Author_Institution :
Sch. of Bus. Adm., Dalian Univ. of Technol., Dalian, China
Abstract :
To analyze the composition factors of difference between produce and sale (DPS) scientifically, the composition of DPS are analyzed. the measurement techniques of influence factors such as leakage rate of city pipeline, leakage rate of water second supply within inner pipeline, self usage percentage of pipeline system running, usage percentage of fire fighting, sale lose and commonweal usage percentage, free usage percentage and system error percentage caused by pipeline scale. For one of the most difficult to be measured factor, namely, the leakage rate of city pipeline, BP neural network is adopted, the structure and steps are designed. The hydraulic pressure, water supply quantity and leakage quantity are taken as sample data, the leakage rate of city pipeline is output by the trained BP neural network. Considering the influence of water supply quantity and node quantity of pipeline, the output value is adjusted by adjusting formula. To validate the validity of above technique, the total DPS is measured by statistic method, the result shows it is almost the same to the sum of varied influence factor value, the obtained structure and cause of DPS by above technique is objective. For reducing DPS of water supply companies, the result obtained by above technique can provide beneficial reference.
Keywords :
backpropagation; hydraulic systems; pipelines; pressure; statistical analysis; water supply; BP neural network; city pipeline; commonweal usage percentage; composition factor; difference-between-produce-and-sale; fire fighting; free usage percentage; hydraulic pressure; influence factor; leakage quantity; leakage rate; measurement technique; pipeline scale; sale lose; self usage percentage; statistic method; system error percentage; water second supply; water supply company; water supply quantity; Artificial intelligence; Cities and towns; Companies; Computational intelligence; Fires; Marketing and sales; Measurement techniques; Neural networks; Pipelines; Quadratic programming; Difference between produce and sale; Measurement method; Neural network; Water supply;