An Example of a Root Cause Analysis

Before you can begin conducting an RCA, you must establish the criteria that standardizes when an RCA should be performed. For example, the manager of the plant, where a cooling water pump exists, might specify criteria as:

• If any pump fails one or more times per week for more than two successive weeks, regardless of the cost for repairs, an RCA study should be conducted.

  -or-

• If any pump fails catastrophically (e.g., causes the system to shut off), an RCA should be conducted.

Now, assume that in January 2015, the bearing temperature alarm, on the outboard bearing of a cooling water pump, sounded. By the time the operator arrived, the motor bearing temperature had increased to a point that the protective bearing temperature circuit caused the pump to shut down, resulting in loss of production. According to the criteria that was developed, these conditions dictate that an RCA should be performed. The first step is to create the analysis.

If water pump failure meets the criteria for requiring an RCA, the first step is to create the analysis. The user who will serve as the Principal Analyst can complete this step and initiate the development of the RCA for the recurring failures of the cooling water pump.

1. Access the RCA Overview page.

2. In the upper-right corner of the page, select New Analysis.

   A blank datasheet appears in the Analysis Summary workspace.

   

3. Enter the values in the fields on the datasheet. For this example, you should create root cause analysis that contains the information as it appears in the following table:

   Field	Value
   Analysis Name	Cooling Water Pump
   Analysis Description	An analysis to determine the root cause of the pump overheating.
   Start Date	01/06/2018
   End Date	01/15/2018
   Comments	Comments
   Analysis Type	Mechanical
   Event Narrative	Investigate what caused the motor temperature in the water pump to increase and overheat.
   Cost	20,000
   Frequency	1 Number 1 represents the number of failures that occurred for the cooling water pump

4. Select .

   The Root Cause Analysis is saved.

1. In the workspace, select the Linked Assets tab.

   The Linked Assets section appears.

   

2. Select .

   The Asset Finder window appears.

   

3. Search for PUMP and select it.

   PUMP is listed in the Linked Assets section.

1. Access the Team Members section.

2. Select .

   The list of human resources appears in the left section and the list of team members appears in the right section.

   

3. From the list of human resources, select the check box to the left of the following names to add as team members:

   • The analyst, Suzie Jones, to help gather information surrounding the failure.

   • The mechanic, Pete Handy, because he will repair the cooling water pump.

   • The engineer, Bob Speck, because he is an expert on systems and safety guidelines.

4. Select .

   The selected names are added to the list of team members in the right section.

   The Principal Analyst is Jane Doe, who is an expert on facilitating team discussion and documenting discussion notes.

   Note: By default, the human resource that created an analysis will be added automatically as a team member and assigned the Principal Analyst role. You can change the Principal Analyst for an RCA.

1. In the workspace, select the Critical Success Factors tab.

   The CSF section appears.

   

2. In the upper-left corner of the workspace, select .

   The Available CSFs and the CSFs for this Analysis subsections appear. The Available CSFs subsection contains the following Critical Success Factors that an administrative user has pre-configured for use with RCA:

   • A cross-functional section of plant personnel/experts will participate in the analysis.
   • A disciplined RCFA approach will be utilized.
   • A measurement process will be used to track the progress of approved recommendations.
   • All analysis hypothesis will be verified or disapproved.
   • Management agrees to fairly evaluate the analysis team findings and recommendations.
   • No one will be disciplined for honest mistakes.

   

1. In the workspace, select the Definition tab.

   The Definition section appears.

2. In the Charter field, enter the following text:

   To identify the root causes of the cooling water pump failure. This includes identifying deficiencies in or lack of management systems. Necessary recommendations for root causes will be communicated to the management for rapid resolutions.

3. Select .

   The Charter field is updated.

You can preserve failure event data by creating an RCA Preserve Record, which contains a data-collection task that should be completed by the team member to whom the Preserve Record is assigned. Each task can be categorized according to the 5 P's (i.e., Parts, Paper, People, Position, Paradigms), which indicates the type of data that should be collected.

1. In the left pane, select the Preserve Records tab.

   The Preserve Records workspace appears.

   

2. In the upper-left corner of the workspace, select .

   The New Preserve Record window appears.

