Use statistical analysis to find a cure for a “sick” wastewater treatment plant.
Wastewater treatment plants (Figure 1) can function improperly for a variety of reasons: Plants sometimes receive materials they are not designed to handle; and some facilities are poorly specified, designed, constructed, operated, and/or maintained. If your plant is “almost working” or “often working,” then your plant is not working, and much effort may be necessary to move from “almost” or “often” to “actually.”
I have found that operators of poorly functioning plants are pressured to accept any effluent sent to them to maintain the capacity of the primary production process. As manufacturing facilities increase production volumes and product diversity, plant owners often fail to upgrade the effluent treatment plant to handle the changes. Even at plants that have not increased production, wastewater treatment plants are often required to handle troublesome liquids that they were not designed to process. Owners prioritize getting the main process back online as quickly as possible, rather than accepting the delay and expense of proper disposal.
I have seen the entire contents of hypochlorite scrubbers at a coking works dumped directly into a biological effluent treatment plant by night shift crews, which killed all of the biological life. I also know of a facility that dumped tons of out-of-spec concentrated detergent directly into the drains of the oil-water separators, totally and irreversibly destroying the effectiveness of downstream granular activated-carbon filters. Both of these examples cost tens of thousands of dollars and took several weeks to fix, only to save operations a few minutes and a few hundred dollars.
Sick process syndrome
Throughout my career conducting several hundred waste minimization analyses at manufacturing facilities, I saw many processes that, at best, could be described as “barely in control.” I dubbed these sick processes, and they can be recognized by the following characteristics:
- high and variable rates of process failure
- staff constantly searching for technical fixes
- control taken away from operators
- many theories about the cause of the problem
- denial of problems
- a history of failed attempts to improve control
- staff considering the process an art more than a science.
The way to cure the sick process is to get past denial and folk wisdom, and instead use statistical analysis to accurately characterize the problem and verify that the attempted cure worked.
When you begin a statistical analysis of a wastewater treatment plant, typically very little concrete information is available. In some cases, though, there is so much data that personnel feel paralyzed by the information overload.
Too little data. It is preferable to start with at least 20 analysis results of various quality parameters. However, usually fewer than ten are available and plants are reluctant to pay for additional testing.
A standard set of analyses of incoming and out-going effluent includes chemical oxygen demand (COD), biochemical oxygen demand (BOD), total suspended solids (TSS), ammonium (NH4), pH, and temperature, as well as data on flowrates. If these data are not available, or if very little data is available, use a site test kit (
Would you like to reuse content from CEP Magazine? It’s easy to request permission to reuse content. Simply click here to connect instantly to licensing services, where you can choose from a list of options regarding how you would like to reuse the desired content and complete the transaction.