Research on big data analysis model of multi energy power generation considering pollutant emission—Empirical analysis from Shanxi Province.

With the development of the integrated energy Internet, energy structure optimization and emission reduction have led to higher requirements for developing various energy sources to enable coordinated and sustainable development. However, data-mining methods are rarely used to study the coordination of multi-energy generation in published research results. In this study, from the perspective of power industry emissions, coordinated generation of various energy sources, and balance of power generation and consumption, a data-mining algorithm was used to analyze the development of thermal power, hydropower, wind power, waste heat, gas, and other power sources. The chi-square automatic interaction detection tree (CHAID), logistic regression, and two-step clustering methods were applied. The results show that: a) CO 2 and SO 2 emissions were mainly affected by thermal power generation, whereas NO x emissions were jointly affected by thermal power, garbage power, and gas-fired power, and the emissions of various pollutants increased with an increase in power consumption. The optimal power-generation scheme under minimum emission can be obtained. b) There was a strong correlation between thermal power generation and residential electricity consumption, and renewable energy (wind energy, photovoltaic, hydropower) exhibited the highest correlation with the electricity consumption of the tertiary industry, which indicates that renewable energy generation can be promoted by managing electricity consumption in the tertiary industry. c) When the electricity demand of all users was small, the proportion of renewable energy power generation increased; in contrast, the thermal power generation was larger. This indicates the importance of improving the sustainable and stable power supply of renewable energy. This study provides a data analysis model for the coordinated development of multiple energies, which will contribute to the decision-making basis for controlling power emissions, improving the utilization rate of renewable energy, and optimizing the energy structure. [Display omitted] • Rich data (8760 h): multi-energy power generation, consumption and emission. • Introduces a complete big data analysis framework to study multi-energy generation. • Three data mining methods are used to analyze power supply, demand and emission. • The conclusions of this paper have a positive effect on energy structure. [ABSTRACT FROM AUTHOR]