   

3. In the Assigned to list, select the team member to whom you want to assign the Preserve Records, and then enter values in the available fields on the Preserve Records datasheet.

   For this example, you should create RCA Preserve records that contain the information as it appears in each of the following tables:

   RCA Preserve Record 1
   Field	Value (s)
   Assigned To	Pete Handy
   Data Category	Parts
   Data to Collect	The pieces of the cooling water pump that could have caused the pump to overheat.
   Collection Strategy	Document (pictures and observations) the condition of the components of the pump.
   Date Due	01/08/2018
   Send Alert on Due Date?	True
   Days Before Due Date to Be Notified	1
   Frequency of Alert After Due Date	Daily
   Alert Email Body	This is a reminder to gather and document the pieces of the pump that could have caused the pump to overheat and report back to the team.
   Completed?	No
   Date Completed	N/A

    

   RCA Preserve 2
   Field	Value (s)
   Assigned To	Suzie Jones
   Data Category	Paper
   Data to Collect	Standard operating procedures and frequent failure information
   Collection Strategy	Research
   Date Due	01/9/2018
   Send Alert on Due Date?	True
   Days Before Due Date to Be Notified	1
   Frequency of Alert After Due Date	Daily
   Alert Email Body	This is a reminder to gather the standard operating procedures and frequent failure information and report back to the team.
   Completed?	No
   Date Completed	N/A

    

   RCA Preserve Record 3
   Field	Value (s)
   Assigned To	Bob Speck
   Data Category	People
   Data to Collect	Operating procedures used by those who maintain the pump.
   Collection Strategy	Interview the floor operators who are assigned to the pump.
   Date Due	01/10/2018
   Send Alert on Due Date?	True
   Days Before Due Date to Be Notified	1
   Frequency of Alert After Due Date	Daily
   Alert Email Body	This is a reminder to interview operators assigned to the pump and document their operating procedures and report back to the team.
   Completed?	No
   Date Completed	N/A

    

   RCA Preserve Record 4
   Field	Value (s)
   Assign To	Bob Speck
   Data Category	People
   Data to Collect	Environment surrounding the pump.
   Collection Strategy	Tools, pictures, diagrams
   Date Due	01/10/2018
   Send Alert on Due Date?	True
   Days Before Due Date to Be Notified	1
   Frequency of Alert After Due Date	Daily
   Alert Email Body	This is a reminder to investigate the environment where the pump exists and report back to the team.
   Completed?	No
   Date Completed	N/A

    

   RCA Preserve Record 5
   Field	Value (s)
   Assigned To	Jane Doe (Principal Analyst)
   Data Category	Paradigms
   Data to Collect	Common assumptions made when a pump fails.
   Collection Strategy	Research and interview.
   Date Due	01/10/2018
   Send Alert on Due Date?	False
   Completed?	No
   Date Completed	N/A

After each team member has collected the data that is specified in the RCA Preserve Record, they should record their findings for the rest of the team through artifacts, including Reference Documents and hard copies of documents. After all the RCA Preserve Records have been completed, the team can begin to analyze the data.

After the team has preserved the failure data, the team can begin to analyze the data. To help you visualize the event and organize the team's hypotheses and analysis, you should create an Event Diagram or a Logic Tree depending on your requirement. For our example, we will create a Logic Tree diagram.

1. In the left pane, select the Logic Tree tab.

   A blank datasheet appears.

   

2. Enter the following values:

   Field	Value (s)
   Event End Date	01/15/2018
   Event ID	RCA-01/15/2018 Note: The Event ID field is populated automatically with the date the is created and is disabled so that value cannot be changed. If you choose to re-create this example RCA Event, your Event ID value will be different from the value listed here.
   Event Start Date	01/06/2018
   Label	Overheating in Pump 1
   Long Description	Investigate the cause of a cooling water pump to overheat and shut down the system.

3. Select .

   The Failure Event node appears in the design canvas in the Logic Tree workspace.

4. In the upper portion of the workspace, in the Palette section, select , and then use the drag-and-drop method to add the icon to the Failure Event node in the design canvas.

5. Select the Failure Mode that you added.

   The Properties window appears.

6. Enter the following information to define the Failure Mode, and then select :

   RCA Failure Mode ID	Value (s)
   No Spare Parts	Label: No Spare Parts Description: No spare parts caused the pump to be out of commission for a longer amount of time.

