Multi-scale energy efficiency recognition and diagnosis scheme for ethylene production based on a hierarchical multi-indicator system

It is essential to monitor and diagnose the energy efficiency level of ethylene production process to detect any abnormities in the production and adjust production parameters appropriately to ensure the energy efficiency. However, few studies have been carried out concerning the diagnosis of abnormal energy efficiency in the ethylene production with appropriate energy efficiency indicator system from multiple dimensions due to the incompatibility of the diagnosis algorithm and the complexity of this petrochemical production. Therefore, to achieve a comprehensive energy efficiency diagnosis and anomaly analysis for the ethylene production, a multi-scale energy efficiency recognition and diagnosis scheme based on a hierarchical multi-indicator system is proposed. Firstly, a novel recognition scheme for abnormal energy efficiency is conducted from a single-indicator scale to a multi-indicator scale based on a three-level indicator system with respect to the energy and materials flows in the system, process, and equipment level of the ethylene production, so that the energy efficiency of the ethylene production can be monitored in different levels and granularities. Secondly, an integration-decomposition diagnosis method to analyze the energy efficiency is proposed based on the Malmquist produc-tion index, where the overall efficiency is decomposed into production efficiency, pure production technical efficiency, and scale efficiency to form a multi-perspective in-depth diagnosis and exploration. Thirdly, a prin-cipal component analysis is converted into a preliminary abnormal cause diagnosis algorithm based on statistics so that the multi-scale and multi-level energy efficiency recognition and diagnosis for the ethylene production is achieved. Finally, the effectiveness of the diagnosis scheme is verified by applying it to a Chinese ethylene plant, and the results show that the abnormal energy efficiency can be identified effectively and the consistency of the diagnosis results with the actual situation is approximately 0.818 at least.