT protocol has proven to be a successful training approach, and enhancements have been an ongoing process to make this TPI tool even more effective.