7. Add another Failure Mode using the following information:

   RCA Failure Mode ID	Value (s)
   Motor tripped due to high temperature.	Label: Motor tripped due to high temperature. Description: The motor shut off because the temperature became too hot and the system was shut down.

8. Diagram the events that led up to and immediately followed the failure event. For this example, you would create the Event Diagram as shown in the following image. The team should note the trend in the failures that occurred in the past. In this way, the team may be able to identify a pattern in the events prior to the RCA that may indicate the underlying cause(s) for these failures.

   

   This example contains nine RCA Sequence Nodes, which are displayed on the Event Diagram.

   Sequence Nodes in Event Diagram

   For each RCA Sequence Node, the following table lists the type of node and the values that exist in the RCA Sequence Node.

   Type of Node	Values
   Start/Finish	Label: Pump place into service Event Start Date: 1/1/2014 12:00 AM Event End Date: 1/1/2014 12:00 AM
   Process	Label: Cooling water pump turned on daily Event Start Date: 1/1/2014 12:00 AM Event End Date: 1/6/2015 12:00 AM
   Operation	Label: Bearing Temperature Alarm sounded Event Start Date: 1/6/2015 12:00 AM Event End Date: 1/6/2015 12:00 AM
   Operation	Label: Pump Overheated Event Start Date: 1/6/2015 12:00 AM Event End Date: 1/6/2015 12:00 AM
   Operation	Label: System shut down Event Start Date: 1/6/2015 12:00 AM Event End Date: 1/6/2015 12:00 AM
   Decision	Label: Decided to perform an analysis Event Start Date: 1/6/2015 12:00 AM Event End Date: 1/6/2015 12:00 AM
   Start/Finish	Label: RCA Study began Event Start Date: 1/6/2015 12:00 AM Event End Date: 1/16/2015 12:00 AM

9. At the top of the workspace, in the Palette section, select above the design canvas, and then use the drag-and-drop method to add the icon to the Hypothesis in the design canvas.

10. Select the Hypothesis that you added.

    The Properties window appears.

11. Enter the following information to define each of the Hypotheses required for our example.

    RCA Hypothesis ID	Value (s)
    Level 1
    Mechanical	Label: Mechanical Descriptions: The increase in temperature was due to a mechanical failure. Confidence Factor: Likely (3) State: Hypothesis True
    Electrical	Label: Electrical Descriptions: The increase in temperature was due to an electrical failure. Confidence Factor: Possibly True (2) State: Hypothesis Not True
    Level 2
    Outboard bearing	Label: Outboard bearing Descriptions: The outboard bearing had a mechanical problem. Confidence Factor: Highly Likely (4) State: Hypothesis True
    Inboard bearing	Label: Inboard bearing Descriptions: The inboard bearing had a mechanical problem. Confidence Factor: Somewhat True (1) State: Hypothesis Not True
    Level 3
    Erosion	Label: Erosion Descriptions: The outboard bearing eroded after overuse. Confidence Factor: Somewhat True (1) State: Hypothesis Not True
    Corrosion	Label: Corrosion Descriptions: The outboard bearing corroded and was rusty. Confidence Factor: Somewhat True (1) State: Hypothesis Not True
    Fatigue	Label: Fatigue Descriptions: The outboard bearing was overheated due to a weakness of materials. Confidence Factor: Highly Likely (4) State: Hypothesis Not True
    Level 4
    Misalignment	Label: Misalignment Descriptions: The outboard bearing was not aligned correctly. Confidence Factor: Possibly True (2) State: Hypothesis Not True
    Loose Bolts	Label: Loose Bolts Descriptions: Bolts were loose and caused the outboard bearing to become loose. Confidence Factor: Possibly True (2) State: Hypothesis Not True
    Lubrication Issue	Label: Lubrication Issue Descriptions: The outboard bearing was not receiving the right amount of lubrication. Confidence Factor: Highly Likely (4) State: Cause Physical
    Level 5
    Leak	Label: Leak Descriptions: The lubricant was leaking from the pump. Confidence Factor: Possibly True (2) State: Hypothesis Not True
    Too Much Applied	Label: Too much applied Descriptions: There was too much lubricant, causing, the bearing to operate too quickly and overheat. Confidence Factor: Possibly True (2) State: Hypothesis Not True
    Too Little Applied	Label: Too little applied Descriptions: There was not enough lubricant, causing friction and an increase in temperature. Confidence Factor: Likely (3) State: Cause Human
    Level 6
    Poor Training	Label: Poor training Descriptions: Workers were not properly training on applying lubricant to the bearing. Confidence Factor: Possibly True (2) State: Hypothesis Not True
    Inadequate Procedures Documented	Label: Inadequate Procedures Documented Descriptions: Workers are referencing procedure documentation that is not clear. Confidence Factor: Possibly True (2) State: Hypothesis Not True
    Worker Fatigue	Label: Worker Fatigue Descriptions: Workers are overworked and are not able to perform at an acceptable level. Confidence Factor: Likely (3) State: Cause Physical
    Inadequate Accountability of Workers	Label: Inadequate Accountability of Workers Descriptions: Workers are not being held accountable for the maintenance procedures they are required to perform. Confidence Factor: Highly Likely (4) State: Cause Latent

