Research On Energy Saving Of The Aluminum Industrial Production Based On Technologies Of Pattern Recognition And Data Mining

Process industry is one of the pillar industries of national economy, and it is widespread in metallurgical industry, pharmaceutical industry, chemical industry and some other industries. So this industry is very important. Some industries of the key areas are of vital importance to the national economy and people's livelihood. It is significant to keep these industries have good fast development for protecting national security, maintaining national interests and enhancing the national economic competitiveness by improving the process industry systems with advanced manufacturing, control and management theories and technologies. With the rapid development of the global economy, the contradiction between energy supply and demand has become increasingly prominent, and saving energy has become the subject of all mankind, is a long-term strategic approach for each country to develop the economy. Especially in the metallurgical industry with huge energy consumption, it is of great practical significance to research how to reduce energy consumption in production. This not only can reduce production costs, enhance the competitiveness and the ability to adapt to the market of the enterprise, but also is of great strategic significance to ease the tension of energy, to solve the bottleneck problem of sustainable development and to promote the development of Chinese economy and society [1].Take the aluminum industrial production for example, pattern recognition method can be used to research the energy consumption problem of the aluminum production process. With implement of data mining, the implicit rules, and factors affecting energy consumption can be found out to avoid these factors in the production process and find a production way of energy saving. If these come true, a lot of energy for the scientific arrangements of the production will be saved. The results can direct enterprises to improve the degree of agility, reduce material and energy consumption, enhance the market competitiveness and the national interests of the enterprises in the global economy.This paper has done some research with the background of the project called the research on the methods of process industry's pattern recognition, energy consumption optimization and scheduling, which is a part of the key national natural science fund project: the research on theories and methods of the non-ferrous metallurgical process control for energy saving and consumption reducing (Project number: 60634020). This paper research on the energy saving problem of the aluminum industrial production, and the mainly works are as follows:1. An investigation for the whole process of the aluminum industrial production was done, and the energy consumption situation of the aluminum industrial production was summarized after reading a large number of papers, and the energy consumption situation in 1 ton of aluminum production was analyzed and extracted with modeling for it. Lacking in actual production data, 1000 energy data for the energy consumption status in 1 ton of aluminum production was generated by using Matlab in order to facilitate the simulation.2. An association rule mining algorithm based on correlation analysis was proposed to solve the energy saving problem in 1 ton of aluminum production in the principle of association rule mining in the data mining. The main factors affecting energy were found out in simulation. And the implicit rules were dug out, which showed that the algorithm has feasibility and effectiveness in solving the energy saving problem of the aluminum industrial production.3. The model of energy consumption problem of the aluminum industrial production was set up, and K-means algorithm was employed to solve the energy saving problem of the aluminum industrial production by using C++. It is the first time to use such pattern recognition algorithms to solve the energy consumption optimization problem of the aluminum industrial production. Through simulation, a solution to the problem was found out, which showed the algorithm was feasible.4. A novel pattern recognition algorithm– the threshold cluster analysis algorithm was studied to solve the energy consumption problem of the aluminum industrial production. Though good results were received, the choice of parameters in the algorithm with blindness. Therefore, it is necessary to improve the algorithm. By defining the parameters with the distance between the data, the improved algorithm was proposed, which can find a better solution, while the computing time is a little longer.5. Comparing with other different algorithms, the results are also variable. All the algorithms are feasible and effective in solving the energy saving problem of the aluminum industrial production. However, the appropriate algorithm should be selected in practice with their own advantages and disadvantages.At last, the works of this paper were concluded, and the prospective of future research on the energy saving of the aluminum industrial production based on pattern recognition and data mining was discussed.