Intelligent optimization method for operation process of oil pumping unit

The invention provides an intelligent optimization method for the operation process of an oil pumping unit. The method comprises the following steps: firstly, determining production efficiency influence factors and property variable of the oil pumping unit; secondly, obtaining the sample data of the oil pumping unit; thirdly, performing dimensionality reduction on load data; fourthly, constructing a network input variable set through non-load variable and a load new principal element; fifthly, constructing a sample value of input variable; sixthly, performing normalization processing; seventhly, selecting training and testing samples; eighthly, constructing a feedforward neural network and training; ninthly, constructing a liquid production capacity preference function; tenthly, constructing parent and descendant population; eleventhly, performing genetic cross and variation computation on parent population individuals, so as to produce descendant population individuals; twelfthly, solving a fitness function of the population individuals; thirteenthly, dividing the individuals into non-dominated sets in different levels; fourteenthly, selecting individuals from the non-dominated sets, so as to construct a new parent population and circulating, and finally obtaining an optimized production efficiency influence factor value. After optimization is performed, the fact that the power consumption is the minimum in the condition that the liquid production capacity is basically fixed can be guaranteed.