Method for predicating pH change on basis of neural network model to realize SBR short-cut nitrification

Abstract

The invention provides a method for predicating pH change on the basis of a neural network model to realize SBR short-cut nitrification and belongs to the field of waste water treatment methods. In an SBR system, the aeration time is controlled by using a real-time control strategy; with long-time stably operating SBR data as basic data, a three-layer BP neural network prediction model is established to predicate a pH change curve in advance; according to a predicated pH change point, an aeration stopping time is set in advance to prevent further oxidation of NO2-N. The method for predicating pH change on the basis of the neural network model to realize SBR short-cut nitrification comprises the specific steps: data collection: collecting data through an online instrument; data selection processing: carrying out selection preprocessing on the data collected in the first step, selecting enough continuously stable samples from the data and using pH as a variable parameter; model establishment: modeling the selected data, selecting the BP neural network prediction model for training, correction and testing, and carrying out process sensor pH monitoring and online pH predication on the BP neural network prediction model after the BP neural network prediction model satisfies accuracy requirements.

Description

technical field [0001] The invention relates to a sewage treatment technology, in particular, the neural network model can be used to predict the pH change point of the SBR ammonia oxidation process, and the aeration stop time can be set in advance, so that the SBR system can gradually realize short-range nitrification, and is suitable for denitrification of urban domestic sewage The invention belongs to the field of wastewater treatment methods. Background technique [0002] The traditional biological denitrification technology mainly includes two stages of nitrification and denitrification. The nitrification reaction includes two steps. The first step is the ammonia oxidation process. The ammonia oxidizing bacteria (AOB) converts ammonia nitrogen into nitrite nitrogen. The second step is the nitrification process. , nitrite nitrogen is further converted to nitrate nitrogen by nitrite oxidizing bacteria (NOB). Denitrification is the conversion of nitrite or nitrate into N ...

AI Technical Summary

Problems solved by technology

[0004] The ammonia oxidation process of urban domestic sewage is nonlinear, time-varying and uncertain, which makes it difficult to establish an accurate mathematical model. Even if it is established, the structure is very complicated and difficult to effectively control

Based on the research, it is found that fuzzy control lacks self-learning ability, poor adaptability, low control precision, and difficult practical application. The predictive control based on neural network combines neural network and predictive control, and combines the rolling idea of ​​predictive control with neural network. The characteristics of the network to accurately describe the nonlinear and uncertain dynamic process are organically combined, which can be well used for predictive control of nonlinear systems

Images