Data-driven modelling for resource recovery: Data volume, variability, and visualisation for an industrial bioprocess

Advances in industrial digital technologies have led to an increasing volume of data generated from bioprocesses, which can be utilised within data-driven models (DDM). However, and variability complications make developing that captures the underlying biological nature bioprocesses challenging. In this study, a framework for is proposed evaluated by modelling bioprocess, treats or agrifood wastewaters whilst simultaneously generating bioenergy. Six were developed predict reduction chemical oxygen demand wastewater bioprocess statistically using both testing (randomly partitioned model development) unseen (new not used during development). The statistical error metrics employed coefficient determination (R2), root mean square (RMSE), absolute (MAE) percentage (MAPE). stacked neural network was best able having highest accuracy on (R2: 0.98; RMSE: 1.29; MAE: 2.27; MAPE: 4.08) 0.82; 2.57; 1.75; 3.68). Data visualisation observe (or confirm) whether new points are boundaries, helping increase confidence model’s predictions future data.