This paper proposes a circular seal identification method which based on the average relative error. First it spread out the circular seal into rectangular ones, so as to gain the seal of rectangular image, then it projects the rectangular seal to get projection curve. using projection curve of the reserved seal and the cyclic shift of detected seal projection curve to achieve registration. It calculated the average relative error that according to the registration after the reserved seal and detected seal. With the average relative error of the seal get the authenticity. The experimental results show that this method can simplify the computation and improve the seal authenticity identification effectively, which can also be applied to the actual.

According to the prediction of energy-saving indicators of oilfield, proposed a prediction model based on QPSO optimized LS-SVR. In order to improve the prediction accuracy and speed, described the complex nonlinear relationship of predictors and factors by using LS-SVR, and optimized the parameters of LS-SVR through improved QPSO. The training data is oil production and liquid production of production data of oil production plant. The prediction result shows that, the model can achieve higher accuracy, so the method is effective and feasible.

Up to now, the common method of reservoir well group that is dynamic connectivity, it mainly includes tracer testing, stress testing, well testing, and numerical simulation. The implementation of these methods is more complex, expensive, high cost, and will affect the normal production of the oilfield. Because of the convenient injection and dynamic data it can get convenient. This paper presents a method that using dynamic reservoir development data inverse well group connectivity. CART algorithm analysis and extraction of potential knowledge from the oilfield development. It establish direct mapping of logging data and well group connectivity relationship. Experiments show that using dynamic data to study well group connectivity relationship can greatly reduce the cost and as a result has a higher accuracy.

This paper introduces method of gas energy saving target forecast that a quantum particle swarm optimization algorithm and BP neural network, using BP neural networks and quantum particle swarm global search ability strong advantage, through the method that improved average optimal position. It solved the BP neural network is being trapped in local minima and slow convergence speed problem. It realized target forecast that based on BP neural network and quantum particle swarm field energy saving. With the production water injection pump unit consumption data for training data, prediction results show that this method can achieve good prediction effect. It can be applied to practice.