Determining maintenance for a machine

1. A computer-implemented method of determining a maintenance task for a machine, wherein the method is performed using one or more processors or special-purpose computing hardware, the method comprising:

accessing a target sensor log corresponding to a first machine;

accessing one or more prior sensor logs corresponding to the first machine and one or more prior sensor logs corresponding to a plurality of second machines which are of the same type as the first machine;

accessing a plurality of computer readable logs corresponding to the first machine and the second machines, the computer readable logs for each second machine comprising a maintenance log comprising a plurality of maintenance task objects, each maintenance task object comprising a time and a maintenance task type;

determining a set of statistical metrics characterising a difference between the target sensor log and each prior sensor log;

selecting a sub-set of the prior sensor logs in dependence upon the statistical metrics;

identifying, based on the maintenance logs of the plurality of second machines, a set of anomalies of the plurality of second machines;

determining a subset of the anomalies that did not reoccur within a predetermined amount of time following the times of the corresponding maintenance task objects;

classifying each of the subset of the anomalies as a rectified anomaly;

selecting, based on the sub-set of prior sensor logs and the classifying each of the subset of the anomalies as a rectified anomaly, correlated maintenance tasks objects from the plurality of maintenance task objects, each of the correlated maintenance tasks objects being responsible for previously rectifying a respective anomaly of the set of anomalies of the plurality of second machines, each of the correlated maintenance task objects being correlated, based on analysing the maintenance logs, to a respective prior sensor log included in the subset of prior sensor logs;

selecting a priority maintenance task based on the sub-set of prior sensor logs and the correlated maintenance task objects;

outputting the priority maintenance task; and

carrying out the priority maintenance task on the first machine.

2. The method according to claim 1 , wherein the computer readable logs further comprise, for the first machine and for each second machine:

a message log including a plurality of message objects, each message object comprising a time and a message type.

3. The method according to claim 1 , wherein the computer readable logs further comprise, for the first machine and for each second machine:

a fault log including a plurality of fault objects, each fault object comprising a time, a duration and a fault type.

4. The method according to claim 1 , wherein selecting a sub-set of the prior sensor logs comprises:

ranking each prior sensor log, based on the statistical metrics, according to one or more comparisons between the prior sensor log and the target sensor log; and

selecting the sub-set as a number, N, of the prior sensor logs which are ranked as the N closest to the target sensor log.

5. The method according to claim 1 , wherein selecting a sub-set of the prior sensor logs comprises selecting each prior sensor log having statistical metrics which satisfy one or more thresholds.

6. The method according to claim 1 , further comprising:

accessing a target maintenance log corresponding to the first machine and comprising a plurality of maintenance task objects, each maintenance task object comprising a time and a maintenance task type;

accessing a target fault log comprising a target fault object corresponding to the first machine, the target fault object comprising a time, a duration and a fault type;

accessing a target message log corresponding to the first machine and comprising a plurality of message objects, each message object comprising a time and a message type;

determining a set of log metrics derived from the computer readable logs, the target maintenance log, the target fault log and the target message log; and

wherein selecting a sub-set of the prior sensor logs is based on the statistical metrics and the log metrics.

7. The method according to claim 1 , further comprising presenting the priority maintenance task on a display and/or generating a textual report identifying the priority maintenance task.

8. The method according to claim 1 , further comprising:

receiving an indication of whether the priority maintenance task is effective and;

in dependence upon the priority maintenance task is ineffective:

updating the maintenance log corresponding to the first machine; and

selecting a new priority maintenance task based on re-determining, based on the updated maintenance log corresponding to the first machine, the sub-set of prior sensor logs, the ranking of prior sensor logs and the correlated maintenance tasks; and

outputting the new priority maintenance task.

9. The method according to claim 1 , wherein correlating each prior sensor log included in the subset to one or more correlated maintenance tasks further includes determining a number and type of one or more spare parts and/or one or more consumables associated with each correlated maintenance task.

