Due to the bad operating conditions during wastewater treatment processes which is characterizes of severe random disturbance, strong non-linearity, severe lag, and time-varying properties, it is difficult to build the precise mathematical model. The wastewater treatment is a kind of typical complex process. The actuality of monitor technology of the sewage disposing effluent quality in domestic and overseas is reviewed, and aiming at the difficulty to measure some essential effluent quality parameters, this thesis surveys soft measuring techniques based on statistical regression methods to predict wastewater parameters. The major contents are summarized as follows:1. Build a soft measuring model based on MIMO linear recurrence (MLR). The result of simulation shows that simple linear operation cannot receive good effect if the system has severe non-linearity. Due to the severe complexity and strong non-linearity, it is difficult to receive good result by MLR.2. Utilize principal component analysis (PCA) to lower the dimension of high-dimension variable space, in order to reserve characteristic information of the principal component variable in the low-dimension eigenvector and avoid the redundancy, so the secondary variables are selected. In order to testify the superiority of the PCR, two soft-measuring...
|