A failure prediction system includes: a processor, the processor being configured to: collect torque values of a drive axis of a robot that is operating in accordance with a given work program; derive an evaluation formula approximating a time change of the torque value which is most recent from among the collected torque values set a failure threshold that is the torque value at which it is determined that failure of the drive axis occurs, based on the evaluation formula and the time change of the torque value when the drive axis reached failure in the past; and calculate an estimated value for the torque value when a prediction time set in advance has elapsed in the evaluation formula, and determines whether failure of the drive axis is predicted within the prediction time according to comparison between the estimated value and the failure threshold.

An abnormality determination device includes a control unit for determining an abnormality of a robot, the control unit being configured to calculate a measurement probability distribution which is a probability distribution using disturbance torque acquired during a predetermined period as a random variable. The control unit causes an average of the calculated measurement probability distribution to conform to an average of an evaluation normal model which is a predetermined probability distribution, compares the measurement probability distribution with the evaluation normal model of which the respective averages conform to each other, and determines an abnormality of the robot in accordance with a result of the comparison.

A system determining whether the feed shafts are normal or abnormal, the system including: a command generation unit that moves the feed shafts in a forward direction at a predetermined speed from a lower limit value to an upper limit value of a range of feeding movement by the numerical controller; and a feed shaft abnormality determination unit that monitors a drive torque command during the movement of the feed shafts in the forward direction by the command generation unit, compares a monitoring result during the movement in the forward direction with a normal value of the drive torque command, determines that abnormality occurs when the drive torque command deviates from the normal value, and outputs the determination result.

A failure prediction system according to an aspect of the present disclosure includes: a torque value collection unit which collects torque values of a drive axis of a robot that is operating in accordance with a given work program; an evaluation formula derivation unit which derives an evaluation formula approximating a time change of the torque value which is most recent from among the torque values collected by the torque value collection unit; a threshold setting unit which sets a failure threshold that is the torque value at which it is determined that failure of the drive axis occurs, based on the evaluation formula and the time change of the torque value when the drive axis reached failure in the past; and a prediction determination unit which calculates an estimated value for the torque value when a prediction time set in advance has elapsed in the evaluation formula, and determines whether failure of the drive axis is predicted within the prediction time according to comparison between the estimated value and the failure threshold.

A system determining whether the feed shafts are normal or abnormal, the system including: a command generation unit that moves the feed shafts in a forward direction at a predetermined speed from a lower limit value to an upper limit value of a range of feeding movement by the numerical controller; and a feed shaft abnormality determination unit that monitors a drive torque command during the movement of the feed shafts in the forward direction by the command generation unit, compares a monitoring result during the movement in the forward direction with a normal value of the drive torque command, determines that abnormality occurs when the drive torque command deviates from the normal value, and outputs the determination result.