| With the increment of consciousness on economizing energy, reducing consumption and environmental protection, and the development of telecommunication and computer technology, used separate heat supplying has been replaced by central heating supply. Central heating supply system has been an important form, and acquires more and more application in the north of country. Central heat supplying can utilize the energy source effectively, increase the rate of energy, reduce environment pollution, promote production, and improve people's living. It is an effective measure and important avenue which can accelerate the city's modern foundation.According to the characteristic of big distance of heat source and heat exchange station, it is difficult to collect and transfer data. In the article, the wireless data collection and transfer systems based on GPRS are discussed in this paper. And the construction of data transfers systems, data collection unit and the function of the monitoring center are all introduced.Because of the characteristic of non-linear, time-variant, big time-delay and coupling between heat stations, many parameters change with the practical instance. It is difficult to build exact math model, and makes the use of traditional control theory restricted.Then according to the heating supply system, based on the algorithm of B-P neural network and fundamental generalized predictive control (GPC) control, an improved B-P neural network and generalized predictive implicit algorithm control are brought forward. In order to resolve coupling during heating supply system, a new intelligent predictive control algorithm is brought forward. The step is that, first the better B-P algorithm is used to eliminate coupling of system, and then by connecting with implicit GPC controller control each single loop. The algorithm neither solves recursively Diophantine equation on-line and nor calculates inverse of matrix.Finally, based on the combustion system of fluidized bed boiler, simulation of fundamental algorithm and intelligence composite algorithm was conducted by using Mat lab. The simulation results prove the validity of the algorithm. |
