| 摘要 |
<p>PURPOSE:To discover an abnormality of a device in the early stage and to predict a repairing timing of the device, by statistically processing data of the driving history of a plant, automatically calculating a convenient and effective control value for actual use, and monitoring the process amount in real time by the control value. CONSTITUTION:A process data 1d from each sensor is periodically read into a process data input device 1, and recorded in a data recording part 2. Identification numbers and measuring values are preserved in a time series manner in a history data preserving device 3. A data retrieving part 4 retrieves a data 4d using the identification numbers as factors. A limit value calculating part 5 divides the measured values preserved to a maximum time T into 10, thereby obtaining an average value Z and a standard deviation sigmai for every divided group. 3sigmai is added to or subtracted from tone average value Zi, whereby the upper and lower limit values are obtained. A quadratic regression curve is calculated according to the method of least squares for tone upper and lower limit values of the average value to obtain a control value 5d which is then stored in a control value data memory 6. An alarm outputting part 7 generates an alarm when a current value of the process 6d is beyond the control value 5d.</p> |
