Ph Process Modeling And System Design

Abstract Abstract This research belongs to the 揝ubject Supported by ?000 Natural Science Fund of Gansu province?and ??1999 Key Subject of Science and Technology of Electric Power Corp. of Gansu province? The problem of controlling and measuring pH梫alue is found in many chemical processing including waste water treatment, medicine making, fomentation, paper, supply water treatment and so on. It is very important to control the p1-1 value within a certain range in these processes. For example, in order to protect surroundings, the pH value of the waste water must be limited strictly within a certain range, the pH value of water supply for industry must meet some demand for avoiding metal corrosion, and so on. The general method of pH measuring, pH controlling and pH modeling are discussed in this paper by means of studying of the ammonia adding of supply water to the boiler of electrical plant. The method of modeling pH process from a chemical point of view and neural network way is emphasized. pH processes have inherently nonlinearly characteristics because of their logarithm measuring relationship with the concentration of hydrogen ion. The literatures on pH modeling in home and out home shows misunderstanding in which the pH model is simply supposed to a one梠rder linear model. This paper studies the typical chemical models of strong base梥trong acid system, strong base? monoprotic weak acid system, strong base梬eak diacidic system and the pH process of water supply for boiler of electrical plant, and draw a conclusion of a general formulation of chemical model of p1-I. But there are many uncertain factors in the real system, such as the change of buffering, the effect of temperature, various of disturbances and so on; moreover, the pH process is highly nonlinear result the chemical model of pH process often failure in the real system. Furthermore, many p1-I process models cannot get only by chemical analysis way because the process involved in too many factors. So it is necessary to find another nonlinear modeling method which not only reflects the characteristics of the nonlinearity, but also has a good generalization capacity and auto梥tudy and adaptive ability on line. It is well known that the three條ayered HP neural network can approximate any nonlinear continuous function arbitrarily well, and it has the generalization ability. So it is used in this paper to model the nonlinear characteristics of p1-1 process of strong base梥trong acid system and water supply for the boiler of electrical plant. The chemical model verifies the state and dynamic performance of the neural network model of pH process. The result of the simulation demonstrates that the three條ayered neural network has a good ability to model the inherent nonlinear process of pH. In addition, the convergence and overfitting problem is discussed simply. Moreover, the modified PID control algorithm and Three梐rea梟onlinear control algorithm is combined to a new algorithm called three梐rea梟onlinear 桰i Abstract PID algorithm in this paper, and used to the pH control system of the water supply for the boiler of electrical plant. Finally, the software and hardware?s designs of a real pH control system of ammonia adding to water supply for boiler of elec...