The Research And Application Of Industrial Circulating Cooling Water Of Narx Intelligent Decision Support System Based On Artificial Neural Network

Water is a kind of basic natural resources, which is indispensable to develop the national economy. Human demand for water has far exceeded the degree of water that can load, so pay attention to cherish, conservation of water resources is imminent.With the development of national economy and the increasing of industrialization significantly, the industrial water consumption is rising sharply. The most effective way to save water in industrial water is circulating cooling water. However, most of the circulating cooling water can be only used with increasing pollution and filling water. Sometimes, water and potion will be still wasted by the adoption of increasing coagulant dosage. Therefore it has an important theoretical significance and economic value to develop a set of industrial circulating cooling water intelligent aided analysis platform.First of all,corrosion and scaling are the common faults of the circulating cooling water system. Circulating cooling water quality continues to deteriorate, easily lead to the formation of scaling and corrosion, thereby affecting the safety of production of the whole system. Although there have been proposed water quality prediction model, but the model which can be applied to industrial cooling water quality prediction is still rare, the reason of formation process of corrosion and scaling is too complicated.The paper regards Circulating cooling water system of China petroleum and chemical co., LTD. Tianjin branch as the research object, taking into account the impact of changes in water quality and interference on the basis of random factors, starting from the continuity of water quality, water quality indicators in historical time series data as a sequence of random variables, using NARX dynamic neural network to predict the rate of corrosion and adhesion of industrial circulating cooling water quality and using the simulation test of Matlab for water quality prediction model which is established, through the contrast of the actual value and predictive value, proving that the model has better prediction performance.Secondly,because the design experience of the operator plays a key role in the traditional way of manual dosing, the water quality of circulating cooling water system is related to the change of process parameters. however the medicine added the system doesn’t change, so it cannot guarantee drug dosing reasonable and accurate, so in this paper, using the method of VB and Matlab mixed programming, combining with the water quality prediction model and neural network expert system, developing the intelligent decision system of industrial circulating cooling water. This system has realized the collection, reading and analysis of water quality data, reducing the frequency of artificial detection of water quality, Implementing on the prediction of the corrosion rate and adhesion rate of water quality prediction, and giving different experts’ opinions for different forecast results, not only meeting the requirements of "prevention first" of circulating cooling water treatment, and reaching the maximum use of drugs, making the operation of industrial circulating cooling water system stable and reliable.