Forecasting Wastewater Characteristics for Effluents from Complex Biopharmaceutical Processes: An Innovative Custom Solution using Open Source Probability and Statistics Tools

摘要

This paper describes the process developed for a predictive model and simulation of complex biopharmaceutical manufacturing processes used to produce monoclonal antibody (mAb) type therapeutic proteins (drug substance AKA "DS"). The focus of this work is forecasting both the untreated wastewater characteristics, compliance risks, resulting treatment requirements, and the final treated wastewater characteristics for the effluent that is generated by the manufacturing process. The model development involves using the bill of materials (BOM) for each DS, creating mass balances around the BOM and step by step manufacturing processes, and converting raw materials to pollutant equivalents (e.g. Perfusion media components to ThOD→COD→BOD). Pollutant masses are then fractionated into components (soluble vs particulate, biodegradable vs non-biodegradable, etc.). Since multiple processes can occur simultaneously or staggered in parallel, the production schedules are evaluated, probabilities are assigned, and the data is run through a Monte Carlo Simulation (MCS) software package. Compliance risks are then assessed, and treatment, whether using existing or recommended technologies, is then evaluated using a standard Activated Sludge Model (ASM) with kinetic and stoichiometric parameters adjusted for biopharmaceutical wastewater characteristics.