Optimisation of energy usage and carbon emissions for an advancedanaerobic digester plant

Abstract

In this thesis Northumbrian Water Limited’s (NWL) Advanced Anaerobic Digester (AAD) plant at Howdon was used to investigate modelling and optimisation opportunities based on energy prices, demands and their new greenhouse gas emissions pledge. It is believed this site is the first in the UK with a mixed operational strategy for biogas and biomethane produced on site: to burn in Combined Heat and Power (CHP) engines to create electricity, burn in Steam Boilers for onsite steam use or inject the biomethane into the national grid - Natural Gas can be imported to make up shortfalls in biomethane if required. Initially, a realistic model for the gas distribution on site was developed using a novel mixed integer linear programming (MILP) approach. Retrospective Optimisation (RO) using historical plant data was performed, with results indicating the plant operated optimally within accepted tolerance 98% of the time. However, improving plant robustness (such as reducing unexpected breakdown incidents) could yield a significant increase in gas revenue of 7.8%. Next, the gas distribution model is developed further as a realistic MILP model for energy and carbon management where operators are provided with a visual daily operational schedule based on varying tariffs. The results indicate that biomethane injection should be maximised for the highest financial gain, with the driving force for optimising the remaining operations being the site electricity demand and whether the electricity purchased from the grid generates carbon emissions, based on the new carbon performance commitment. Using the developed energy and carbon model a sensitivity analysis was performed on electricity tariffs, natural gas prices, the volume of biogas production and the Biomethane Upgrade Plant (BUP) processing limits. The results reinforce the understanding that maximising biomethane injection into the national grid is the most cost-effective operational strategy. Second to this, the optimal operation of the CHP engines is subject to the available excess biogas available after BUP processing and the current daily energy prices. To ensure the site always maintains a positive revenue, operators should ensure that at least 20,000 Nm3 /day of raw biogas can be processed and injected into the national grid. Finally, an investigation into the unique modelling problem regarding the three on site Anaerobic Digesters (ADs) was performed. A key parameter used in the current optimisation model is the amount of biogas that is produced on site each day, however currently an average daily value is used based on historical data. To improve the optimisation, it would be better to provide a more accurate prediction based on current state of the ADs and the expected sludge processing volumes into the ADs. The lack of individual gas flow data for each AD posed an interesting challenge in predicting the total biogas flow produced on site. Multiple linear models of the onsite AD’s were investigated but were not accurate enough to be used on site. A NARX (Nonlinear autoregressive with external input) Neural Network was developed to model all three anaerobic digesters as a single process for the day ahead prediction of biogas production. The resulting optimal NARX model can accurately predict the biogas production on a day-ahead basis over 95% of the time.