应用人工神经网络方法预测油井堵水效果

堵水措施效果测对于多油层油田开发具有重要意义，统计了萨尔图、喇嘛甸油田已堵水井的增油降水效果及相应地质条件、开采状况，对影响堵水效果的因素进行了分析，并归纳为6个表征参数，利用人工神经网络方法建立起了油井堵水的效果预测模型，采用已堵水井影响因素的表征参数，对模型进行训练。应用于13口堵水井的效果预测，结果表明，采用人工神经网络方法克服了传统方法的局限性，建立的模型预测精度高，可靠性强，是一种有效的预测方法。

Applying Artificial Neutral Network to Predict Water Plugging Results of Oil Wells

The prediction of water plugging results is important to the development of multi layered oil field. This paper gives the statistical data about the oil incremental and water decreasing results of the plugged wells in Saertu and Lamadian oil fields, the relative geologic condition and production condition. It analyzes the factors that will influence plugging results, and the factors are generalized as 6 characterization parameters. The prediction model of oil well water plugging results is established based on artificial neutral network. The characterization parameters of the influenced factors for the plugged wells are used to train the model, and the model is applied to predict the results of 13 water plugging wells. The results show that the artificial neutral network method get rid of the limit of the traditional method. The established model is more accurate, reliable and effective in prediction the results.