10. 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 a target sensor log corresponding to a first machine;

accessing one or more prior sensor logs corresponding to the first machine and one or more prior sensor logs corresponding to a plurality of second machines which are of the same type as the first machine;

accessing a plurality of computer readable logs corresponding to the first machine and the second machines, the computer readable logs for each second machine comprising a maintenance log comprising a plurality of maintenance task objects, each maintenance task object comprising a time and a maintenance task type;

determining a set of statistical metrics characterising a difference between the target sensor log and each prior sensor log;

selecting a sub-set of the prior sensor logs in dependence upon the statistical metrics;

identifying, based on the maintenance log for each of the plurality of second machines, a set of anomalies of the plurality of second machines;

determining a subset of the anomalies that did not reoccur within a predetermined amount of time following the times of the corresponding maintenance task objects;

classifying each of the subset of the anomalies as a rectified anomaly;

selecting, based on the sub-set of prior sensor logs and the classifying each of the subset of the anomalies as a rectified anomaly, correlated maintenance tasks objects from the plurality of maintenance task objects, each of the correlated maintenance tasks objects being responsible for previously rectifying a respective anomaly of the set of anomalies of the plurality of second machines, each of the correlated maintenance task objects being correlated, based on analysing the maintenance logs, to a respective prior sensor log included in the subset of prior sensor logs;

selecting a priority maintenance task based on the sub-set of prior sensor logs and the correlated maintenance task objects;

outputting the priority maintenance task; and

carrying out the priority maintenance task on the first machine.

11. An apparatus for determining a maintenance task for a machine, the apparatus comprising one or more processors or dedicated hardware configured to:

access a target sensor log corresponding to a first machine;

access one or more prior sensor logs corresponding to the first machine and one or more prior sensor logs corresponding to a plurality of second machines which are of the same type as the first machine; and

access a plurality of computer readable logs corresponding to the first machine and the second machines, the computer readable logs for each second machine comprising a maintenance log comprising a plurality of maintenance task objects, each maintenance task object comprising a time and a maintenance task type; the apparatus further comprising:

a statistical metric determining module configured to:

determine a set of statistical metrics characterising a difference between the target sensor log and each prior sensor log; and

identify, based on the maintenance log for each of the plurality of second machines, a set of anomalies of the plurality of second machines;

a ranking module configured to select a sub-set of the prior sensor logs in dependence upon the statistical metrics;

a maintenance task determining module configured to control a fault maintenance correlation module to:

determine a subset of the anomalies that did not reoccur within a predetermined amount of time following the times of the corresponding maintenance task objects; and

classify each of the subset of the anomalies as a rectified anomaly;

a fault maintenance correlation module configured to select, based on the sub-set of prior sensor logs and the classifying each of the subset of the anomalies as a rectified anomaly, correlated maintenance tasks objects from the plurality of maintenance task objects, each of the correlated maintenance tasks objects being responsible for previously rectifying a respective anomaly of the set of anomalies of the plurality of second machines, each of the correlated maintenance task objects being correlated, based on analysing the maintenance logs, to a respective prior sensor log included in the subset of prior sensor logs; and

a maintenance task determining module configured to:

select a priority maintenance task based on the sub-set of prior sensor logs, and the correlated maintenance task objects; and

output the priority maintenance task to be carried out on the first machine.

12. The system according to claim 11 , wherein the ranking module is configured to:

rank each prior sensor log, based on the statistical metrics, according to one or more comparisons between the prior sensor log and the target sensor log; and

select the sub-set as a number, N, of the prior sensor logs which are ranked as the N closest to the target sensor log.

13. The system according to claim 11 , wherein the ranking module is configured to select a sub-set of the prior sensor logs by selecting each prior sensor log having statistical metrics which satisfy one or more thresholds.

14. The system according to claim 11 , the apparatus further configured to:

access a target maintenance log corresponding to the first machine and comprising a plurality of maintenance task objects, each maintenance task object comprising a time and a maintenance task type;

access a target fault log comprising a target fault object corresponding to the first machine, the target fault object comprising a time, a duration and a fault type;

access a target message log corresponding to the first machine and comprising a plurality of message objects, each message object comprising a time and a message type;

the system further comprising a log metric determining module configured to determine a set of log metrics derived from the computer readable logs, the target maintenance log, the target fault log and the target message log;

wherein the ranking module is further configured to select a sub-set of the prior sensor logs based on the statistical metrics and the log metrics.