Application of DEA and statistical inference to model the determinants of biomethane production efficiency: A case study in south China

- Investigated the determinants of efficiency in biomethane co-digestion pilot project in South China.
- Methodology involved application of data envelopment analysis and multiple linear regression.
- Variables affecting output fluctuation were found and determinants of efficiency analyzed.
- The pilot project was found to suffer from scale inefficiencies, implying sub-optimal operation.
- Based on the findings, engineering management and policy recommendations were provided.

Global interest in the conversion of organic biowaste to biomethane is increasing rapidly. As new projects are built, managers must ensure that biomethane engineering processes are operating efficiently. In China, increasing biogas energy output is an integral part of the central government's 13th Five Year Plan. However, many biogas plants that convert various organic waste types to energy in China operate inefficiently. In this context, the objective of this research is to investigate the determinants of efficiency in a major biogas engineering pilot project that converts municipal, industrial and agricultural waste into biomethane vehicle fuel. The methodology involved applying data envelopment analysis and multiple linear regression to determine factors statistically significant for fluctuations in performance efficiency. The results provided important insights. First, variables statistically significant for the production of primary outputs were isolated. The variables most influential for biomethane production included bagasse input, fish waste input, and cassava input. As for solid fertilizer output, the most influential variables included manure input, fish waste, other input, and FeCl2. Secondly, the surveyed case was found to have significant scale inefficiencies, which has important implications for optimization of industrial scale co-digestion projects. As time progressed, the project experienced decreasing returns to scale, indicating that although overall inputs increased, production per unit of input decreased. Third, specific input/output targets and slacks were computed in order to identify changes required for the project to become efficient at certain points in time over the survey period. Fourth, possible determinants of efficiency were analyzed. The paper concludes with several engineering management and policy suggestions to enhance biomethane conversion efficiency.