Intelligent Automatic Control System Design For SBBR One-stage Completely Autotrophic Nitrogen Removal Process

There were many problems in single-stage autotrophic denitrification, such as difficult to start up and stable operation, high require of automatic control accuracy. In order to solve these problems, SBBR one-stage completely autotrophic nitrogen removal process was put as the research object. first of all, through the experiments, the dissolved oxygen was determined when total nitrogen removal rate was largest with the low concentration of influent NH+4 –N, Then, on the basis of the SBBR one-stage completely autotrophic nitrogen removal process, PID controller based artificial neural network was designed for SBBR one-stage completely autotrophic nitrogen removal intelligent control system to pursuit the quickly start and hight efficiency and stable operation of the SBBR one-stage completely autotrophic nitrogen removal system.The main research results and conclusions are as follows:①The start research results of the SBBR one-stage completely autotrophic nitrogen removal intelligent control system showed that the dissolved oxygen was the the main influencing factors of SBBR one-step completely autotrophic nitrogen removal process, when the temperature was 30 ℃, p H was 8.0, the dissolved oxygen with total nitrogen removal rate was the largest changed while the influent N H+4-N and COD concentration was changed, thus when the influent NH+4-N concentration was 250 mg/L and the concentration of COD was 100mg/L, the dissolved oxygen value would be 1.42mg/L, when influent NH+4-N concentration was 100mg/L and the concentration of COD was 50mg/L, dissolved oxygen would be 1.36mg/l, when influent NH+4-N concentration was 50 mg/L and the COD concentration was 50mg/L, and dissolved oxygen would be 1.32mg/L.②A SBBR one-stage completely autotrophic nitrogen removal intelligent control system was designed based on artificial neural network intelligent control, the system were composed of a RBF neural network feedforward controller and a BP neural network PID controller. The artificial neural network intelligent control system and simple PID control were simulinked using MATLAB to analysis their performance in SBBR one-stage completely autotrophic nitrogen, the simulation results showed that: artificial neural network intelligent controller overshoot was 10.4%, rise time was 159 s, static error was 0; PID control overshoot was 15.2%, rise time was 183 s, regulating time was 230 s, static error was 0. Compared with the PID controller, artificial neural network intelligent controller overshoot decreased by 4.8%, rise time reduces 24 s, adjust time reduced 43 s, the application of artificial neural network PID control can improve the transient performance of the control system.③RBF neural network was trained by using projects prophase research and this study start-up phase of the test data, the result showed that when RBF neural network function distribution density spread is 42, the nonlinear correlation coefficient between RBF neural network output data and test data was 0.972, thus RBF neural network fitting effect is good.④Constructing PID control system based on the dissolved oxygen values above correspond to the influent NH+4-N and COD concentration can realize start of the SBBR single-stage autotrophic nitrogen removal system.⑤The operation results of intelligent control of SBBR one-stage completely autotrophic nitrogen removal system showed that the RBF neural network, feedforward control can quickly determine the inverter ’s frequency under different influent NH+4-N and COD concentration, and the dissolved oxygen set value; When the influent NH+4-N concentration was 50 mg/L and the concentration of COD was 50mg/L, the artificial neural network intelligent control system’s rise time and adjusting time was 119 s. The overshoot was 0, the steady-state error was ±0.03mg/L; The effluent total nitrogen of SBBR reactor can stably belowed 3mg/L and the total nitrogen removal rate is between 95.2% and 97.8%, the artificial neural network intelligent control can achieve efficient and stable operation of SBBR one-stage completely autotrophic nitrogen removal process.