A multi-modal-perception-based wastewater treatment method and system

Abstract

A multi-modal-perception-based wastewater treatment method, characterized in that the method achieves treatment of wastewater, includes following: Step S1. 17 sensors locate at each process sections. Collecting n sets of data. Generating a 17-dimensional feature vector through multi-source fusion, where a=1, 2, 3...n; Step S2. Calculating theoretical control parameters from the 17-dimensional feature vector V=[v1, v2,...v17]ᵀ through the calculation model, with an LSTM model predicting compensation values; Step S3. Superimposing the theoretical control parameter set Uasm = [u1, u2, ..., up]T with the theoretical control parameter compensation vector Ulstm = [Δu1,Δu2, ...,Δup]T to obtain the decision control parameter set u=[u1, u2,...up]; Step S4. Monitoring the effectiveness of the final control commands and dynamically adjusting the LSTM network and NSGA-II algorithm; Step S5. Performing safety operations on the final control commands being executed. The invention has the advantage of optimizing water treatment.