Method for analyzing operation of a machine

• US 7,966,151 B2
• Filed: 10/24/2006
• Issued: 06/21/2011
• Est. Priority Date: 10/24/2006
• Status: Expired due to Fees

• Current Assignee: Ford Motor Company
• Original Assignee: Ford Motor Company
• Inventors: Coffman, Paul E. Jr., Filev, Dimitar, Tseng, Fling
• Primary Examiner(s): Kundu; Sujoy K
• Application Number:
  US11/552,178
• Publication Number:
  US 20080188973A1
• Time in Patent Office: 1,701 Days
• Field of Search: 702/182, 702/183, 702/184, 702/185, 702/186, 702/188, 700/109, 700/110
• US Class Current: 702/182
• CPC Class Codes: G05B 23/0221 Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods
• CPC Class Codes: G05B 23/0283 Predictive maintenance, e.g. involving the monitoring of a system and, based on the monitoring results, taking decisions on the maintenance schedule of the monitored system; Estimating remaining useful life [RUL] preventive maintenance, i.e. planning maintenance according to the available resources without monitoring the system G06Q10/06

Abstract

A constraint analysis and reliability agent executes a method for analyzing operation of a manufacturing asset, and includes the steps of collecting operation data for a machine over a plurality of predetermined time periods. The operation data includes a plurality of mutually exclusive events that describe operation of the machine. For each of the predetermined time periods, it is determined whether the machine is in an “ON” or an “OFF” state. Data for the “OFF” states is removed from the collected data to generate a filtered data set. Reliability information is then generated based, at least in part, on the filtered data set. This facilitates predictions of future machine operation.

Claims

1. A method for analyzing operation of at least one machine, comprising:

collecting operation data for a machine over a plurality of predetermined time periods, the operation data including the respective time for a plurality of events that together form a set of mutually exclusive events that describe operation of the machine, the set of events including cycle events, starve events, blocked events, and down events;

determining for each of the predetermined time periods whether the machine is in an “

ON”

state or an “

OFF”

state based at least in part on the collected operation data;

removing data for the OFF states from at least a portion of the collected operation data, thereby generating a filtered data set; and

generating reliability information for the machine based at least in part on the filtered data set, thereby facilitating predictions of future machine operation, the step of generating reliability information including;

determining a reliability parameter based at least in part on the down events in the filtered data set;

choosing a probability density function; and

fitting the reliability parameter to the chosen probability density function.

2. The method of claim 1, wherein the machine is considered to be in an ON state for a given one of the predetermined time periods when there is more than one occurrence of a cycle event, a starve event, a blocked event, or a combination thereof during the given predetermined time period.

3. The method of claim 1, wherein the step of determining the reliability parameter includes:

determining a time-to-failure parameter;

determining a time-to-repair parameter; and

mathematically combining the time-to-failure parameter and the time-to-repair parameter to generate the reliability parameter.

4. The method of claim 1, further comprising:

collecting additional operation data for the machine;

determining when the machine is in an ON state and when the machine is an OFF state;

removing data for the OFF states from the additional operation data for the machine and updating the filtered data set;

determining when a failure of the machine has occurred;

determining a time-to-repair the failure;

determining a new reliability parameter from the updated filtered data set;

performing an analysis to determine compatibility between the new reliability parameter and a plurality of probability density functions, thereby generating a plurality of compatibility values;

fitting the new reliability parameter to the probability density function having the best compatibility value when the difference between the best compatibility value and the compatibility value for the chosen probability density function is greater than a predetermined amount; and

fitting the new reliability parameter to the chosen probability density function when the difference between the best compatibility value and the compatibility value for the chosen probability density function is not greater than the predetermined amount.

5. The method of claim 1, further comprising:

defining a work area having a plurality of machines;

collecting corresponding operation data for the machines over the predetermined time periods, the operation data including the time for cycle events, starve events, blocked events, and down events for each respective machine;

determining for each of the machines over each of the predetermined time periods whether the respective machine is in an ON state or an OFF state based at least in part on the respective collected operation data;

removing data for the OFF states from at least a portion of the corresponding collected operation data for each machine, such that the filtered data set includes information for each machine; and

generating reliability information for each machine based at least in part on the filtered data set, thereby facilitating predictions of future operation of each machine.

6. The method of claim 5, wherein the step of generating reliability information for each machine includes:

determining a reliability parameter for each machine based at least in part on the down events for the respective machine in the filtered data set;

choosing a corresponding probability density function for each machine; and

fitting each of the reliability parameters to the respective chosen probability density function.

7. The method of claim 5, further comprising:

determining for each of the predetermined time periods whether the work area is in an ON state or an OFF state based at least in part on whether the machines in the work area are in an ON state or an OFF state, the work area being considered to be in an OFF state for a given one of the predetermined time periods if a predetermined percentage of the machines in the work area are determined to be in an OFF state during the given predetermined time period.

8. The method of claim 7, further comprising removing additional data from at least a portion of the collected operation data to generate the filtered data set, the additional data removed including data for the ON states of the machines during a given one of the predetermined time periods when the work area is determined to be in an OFF state for the given predetermined time period.

9. The method of claim 8, further comprising performing a constraint analysis using the reliability information generated for each of the machines, thereby facilitating identification of bottlenecks within the work area.

10. The method of claim 8, wherein the step of performing a constraint analysis includes:

identifying the machine having the greatest percentage of downtime;

identifying at least one other machine having a percentage of downtime within a predetermined amount of the greatest percentage of downtime; and

analyzing the impact on throughput for each of the identified machines, thereby providing information for reducing constraints.

