Reliability HotWire

Reliability HotWire

Issue 130, December 2011

Reliability Basics

Analyzing Maintenance Data with the Event Log Folio in Weibull++

Companies in the manufacturing world face an ever increasing pressure to make the manufacturing process as cost-effective as possible. Maintenance is an integral part of that process and effective maintenance policies can lead to reduced manufacturing costs. However, in order to determine the appropriate maintenance policies, we must first quantify the reliability and maintainability of physical assets, such as a piece of equipment in an assembly line. This article provides a simple example of how to analyze maintenance data using the event log folio in Weibull++, and how to use the results of this analysis in a reliability block diagram in BlockSim.

Example

Consider a simple example where a piece of equipment consists of two subsystems (A and B) and each subsystem consists of three components (A1, A2, A3, and B1, B2, B3). The dates/times that the equipment failed and the dates/times it was restored were collected and logged in a Weibull++ event log folio, as shown next.

The "F/E" column indicates whether the observation is a failure (F) or a general event (E). A general event represents an activity that brings the system down but is not directly relevant to the reliability of the equipment, such as preventive maintenance, routine inspections and the like. In this case, all observations were identified as failures. The "Level 1" column indicates the highest level within the equipment in which the failure occurred (subsystem A or B), while the "Level 2" column, indicates the component responsible for the observed failure (components A1, A2, A3, B1, B2 or B3).

The next figure shows the Main page of the control panel and the analysis settings.

The steps outlined below describe the process that a user would go through to configure these settings on the control panel.

  1. Select which level(s) to analyze. In general, each repairable system is composed of non-repairable items that are referred to as lowest replaceable units (LRUs). In this example, the level 1 subsystems (subsystem A and B) are not LRUs because in the event of failure, these subsystems are not replaced, but rather, only a component within the subsystem is replaced. The level 2 components, on the other hand, are LRUs because they are replaced when they fail. The event log folio allows you analyze up to four levels, but in this example, we only need to go down two levels.

  2. Select how to analyze the failures and events. If the log contains information about events that are not considered to be failures (e.g., downtime due to inspections, maintenance procedures and the like), the non-failure events are identified with an E in the "F/E" column and the user has the option to analyze failures and events separately or jointly. This setting doesn’t apply in this example because all the observations are failures.

  3. Enter the date/time the data collection period started and ended. The dates/time entered in the System area of the control panel will affect how the software converts the recorded data to time-to-failure and time-to-repair data. The times-to-failure of each component will be calculated by the following equation:

    Time-to-Failure = Date/Time of Failure - Date/Time Last Restoration - System Downtime Since Last Restoration

    (1)

    The time-to-repair of each component will be calculated by the following equation:

    Time-to-Repair = Date/Time of Restoration - Date/Time of Failure

    (2)

    The last suspension time for each component will be calculated as:

    Last Suspension = System End Date/Time - Date/Time of Last Restoration

    (3)

    In this example, the equipment is not new (i.e., the equipment has operated for an unknown period of time before the system start date); therefore, the time to first occurrence of an event will be treated as a suspension. If the equipment were new (i.e., equipment started operating on the same calendar date as the system start date), then the first occurrence of an event would be considered a failure. The first failure (for new equipment) or the first suspension (for non-new equipment) for each component will be calculated as:

    Time to First Occurrence of Event = Date/Time of First Failure - System Start Date/Time

    (4)
  4. Enter the operating shift schedule of the system. To do this, click the Shift Pattern icon on the control panel. This setting indicates how many hours a day are accumulated by the equipment. In this example, the equipment was scheduled to run two shifts, one from 8 AM to 5 PM, and another from 5 PM to 12 AM, for seven days a week. The following figure shows the shift schedule.

  5. Select a failure distribution and a repair distribution for the components to be analyzed. The Analysis page of the control panel displays these options, as shown next.

  6. The "quick defaults" setting is to use the 2-parameter Weibull with maximum likelihood estimation when there are two or more unique failure or repair times, or 1-parameter exponential when there is only one unique failure or repair time. In this example, we have a small data set and few suspensions, so select the Prefer RRX if sufficient data option and select all available distributions to be considered for the data set. The software will evaluate the fit of each selected distribution to the times-to-failure and times-to-repair data sets and then the results will be based on the one that provides the best fit.

  7. On the Main page of the control panel, click the Calculate icon to analyze the data set. The following report shows the results of the analysis.

Now that the failure and repair distributions of the equipment have been estimated, the results can be transferred to BlockSim and used in reliability block diagrams.

Reliability Block Diagrams

To export the results of the event log folio to BlockSim, click the Transfer to BlockSim Template icon on the control panel of the event log folio. You will be prompted to select which block to transfer, as shown next. (Each block contains the estimated failure and repair distributions of its corresponding component.)

Select all the blocks and click OK. Next, you will be prompted to specify a name and location to save the new BlockSim file. When you open the new file in BlockSim, it will include a template that contains all six blocks of the equipment, as shown next.

Now you can use the blocks in the template to create a reliability block diagram, run a simulation of the equipment and obtain results, such as the probability of failure in a given period of time and the average availability. For example, the following figure shows an RBD where the equipment is represented by two blocks (subsystems A and B) and each block has a subdiagram of its corresponding components.

As an example, a simulation of the equipment shows that for 5,840 hours of operation, the average availability is calculated to be 96.18%, as displayed next. Note that your results may vary, depending on the settings of the simulation.

Conclusion

This article described the process of using the event log folio in Weibull++ in order to convert maintenance log records to times-to-failure and times-to-repair data and then create failure and repair distributions. Once those distributions were obtained, the analysis results were transferred to BlockSim where the performance of the system can be evaluated using simulation.