猪舍有害气体测定与温度智能控制算法

为解决H2S传感器与NH3传感器的交叉敏感问题，充分利用传感器输出信号所包含的气体分子反应过程频谱特性，将小波变换与遗传算法相结合用于猪舍有害气体测定特征提取，较好地提高了有害气体测定的有效性。结果表明，该方法使BP神经网络的定性测定准确率达92%，定量测定的平均测定精度达87%。设计模糊控制算法对猪舍温度进行智能控制，Matlab仿真试验表明该控制算法使系统反应时间短，稳态误差低，较好满足了猪舍温度控制的要求。

Harmful gases determination and temperature intelligent control algorithm in piggery

By adequately utilizing spectrum features during the reaction of gas molecules from sensor output signals, and using a method combining wavelet transform with genetic algorithm for harmful gases determination in piggery, the sensor cross-sensitivity problem from NH3 and H2S sensors was solved, and the effectiveness of determination of harmful gases was improved. Experimental results showed that accuracy of qualitative determination by BP neural network with the proposed method reached 92%, and average accuracy of quantitative determination reached 87%. A fuzzy control algorithm was designed for intelligent temperature control in piggery, and the Matlab simulation results show that system response time becomes shorter and steady-state error is lower by this control algorithm, resulting in satisfaction for requirement of piggery temperature control.