矿井突水已成为影响矿山安全生产的主要危害之一，快速准确识别突水水源类型是矿井突水灾害治理的关键步骤。提出了1种基于PCA-GA-RF的矿井突水水源识别模型；基于安徽省颍上县谢桥煤矿的88组水样实测数据，遵循分层随机抽样的原则，按照7∶3的比例将其分为62组训练样本和26组预测样本，经PCA提取4个主成分，构建PCA-GA-RF模型，并与PCA-RF、PCA-ABC-RF和PCA-FA-RF模型对比。结果表明：PCA-GA-RF模型判别结果准确率为96.153 8%，与其他模型相比准确率、精确率、召回率和F₁值（精确召回率）最高，具有优越性。

Mine sudden water has become one of the main hazards affecting the safety production of mines, and rapid and accurate identification of the type of sudden water source is a key step in the management of mine sudden water disaster, so a PCA-GA-RF-based mine sudden water source identification model is proposed. Based on the measured data of 88 groups of water samples from Xieqiao Coal Mine in Yingshang County, Anhui Province, and following the principle of stratified random sampling, it was divided into 62 groups of training samples and 26 groups of prediction samples according to the ratio of 7:3, and the four principal components were extracted by PCA to construct the PCA-GA-RF model, and compare it with the PCA-RF, PCA-ABC-RF and PCA-FA-RF models. The results show that the PCA-GA-RF model discriminates the results with an accuracy of 96.153 8%, which is superior with the highest accuracy, precision, recall and F₁ value compared with other models.