Method for predicting concentration of disinfection by-product haloacetic acid in water supply system

Abstract

The invention discloses a method for predicting the concentration of a disinfection byproduct haloacetic acid in a water supply system. The method comprises the following steps: S1, dividing a haloacetic acid and water quality parameter database into a training set and a prediction set; S2, establishing a neural network model by using a neural network newrb function in MATLAB (Matrix Laboratory), wherein the format of the neural network model is net = newrb (P, T, GOAL, SPREAD, MN, DF); S3, training the built neural network model, verifying and checking the prediction performance of the neural network model through the model, directly entering the step S5 when the correct rate is high, and performing the step S4 when the correct rate is low; S4, when the test accuracy is low, adjusting the number of SPREAD and neurons, and repeating the step S2 until a satisfactory result is obtained; S5, stopping. The prediction effect of the RBF neural network model is much better than that of a linear model or an lg linear model, the time consumed in the RBF neural network operation process is very short, and the operation time of each neural network model is generally 10-13s.

Description

technical field [0001] The invention belongs to the technical field of detecting the concentration of chloroacetic acid, a by-product of disinfection in a water supply system, and in particular relates to a method for predicting the concentration of haloacetic acid, a by-product of disinfection in a water supply system. Background technique [0002] Disinfection is an essential means to ensure the hygiene and safety of drinking water. However, many disinfection by-products (DBPs) are produced during the disinfection process, among which haloacetic acids (HAAs) are one of the DBPs with the highest detection rate and the largest amount in drinking water. There are nine common haloacetic acids: monochloroacetic acid (CAA), monobromoacetic acid (BAA), dichloroacetic acid (DCAA), dibromoacetic acid (DBAA), bromochloroacetic acid (BCAA), trichloroacetic acid (TCAA) , Bromodichloroacetic acid (BDCAA), dibromochloroacetic acid (DBCAA), tribromoacetic acid (TBAA). Experiments have p...

AI Technical Summary

Problems solved by technology

However, since the formation of HAAs is a very complicated process, many studies have shown that the influence of water quality operating parameters on the formation of HAAs is not represented by a simple linear or log linear relationship, but more of a weak linear or nonlinear relationship.

As a result, linear model predictions for many HAAs are suboptimal

For example, Song Qianyun et al. used the log linear regression model to predict the formation of HAAs after chlorination of Qiantang River source water. The results showed that the prediction accuracy for BDCAA and TCAA was only 47.2% and 58.3%.

[0004] There are also a small number of literatures that use feedforward-backpropagation neural network (referred to as BP network) to predict the formation of HAAs after chlorination of water bodies. The process converges slowly and requires a long training time. For some complex problems, the training time required by the BP algorithm may be very long

[0005] Moreover, most of the existing HAAs models are established based on the HAAs data obtained after simulated disinfection of water source water, and a small part is established by water plant water (factory water that has not been transported by the pipeline network), but from source water to real users Drinking water has undergone many water treatment steps during the period, and it has also been transported through the pipeline network after disinfection. The composition and distribution of its HAAs are far from those obtained after simulated disinfection of source water.

Moreover, most of the models built based on laboratory simulation disinfection include the indicator of "disinfection time", but the actual pipe network system is difficult to measure the disinfection time (that is, the total time spent from the start of disinfection in the water plant to the delivery to the user end) How much time), so these HAAs models are difficult to actually apply to the actual water supply system, the drinking water network needs to establish its own HAAs model to achieve real prediction and application

Images