Machine fault modelling
Systems, methods, non-transitory computer readable media can be configured to access a plurality of sensor logs corresponding to a first machine, each sensor log spanning at least a first period; access first computer readable logs corresponding to the first machine, each computer readable log spanning at least the first period, the computer readable logs comprising a maintenance log comprising a plurality of maintenance task objects, each maintenance task object comprising a time and a maintenance task type; determine a set of statistical metrics derived from the sensor logs; determine a set of log metrics derived from the computer readable logs; and determine, using a risk model that receives the statistical metrics and log metrics as inputs, fault probabilities or risk scores indicative of one or more fault types occurring in the first machine within a second period.
1. A computer-implemented method of determining a fault probability for a machine, wherein the method is performed using one or more processors or special-purpose computing hardware, the method comprising:
accessing sensor logs corresponding to a machine, each sensor log spanning a first period;
accessing computer readable logs corresponding to the machine, each computer readable log spanning the first period, the computer readable logs comprising:
a maintenance log recording information related to locations, times, and types of respective previous maintenance tasks performed on the machine; and
a fault log recording information related to locations, durations, types, and resolutions of respective faults that have occurred on the machine;
predicting probabilities that respective faults of each type will occur and severities of the respective predicted fault types on the machine during a second period, based on information from the computer readable logs;
selecting a fault type based on the predicted probabilities and the severities;
selecting a maintenance task associated with the selected fault type, wherein the selecting of the maintenance task comprises:
predicting a delay resulting from the selected fault; and
selecting a maintenance task type, from maintenance task types, based on expected reductions, of the delay, resulting from the respective maintenance task types;
and
carrying out the selected maintenance task.
2. The method according to claim 1 , wherein the computer readable logs for the machine further comprise a message log including a time, an identifier of a component sub-system, and warning information related to a fault.
3. The method according to claim 1 , wherein the selecting the maintenance task comprises selecting a maintenance task type having a highest probability, compared to other maintenance task types, of resolving the selected fault type based on previous occurrences of the selected fault type on the machine and previous maintenance tasks performed in connection with the previous occurrences.
4. The method according to claim 1 , wherein the selecting a maintenance task type comprises selecting a maintenance task type, from the maintenance task types, having a highest expected reduction of the delay.
5. The method according to claim 1 , wherein the selecting the fault type comprises selecting a fault type having a highest product of the predicted probability and the severity, compared to other fault types.
6. The method according to claim 1 , further comprising:
diagnosing a fault based on a relationship between pairs of parameters in the sensor logs.
7. The method according to claim 6 , wherein the diagnosing the fault comprises:
determining a parameter curve indicating how a first parameter changes with respect to a second parameter;
in response to detecting that the determined parameter curve deviates by more than a threshold amount from an average parameter curve indicating an expected relationship between the first parameter and the second parameter, determining that a fault is developing in the machine.
8. The method according to claim 7 , wherein the detecting that the determined parameter curve deviates by more than a threshold amount from an average parameter curve comprises determining that a rate of change of the first parameter with respect to the second parameter from the determined parameter curve deviates from a rate of change of the first parameter with respect to the second parameter from the average parameter curve by more than a threshold amount.
9. The method according to claim 1 ,
wherein the carrying out the selected maintenance task comprises carrying out the selected maintenance task on a sub-system, compared to other sub-systems of the machine, having a highest predicted probability that the selected fault type will occur or a highest severity of the selected fault type.
10. The method according to claim 1 , wherein the maintenance log comprises free text notes relating to the previous maintenance tasks, and the method further comprises:
determining frequencies of occurrences of words and phrases relating to the previous maintenance tasks, synonyms of the words and phrases, and misspelled versions of the words and phrases, in the free text notes.
11. A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to perform a method comprising:
accessing sensor logs corresponding to a machine, each sensor log spanning a first period;
accessing computer readable logs corresponding to the machine, each computer readable log spanning the first period, the computer readable logs comprising:
a maintenance log recording information related to locations, times, and types of respective previous maintenance tasks performed on the machine,
wherein the maintenance log comprises free text relating to the previous maintenance tasks; and
a fault log recording information related to locations, durations, types, and resolutions of respective faults that have occurred on the machine;
predicting probabilities that respective faults of each type will occur and severities of the respective predicted fault types on the machine during a second period, based on frequencies of occurrences of words and phrases relating to the previous maintenance tasks, and any of synonyms of the words and phrases, or misspelled versions of the words and phrases, in the free text;
selecting a fault type based on the predicted probabilities and the severities;
selecting a maintenance task associated with the selected fault type; and
carrying out the selected maintenance task.
12. A system, the system comprising one or more processors or special-purpose computing hardware configured to:
accessing sensor logs corresponding to a machine, each sensor log spanning a first period;
accessing computer readable logs corresponding to the machine, each computer readable log spanning the first period, the computer readable logs comprising:
a maintenance log recording information related to locations, times, and types of respective previous maintenance tasks performed on the machine;
a fault log recording information related to locations, durations, types, and resolutions of respective faults that have occurred on the machine;
predicting probabilities that respective faults of each type will occur and severities of the respective predicted fault types on the machine during a second period, based on information from the computer readable logs;
selecting a fault type based on the predicted probabilities and the severities, wherein the selecting of the fault type comprises selecting a fault type based on products of the predicted probabilities and respective severities;
selecting a maintenance task associated with the selected fault type; and
carrying out the selected maintenance task.
13. The system according to claim 12 , wherein the computer readable logs for the machine further comprise a message log including a time, an identifier of a component sub-system, and warning information related to a fault.
14. The system according to claim 12 , wherein the one or more processors or special-purpose computing hardware are further configured to diagnose a fault based on a relationship between pairs of parameters in the sensor logs.
15. The system according to claim 14 , wherein the diagnosing the fault comprises:
determining a parameter curve indicating how a first parameter changes with respect to a second parameter;
in response to detecting that the determined parameter curve deviates by more than a threshold amount from an average parameter curve indicating an expected relationship between the first parameter and the second parameter, determining that a fault is developing in the machine.
16. The system according to claim 15 , wherein the detecting that the determined parameter curve deviates by more than a threshold amount from an average parameter curve comprises determining that a rate of change of the first parameter with respect to the second parameter from the determined parameter curve deviates from a rate of change of the first parameter with respect to the second parameter from the average parameter curve by more than a threshold amount.
17. The system according to claim 12 , wherein the carrying out the selected maintenance task comprises carrying out the selected maintenance task on a sub-system, compared to other sub-systems of the machine, having a highest predicted probability that the selected fault type will occur or a highest severity of the selected fault type.