METHOD FOR PREDICTING DISORDER OF TOWER CRANE BY USING DATA MINING

摘要

The present invention relates to a method for predicting power crane faults by using data mining which comprises the steps of: defining at least two sets of status data for a faulty tower crane; collecting the status data for the fau performing normalization to transform the status data collected from the tower crane in a scheme of converting a value equal to or higher than a threshold value of the status data to 1, and transform other values in proportion to the threshold, so the result is in the range of 0 to 1; estimating the probability of the fault as an output value by using a sigmoid function as a transfer function while adjusting the connection weight in a learning process in advance, and inputting the normalized data to a neural network; predicting the fault period by using the probability of the fault as an independent variable and the time when the probability of the fault reaches 1 as a dependent variable in a regression analysis. According to the present invention, the method for predicting power crane faults by using data mining can accurately predict the time of expected tower crane faults, thus the tower crane will never malfunction during the operation, thereby enabling a harbor to maintain work efficiency.