| The cement calcining system is one of the key subsystems of cement production.With the development of industrial technology,the equipment technology of the cement calcining system has been significantly improved.However,the popularization of intelligent equipment has not brought about the improvement of supporting decision-making methods.As a result,my country’s cement manufacturing industry still faces problems such as low resource utilization,poor product quality,and large waste discharge.The main reasons include:1.The cement industry has the characteristics of multiple equipment and complex production mechanism,which makes it difficult to establish an effective mechanism model;2.The cement industry data is multi-interference,uncertain and multiple information,and the dynamic changes of working conditions make it difficult for process engineers to perceive the dynamic changing working conditions.Therefore,it is of great significance to study the intelligent optimization control of cement calcining system to improve the quality of cement products and reduce production energy consumption.Aiming at the above problems,this paper carries out the research on the multi-objective optimization model and dynamic decision-making algorithm of the cement calcining system.(1)Aiming at the time-varying time delay,uncertainty and nonlinearity of the cement industry production process data,a time series-based convolutional neural network model is proposed.First,study the correlation between production process variables and production indicators on the time scale,construct a time series input layer,and solve the problem of time-varying delays that are difficult to determine.Then,the multi-layer convolution-pooling layer is used to extract the high-dimensional data features of the time series input layer to eliminate data redundancy,and the fully connected layer is used for feature integration.Finally,the adaptive moment estimation algorithm is used to continuously correct the prediction error to solve Uncertainty and non-linear problems of production process data realize the prediction of multiple production indexes of cement calcining system.(2)Aiming at the related and cooperative optimization of production indicators such as electricity consumption,coal consumption and quality of the cement calcining system,as well as the subjectivity and adjustment lag of process engineers,a multi-objective optimization model of the cement calcining system is established.The model takes the energy consumption of the cement calcining system as the optimization target,and comprehensively considers factors such as quality production indicators,equipment capabilities,raw material conditions,and variable constraints.A constrained differential evolution algorithm is proposed.By solving the optimization objective function,it reveals the inherent law between the production controlled variable and the production index,and obtains the initial value of the controlled variable to achieve the purpose of reducing energy consumption and improving product quality.(3)Aiming at the open-loop solution to the problem that it is difficult to track the dynamic change of the cement calcining system,based on the open-loop decision of the controlled variable setting value,considering the industrial dynamic characteristics,a dual-period coupling-based energy consumption prediction and correction algorithm is proposed.And design the optimized performance index based on the energy consumption optimization model and the reference trajectory.By solving the performance index function,the set value of the controlled variable that meets the change of the working condition is obtained.Experimental results prove the effectiveness of the prediction correction algorithm. |