    Diagram each failure mode in detail. For this example, you would create a Logic Tree like the one shown in the following image:

    

    In the Logic Tree in this image, you can see that two Failure Mode nodes are linked to the Failure Event node, and multiple Hypothesis nodes are linked to the Motor Tripped due to high temperature Failure Mode node.

    This example contains four RCA Logic Gates, which represent Logic Gate nodes and are used in the Logic Tree. You must enable Logic Gates to view them on the Logic Tree.

    Example Data in Logic Gates

    In the following table, the Logic Gates are numbered as they appear in the Logic Tree from left to right. For each RCA Logic Gate, the table displays the value in the Gate Type field of the RCA Logic Gate and identifying information about the node (e.g., RCA Failure Mode) to which it is linked.

    RCA Logic Gate	Value Specified	Linked To
    1	Gate Type: OR	RCA Failure Mode: Motor tripped due to high temperature RCA Hypothesis: Mechanical RCA Hypothesis: Electrical
    2	Gate Type:OR	RCA Failure Mode: Motor tripped due to high temperature RCA Hypothesis: Mechanical RCA Hypothesis: Inboard bearing RCA Hypothesis: Outboard bearing
    3	Gate Type:OR	RCA Failure Mode: Motor tripped due to high temperature RCA Hypothesis: Outboard bearing RCA Hypothesis: Corrosion RCA Hypothesis: Erosion RCA Hypothesis: Fatigue
    4	Gate Type:OR	RCA Failure Mode: Motor tripped due to high temperature RCA Hypothesis: Fatigue RCA Hypothesis: Loose Bolts RCA Hypothesis: Lubrication Issue RCA Hypothesis: Misalignment
    5	Gate Type:OR	RCA Failure Mode: Motor tripped due to high temperature RCA Hypothesis: Lubrication Issue RCA Hypothesis: Leak RCA Hypothesis: Too Little Applied RCA Hypothesis: Too Much Applied
    6	Gate Type:OR	RCA Failure Mode: Motor tripped due to high temperature RCA Hypothesis: Too Little Applied RCA Hypothesis: Inadequate Accountability of Workers RCA Hypothesis: Inadequate Procedures Documented RCA Hypothesis: Poor Training RCA Hypothesis: Worker Fatigue

    Now that several likely hypotheses have been introduced, the team will need to rule out hypotheses. To rule out a hypothesis or prove a hypothesis, the hypothesis will need to be tested against the data that is available to the team (i.e., the failure event data and the team of experts).

    As described in the next step, the Principal Analyst should create RCA Verifications and assign them to team members who will be responsible for completing the associated task.

12. Select the Hypothesis that you want to verify.

13. In the lower-left corner of the workspace, select the Verifications tab.

    The Verifications pane appears.

14. Above the list, select .

    The Create Verification pane appears.

    

15. In the Assigned to list, select the team member to whom you want to assign the Verification.

    A blank datasheet appears.

    

16. Enter the details in the available fields based on the following information to define each of the Verification required for our example, and then select .

