Forced Generation Outage Investigations for the Northwest Power Pool

After extensive analysis of data representing 1,767 machine-years of hydroelectric generating unit operation, the following conclusions have been reached. 1. There is no evidence that unit size is a factor in determining the outage rate for hydroelectric units. 2. There is some very inconclusive evidence that runner type and age have some effect upon unit outage rate. 3. There is substantial evidence that different types of runners experience outages with different frequencies, but that those types subject to more frequent outages are generally restored to service more rapidly after a forced outage, thus tending to keep the over-all outage rates for various types of units in balance. 4. It was found that neither the exponential nor the log-normal distributions fit the collected data on duration of outages well enough to be accepted as the true underlying distribution. However the log-normal distribution describes the data well enough for most purposes. 5. It was concluded that the best over-all outage factor for hydroelectric units which could be determined from the present data was 0.723%. The evaluation of loss of system generation capacity due to generator forced outages can be made by using probability methods. The Halperin and Adler mathematics provide a method for the calculation of probability which is identical in most respects to other methods for this calculation. The Halperin and Adler mathematics also provide a method for calculating average duration and interval.