Considering that random error and system error in the gas turbine condition monitoring are not easy to distinguish and identi -fy, this article first put forward using statistical hypothesis test to solve the problem .Build gas turbine model to determine turbine operation's general parameters in the hypothesis test , set the actual collection of measurable parameters as the sample from the overall data, cope with the sample data and finally get the t statistic .According to the t-distribution threshold judgment to complete the monito-ring task and use LabVIEW programming to prove the correctness of the t-test.%本文针对燃气轮机状态监测中随机误差与系统误差不易区分和辨别的难题,首次提出了利用统计学中的假设检验进行状态监测的方案。通过燃气轮机变工况模型确定假设检验中的燃气轮机运行参数的总体均值,并将实际采集的可测参数作为来自总体的样本,进行统计处理,得到统计量t。根据t分布临界值判断并完成监测任务,最后用LabVIEW编程计算实例论证了t检验的正确性。
展开▼