Online measurement of alkalinity in anaerobic co-digestion using linear regression method

Abstract

Alkalinity is a reliable indicator of process stability in anaerobic digestion system. Total alkalinity (TA) and partial alkalinity (PA) are usually monitored offline as indicators for the status of anaerobic digestion process. In order to online monitor TA and PA, the linear regression method was used as estimator to predict alkalinity via software sensor method. Parameters, namely, pH, oxidation and reduction potential (ORP), and electrical conductivity (EC), were used as input variables. EC was the most significant parameter with TA and PA. Multiple linear regression (MLR) models and simple linear regression models with EC were constructed to predict TA and PA in anaerobic co-digestion system. On the basis of the evaluation of prediction accuracy, the applications of linear regression models were better for monitoring PA than TA. MLR models provided higher accuracy for alkalinity prediction than simple linear regression models. The two MLR models based on single-phase anaerobic digestion system were also feasible to predict TA in anaerobic co-digestion systems. However, the accuracy of these models should be improved by calibrating for broad applications of linear regression method in online alkalinity measurement. Keywords: anaerobic digestion, alkalinity, online measurement, model, linear regression DOI: 10.3965/j.ijabe.20171001.2701 Citation: Bai X, Li Z F, Wang X M, He X, Cheng S K, Bai X F, et al. Online measurement of alkalinity in anaerobic co-digestion using linear regression method. Int J Agric & Biol Eng, 2017; 10(1): 176–183.