Health-safety-environment performance evaluation for electric power construction projects based on a fuzzy Petri net-long short-term memory (FPN-LSTM) algorithm

Health-safety-environment (HSE) performance evaluation is a key tool to improve the comprehensive management level of electric power enterprises, and can effectively reduce the probability of engineering accidents and ensure the implementation of electric power construction. However, most existing techniques do not consider the credibility of each evaluation element, and hence have relatively low resolution. A method of HSE performance evaluation for electric power construction projects is proposed based on FPN-LSTM. Firstly, a fuzzy Petri net (FPN), which can be regarded as the HSE performance evaluation model for electric power construction projects, is established based on an on-site HSE data statistics table. Secondly, the confidence values of initial, intermediate, and termination places in the FPN model are determined based on the percentage difference of node data and linear interpolation calculation. The HSE performance evaluation of electric power construction projects can then be carried out. Thirdly, a long short-term memory (LSTM) algorithm is introduced to train and update the confidence in FPN, thereby maximizing the optimization of HSE performance evaluation results. The HSE performance evaluation of the Luxi Branch was used to verify the FPN-LSTM model and compare it with existing models. The FPN-LSTM model can accurately and systematically reflect the HSE performance level of an entire electric power construction project, as well as precisely and effectively clarify the implementation and distribution characteristics of evaluation indicators at all levels. In summary, the FPN-LSTM method offers a scientific, systematic, and precise decision-making tool for HSE managers.