CAJ | 학술논문

为了实现对煤与瓦斯突出快速、准确和动态预测，提出了一种基于主成分分析(PCA)和案例推理(CBR)的煤与瓦斯突出预测方法。考虑煤与瓦斯突出多种影响因素，利用案例推理技术对煤与瓦斯突出危险性进行预测。同时采用一种基于PCA的案例描述特征权值确定方法，以提高案例检索效率以及煤与瓦斯突出预测准确率。利用实测数据对所提方法进行验证，实例验证结果表明，所提方法预测结果的准确性和稳定性更高，预测平均误差和最大误差分别仅为0.154%和0.77%，远小于模糊神经网络方法和专家给定权值的案例推理方法。
위료실현대매여와사돌출쾌속、준학화동태예측，제출료일충기우주성분분석(PCA)화안례추리(CBR)적매여와사돌출예측방법。고필매여와사돌출다충영향인소，이용안례추리기술대매여와사돌출위험성진행예측。동시채용일충기우PCA적안례묘술특정권치학정방법，이제고안례검색효솔이급매여와사돌출예측준학솔。이용실측수거대소제방법진행험증，실례험증결과표명，소제방법예측결과적준학성화은정성경고，예측평균오차화최대오차분별부위0.154%화0.77%，원소우모호신경망락방법화전가급정권치적안례추리방법。
In order to realize the accurate,speed and dynamic prediction of coal and gas outburst,a prediction meth?od based on principal component analysis(PCA)and case-based reasoning(CBR)was proposed. Considering multiple influencing factors of coal and gas outburst,the hazard prediction is done with CBR technology. At the same time,a method based on PCA is used in weights allocation for case retrieval and matching to improve the retrieval efficien?cy and prediction precision. The proposed method was validated using practical measured data. The simulation ex?ample shows that the proposed method provides more accurate and robust prediction results and the average predic?tion error and maximum prediction error are 0.154%and 0.77%,respectively. The prediction errors are much less than that obtained from the fuzzy neural network method and the CBR method using weights given by experts.