Forecasting COD load in municipal sewage based on ARMA and VAR algorithms

Abstract

Due to different sources and the water using habits, the influent COD of municipal sewage fluctuates sharply over time. To ensure the treatment quality of sewage, the wastewater treatment plants (WWTP) often over-aerate the air and over-add the chemicals. This results in a waste of energy consumption and increases the operation cost for WWTP. With the rapid expansion of industrialization and urbanization, the quantity of municipal sewage has increased by years. Energy saving and sustainable water management for municipal WWTP are becoming an urgent issue that needs to be solved. This paper proposes a COD load forecasting model for municipal WWTP using hybrid artificial intelligence algorithms. The auto-regressive moving average (ARMA) algorithm is used for sewage inflow forecasting, and the vector auto-regression (VAR) algorithm is used for COD forecasting. The real-time data from a municipal WWTP is collected for model verification. Besides the proposed ARMA + VAR model, the BPNN, LSSVM, and GA-BPNN based COD load forecasting models are also studied as the contrasting cases. The accuracy of the forecasting performance of the ARMA + VAR model is as high as nearly 99%, which reveals its superior to the other forecasting models for future application in the wastewater treatment plants.