Automated machine learning-based accurate dosing method for sewage treatment plant

Abstract

The present invention belongs to the technical field of accurate dosing during sewage treatment by sewage treatment plants. Disclosed is an automated machine learning-based accurate dosing method for a sewage treatment plant. The method comprises: acquiring historical data of a sewage treatment plant, and constructing a data set; training an automated machine learning model, and on the basis of influent water data and dosing data, predicting effluent water data; on the basis of a dosing rule, compiling statistics to obtain a dosage corresponding to each piece of influent water data, so as to obtain a dosing sample data set; extracting data from the dosing sample data set, and inputting same into an effluent water prediction model, so as to obtain corresponding effluent water quality data; comparing the effluent water quality data with a water quality standard-reaching threshold value; and adjusting the dosage to continuously update the dosing sample data set of the sewage treatment plant, until all effluent water quality data intervals corresponding to all samples fall within a standard-reaching range. The present invention can more flexibly adapt to different sewage treatment environments, quickly respond to a change in influent water quality, and improve the real-time performance and precision of dosing decisions, thereby greatly improving the efficiency and reliability of a sewage treatment system.

Images