Research On DO Control In Process Of Water Treatment

Water pollution is an urgent problem in the world at present; and people come to have an agreement to meet water demand by water treatment. The technology of activated sludge is an economic, simple and convenient method to improve water quality. To exactly control DO in the biochemical pool can get good treatment results. But the biochemical reaction process is a complex biochemical and nonlinear system, in the real water treatment, the stream is fluctuant, it is very difficult to exactly control DO. In order to resolve this problem, radial basis function neural network (RBFNN) is adopted in the design. By using this method, considering the COD value of and the quantity of the inputting water as the input signals, the DO value of the output water can be concluded exactly. Gas explosion can be controlled by modulating the roll speed of the AC motor, which is used to drive the fan, so as to achieve the invariable DO value control. The research and design is carried out according to the water processing system of DaLian Lvyuan Medicine Company. The research is as follows.First, the actual research was carried out in the water processing workshop. The analyses of the system are done. The control object and the tactics are confirmed. The dissolved oxygen in the biochemical reaction pool is decided as the controlled object. The relational datas of the former six months are used. The radial basis function neural network (RBFNN) is adopted to prognosticate the DO value of output water and the relational conclusions are validated.And then, the hardware of the system is designed. The TI Company's Digital signal processing (DSP) called TMS320LF2407 is applied. Make use of its excellent ability of controlling motor. To control the speed of the fan, so as to control DO in the process of the water processing. The power supply, expending of the memorizers, JTAG interface, the measure of the motor's speed, the real time display of the datas are designed. And the design of the system software is completed.Last, the circuit of the system hardware is put up. The modulation and the simulation of the system are carried out by using the ICETEK-5100PP simulator.Through research on the control method based on neural network, this paper attempts introducing the technology of neural network intellectualized control and the digital signal processor into waste water treatment. The results simulation shows that the system is well self-learning, adaptive and robust, and can realize accurate control on nonlinear biochemical reaction processing.