With the continuous progress of science and technology and the intensification of the pace of national industrialization, the demand for energy is also increasing, and the introduction of temperature control technology can realize the reasonable control of energy under the premise of meeting the needs of production or construction. This paper analyzes the smelting process of LF refining furnace and establishes the temperature forecasting model and alloy charging model. Then the differential variational operator and immune cloning operator are introduced on the basis of artificial bee colony to propose an improved artificial bee colony algorithm, and artificial neural network is used to establish an intelligent model for steel temperature forecasting, followed by simulation experiments. The experimental results show that the steel temperature forecasting model in this paper has higher forecasting accuracy and stronger generalization ability than the traditional mechanism model and neural network model for steel temperature forecasting. Pareto’s law is applied to determine the main factors affecting energy consumption, and the main factor is smelting power in the electric furnace process. The model was put into use with an average power saving of about 10.8kW-h per ton of steel, which reduced the production cost.