11. The method of claim 1, further comprising:

determining for each of the predetermined time periods whether the machine is in an “

auto and maintenance”

state;

determining for each of the predetermined time periods whether the machine is in an “

auto and faulted”

state; and

removing data for the “

auto and maintenance”

states and data from the “

auto and faulted”

states from the collected operation data prior to the step of removing data for the OFF states.

12. A method for analyzing operation of at least one machine, comprising:

collecting operation data for a machine over a plurality of predetermined time periods, the operation data including the time for cycle events, starve events, blocked events, and down events;

determining for each of the predetermined time periods whether the machine is in an “

ON”

state or an “

OFF”

state based at least in part on the collected operation data;

removing data for the OFF states from at least a portion of the collected operation data, thereby generating a filtered data set;

determining a time-to-failure parameter using at least the down events from the filtered data set;

determining a time-to-repair parameter using at least the down events from the filtered data set;

choosing a probability density function; and

fitting the determined parameters to the chosen probability density function, thereby generating reliability information for the machine to facilitate predictions of future machine operation.

13. The method of claim 12, further comprising:

collecting additional operation data for the machine over a plurality of additional predetermined time periods;

determining for each of the additional time periods whether the machine is in an ON state or an OFF state based at least in part on the additional collected operation data;

removing data for the OFF states from at least a portion of the additional collected operation data, thereby generating an additional filtered data set;

aggregating the filtered data set and the additional filtered data set, thereby generating a new filtered data set;

determining a new time-to-failure parameter using at least the down events from the new filtered data set after the machine has had a failure;

determining a new time-to-repair parameter using at least the down events from the new filtered data set after the failure has been repaired;

determining whether the new determined parameters should be fit to the chosen probability density function or another probability density function chosen from a predetermined list of probability density functions; and

fitting the new determined parameters to one of the probability density functions, thereby generating reliability information for the machine to facilitate predictions of future machine operation.

14. The method of claim 13, wherein the time-to-repair parameter and the additional time-to-repair parameter are each a mean time-to-repair value, and the time-to-failures parameter and the additional time-to-failure parameter are each a mean time-to-failure.

15. The method of claim 12, wherein the machine is considered to be in an ON state for a given one of the predetermined time periods when there is more than one occurrence of a cycle event, a starve event, a blocked event, or a combination thereof during the given predetermined time period.

16. The method of claim 15, further comprising:

determining for each of the predetermined time periods whether the machine is in an “

auto and maintenance”

state;

determining for each of the predetermined time periods whether the machine is in an “

auto and faulted”

state; and

removing data for the “

auto and maintenance”

states and data from the “

auto and faulted”

states from the collected operation data prior to the step of removing data for the OFF states.

17. The method of claim 12, the machine being part of a work area including a plurality of other machines, the method further comprising:

collecting operation data for the other machines over the predetermined time periods, the operation data for the other machines including the time for cycle events, starve events, blocked events, and down events for each machine;

determining for each of the other machines over each of the predetermined time periods whether the other machine is in an ON state or an OFF state based at least in part on the collected operation data; and

determining for each of the predetermined time periods whether the work area is in an ON state or an OFF state based at least in part on whether the machines in the work area are in an ON state or an OFF state.

18. The method of claim 17, wherein the work area is considered to be in an OFF state for a given one of the predetermined time periods if at least half of the machines in the work area are determined to be in an OFF state during the given predetermined time period.

19. The method of claim 18, further comprising:

removing additional data from at least a portion of the collected operation data to generate the filtered data set, the additional data removed including data for the ON states of the machines during a given one of the predetermined time periods when the work area is determined to be in an OFF state for the given predetermined time period.

20. The method of claim 19, further comprising:

determining for each of the machines a constraint parameter related to the time each machine is in an OFF state; and

ordering each of the machines according to the determined constraint parameter, thereby providing information regarding bottlenecks in the work area.

21. A method for analyzing operation of at least one machine, comprising:

defining a work area having a plurality of machines;

collecting corresponding operation data for the machines over a plurality of predetermined time periods, the operation data including the respective time for a plurality of events that together form a set of mutually exclusive events that describe operation of the machine, including the time for cycle events, starve events, blocked events, and down events for each respective machine;

determining for each of the machines over each of the predetermined time periods whether the respective machine is in an “

ON”

state or an “

OFF”

state based at least in part on the respective collected operation data;

removing data for the OFF states from at least a portion of the corresponding collected operation data for each machine, thereby generating a filtered data set that includes information for each machine; and

generating reliability information for each machine based at least in part on the filtered data set, thereby facilitating predictions of future operation of each machine, the step of generating reliability information for each machine including;

determining a reliability parameter for each machine based at least in part on the down events for the respective machine in the filtered data set;

choosing a corresponding probability density function for each machine; and

fitting each of the reliability parameters to the respective chosen probability density function.

22. The method of claim 21, further comprising determining for each of the predetermined time periods whether the work area is in an ON state or an OFF state based at least in part on whether the machines in the work area are in an ON state or an OFF state, the work area being considered to be in an OFF state for a given one of the predetermined time periods if a predetermined percentage of the machines in the work area are determined to be in an OFF state during the given predetermined time period.

23. The method of claim 22, further comprising removing additional data from at least a portion of the collected operation data to generate the filtered data set, the additional data removed including data for the ON states of the machines during a given one of the predetermined time periods when the work area is determined to be in an OFF state for the given predetermined time period.