Mitigating the Growing Risks of Fenceline Monitoring Chemical Plants

Discharge of total dissolved solids (TDS) represents a limiting factor for the wastewater treatment plant (WWTP) at Michigan Operations in Midland as permits regulate it. These compliance requirements involve mass loading and concentration balance downstream in the river. Whereas the first is easier to control and predict, the latter heavily depends on uncontrolled conditions, such as weather and background river TDS. Additionally, the TDS concentration significantly affects the compliance requirements for Whole Effluent Toxicity (WET). Intuitively, this aspect substantially affects the normal operation of those production processes that discharge wastewater with high TDS concentration.

The control strategy consists of utilizing a pond for equalizing and optimizing the discharge flow as the river and weather conditions allow. Furthermore, coordination with sender plants is critical to ensure appropriate management of incoming streams to the WWTP. and involves coordination with the sender plants. In this context, the uncertainty of the uncontrollable variables poses a severe limitation and quantifying the risk of a compliance violation can improve decision-making on the control strategy and production coordination. The risk of exceeding discharge permits can be quantified using stochastic simulations. Stochastic simulations allow recreating all the possible outcomes that drive the control strategy decision by propagating the uncertainty of the input variables via deterministic models to obtain a probability distribution of the output of interest.

The analysis of the extensive historical records of production, weather, and river conditions can help identify the variables that exert a strong impact on the TDS discharge, understand daily fluctuations, and evaluate the influence of environmental conditions. Moreover, data analysis allows randomly generating inputs from well-defined probability distributions over the domain of interest, considering seasonality effects, correlations, and time-series dependencies. This study is also helpful for comparing and validating the deterministic models shaping the relationship between inputs and outputs. This setup allows modeling interdependent and non-linear relationships governed by mass balances.

The models were implemented in an interactive and friendly user-interface web application built in RShiny, allowing WWTP personnel to examine multiple scenarios and quantify their associated risks and it provides a complete solution for long-term and short-term risk assessment.