Maintenance support system and maintenance support method

A maintenance support system for monitoring a plurality of steam traps in a steam plant, comprising:

• a plurality of detectors connected to a plurality of steam traps in a steam plant, the plurality of detectors configured to detect a state of the plurality of steam traps, the plurality of detectors configured to detect detection signals corresponding to at least one of a temperature, an ultrasonic wave, and a pressure of a steam trap;
• and
• one or more processors that are programmed or configured to:
• deliver maintenance information that indicates an operational status of each steam trap of the plurality of steam traps to maintenance personnel every preset time;
• generate determination results based on detected detection signals that are detected each preset time by the plurality of detectors, the determination results being results of determination regarding a state of the plurality of steam traps;
• wherein delivering the maintenance information comprises delivering predictive information to the maintenance personnel at a specified time between a previous delivery of the maintenance information and a next delivery of the maintenance information, the predictive information providing predictions regarding the operational status of each steam trap of the plurality of steam traps at the next delivery of the maintenance information,
• wherein the maintenance information is generated based on the determination results that are based on the detection signals detected during each preset time, and
• wherein the predictive information is generated from the determination results that are based on the detection signals detected during a period from the previous delivery of the maintenance information to the specified time and a normal range, wherein the normal range comprises a normal upper limit value and a normal lower limit value;
• and
• set, for each steam trap, the normal range by:
• (i) determining, for that steam trap, based on the detection signals, determination target state values during a predetermined first time period;
• (ii) determining, based on the determination target state values, bad state determination results indicating whether that steam trap is in a normal state or a bad state according to an acceptable range including an acceptable upper limit value and an acceptable lower limit value different from the normal upper limit value and the normal lower limit value, wherein that steam trap is determined to be in the normal state if a determination target state value is within the acceptable range and determined to be in the bad state if the determination target state value is outside the acceptable range;
• (iii) setting, based on a coefficient and the determination target state values having bad state determination results indicating that that steam trap is in the normal state, the normal range for that steam trap;
• (iv) determining, based on whether the determination target state values are in the normal range, predictive results indicating whether that steam trap is in the normal state or a state indicating an inclination toward the bad state;
• (v) comparing the predictive results to the bad state determination results corresponding to determination target state values for a predetermined second time period after a third time period including the determination target state values from which the prediction results were determined, to determine correctness determinations indicating whether the predictive results are correct or incorrect;
• (vi) selecting, according to a purpose of that steam trap, an index for an accuracy determination from a first index α, a second index β, and a third index γ, wherein the first index α indicates a first ratio, relative to all of the correctness determinations, of cases in which the predictive result indicates the normal state, and the bad state determination results for the predetermined second time period include a result indicating the bad state, wherein the second index β indicates a second ratio, relative to all of the correctness determinations, of cases in which the predictive result indicates the state indicating the inclination toward the bad state, and the bad state determination results for the predetermined second time period do not include a result indicating the bad state, and wherein the third index γ equals a sum of the first index α and the second index β such that γ=α+β and the third index γ indicates a ratio of all cases of incorrect predictive determination;
• (vii) reducing, by performing the accuracy determination based on the first index α, cases in which a bad state could not be predicted, reducing, by performing the accuracy determination based on the second index β, cases in which unnecessary countermeasures are performed, or reducing, by performing the accuracy determination based on the third index γ, the number of cases of incorrect predictive determination;
• (viii) adjusting the coefficient to reduce the selected index;
• and
• (ix) repeating steps (iii)-(viii) until the selected index satisfies a given value associated with that index.