Algae Removal Experiments And Electric Pulse Parameter Optimizing Based On Neural Network And Genetic Algorithm

Circulating cooling water system is the essential equipment for heat transfer oflarge-scale installations in industrial production. Due to good conditions for microbialbreeding in circulating cooling water system, bacteria and algae breed rapidly. Thecombination of microbial metabolites and impurities in water causes lots of biologicalslime and corrosion in the circulating cooling water system, which results in decline ofheat transfer efficiency. So sterilization and algae removal are both important to ensurethe normal operation of circulating cooling water equipment.At present, most studies in the domestic and foreign focus on sterilization ofhigh-voltage and low-frequency electromagnetic pulse, while there is less study onmicroorganism killing of high-frequency and low-voltage electromagnetic pulse.Moreover, there is no rational basis for choosing electromagnetic parameter in watertreatment of electromagnetic pulse. It is necessary to find a solution to optimizingelectromagnetic parameter. The main research works are as follows:①Neural network and genetic algorithm, which are both hotspots, are introducedand combined together. The first reason is that the method of GA optimizes neuralnetwork’s initial weights and threshold, which helps to establish an accurate predictionmodel on algae removal. The second reason is that when the method of GA is unitedwith neural network prediction model, it is possible to find out the optimum electricparameters for algae removal. The program of above two methods has been compiledsuccessfully and is proved to be effective in both theory and experiment analysis..②Algae removal experiments of high-frequency and low-voltage electromagneticpulse are carried out in lab. Chlorophyll a content in water is an important indicator forwater quality and it is almost proportional to algae density in water. So we use alterationrate of chlorophyll a instead of algae density in water as an indicator. By orthogonalexperiment design,50groups of experiments were performed. In experiments,chlorophyll a alteration rate was tested by water quality analyzer(manta2) with varyingthe three parameters such as voltage, frequency and processing time. Experiment resultsindicated that high-frequency and low-voltage electromagnetic pulse could kill algae inwater effectively. Through variance analysis of experiment data, we found that theprocessing time makes the greatest algae removal effect, after that the electric parameteris voltage and finally frequency. ③Based on50groups of experiment results, a neural network prediction modelon algae change rate was established via GA optimizing neural network’s initial weightsand threshold. Simulation results showed error of the prediction model is within7%andthe optimum group of electromagnetic parameters was as follows:U66V,P204KHz,T76min.