    RCA Verification ID	Values Specified
    Inspect wiring to make sure that the cause was not electrical.	Due Date: 1/10/2015 Method: Inspect wiring to make sure that the cause was not electrical. Completed: No
    Perform a metallurgical analysis to determine why the bearing failed.	Due Date: 1/10/2015 Method: Perform a metallurgical analysis to determine why the bearing failed. Completed: False

    The following image shows an example of an RCA Verification that contains a task to test the Mechanical hypothesis.

    

    The task defined in this RCA Verification instructs the team member to inspect the wiring for the cooling water pump to ensure that the problem was not electrical, but mechanical.

    When a hypothesis has been tested, you must change the state of the associated RCA Hypothesis, as described in the next step.

17. Select the Hypothesis whose state you want to change.
18. At the top of the workspace, in the States panel, select one of the following options, and then select :
    • Hypothesis True if the hypothesis has been confirmed. After you determine a hypothesis to be true, you can identify the root cause of the Failure Event.
    • Hypothesis Not True if the hypothesis has been ruled out.

    For example, the following image shows the Logic Tree after Pete Handy has completed the RCA Verification to which he was assigned. Pete Handy concluded that the wiring did not the cause the pump to overheat, so the problem must be mechanical.

    

    After the state of the Mechanical RCA Hypothesis is set to True, the team can continue down the Logic Tree until they identify the root cause. Causes can be further identified as physical, human, or latent. For example, the following image shows the Logic Tree after ALL the RCA Verifications have been completed. The team has concluded that inadequate accountability of workers has led to insufficient lubricant being applied to the outboard bearing, which caused the motor's temperature to increase and the pump to overheat.

    

    In this case, the root cause for the pump failure would be identified as latent (i.e., inadequate accountability of workers), which lead to human (i.e., too little applied) and physical (i.e., lubrication issue) causes. Now that a hypothesis has been verified and the root cause has been determined, you can create RCA Recommendations to communicate your findings on how to prevent this failure in the future.

    If the hypothesis is not true but you cannot rule out the hypothesis and set the state to False, you can assign a probability to the hypothesis to include the team in determining the possibility of the hypothetical cause.

Now that the team has determined the root cause of the cooling water pump failure to be inadequate accountability of workers, you will want to record your findings and create RCA Recommendations. After recommendations have been implemented, you can share the results of the RCA with others. You can access the RCA Overview page to view the financial effects of implementing the recommendations. For our example, we will create one Recommendation record.

1. Access the Root Cause Analysis for which you want to create a recommendation.

2. In the upper-right corner the workspace, select the .

   The Recommendations pane appears.

   

3. Select .

   The RCA Recommendation datasheet appears.

4. On the General Information tab, enter the following details in the available fields:

   Field	Description
   Alert Responsible Party When Due?	True
   Analysis ID	Cooling Water Pump
   Assigned to Name	Suzie Jones
   Author Name	Jane Doe
   Business Impact	Production (PRODUCTION)
   Cause	Inadequate accountability of workers.
   Cause Type	Cause Latent
   Days Before Due Date To Be Alerted?	4
   Final Approver Name	Jane Doe
   Frequency of Alert After Due Date	Daily
   Implementation Alert Text	You have less than 4 days to complete the new required report.
   Mandatory Date	2/28/2015
   Recommendation Description	This report will require workers to document each time they perform a routine procedure, which will hold them accountable for pump maintenance.
   Recommendation Headline	Create a new performed maintenance report.
   Recommendation ID	REC-1376
   Recommendation Priority	Medium (Medium)
   Required Equipment Status	On-Line (ON-LINE)
   Reviewer Name	Jane Doe
   Status	Created (CREATED)
   Target Completion Date	Monday, February 20, 2015

5. If you want to schedule a recommendation alert email to be sent to a specified recipient, select the Alert tab, and enter the necessary information on the datasheet.

6. Select .

   The RCA Recommendation is saved to the database and linked to the analysis for the selected Root Cause analysis.

After you have successfully identified the root causes associated with the failure that was investigated, you can communicate the findings and propose recommendations to the appropriate individuals within your organization. You can communicate the findings via a hyperlink to the analysis or via a summary report through email to selected recipients.

After the RCA findings have been communicated and the recommendation has been implemented, the team will want to track the effectiveness of the recommendation to see if the new report has prevented the cooling water pump from failing thus far. Tracking for results involves creating RCA Tracking Items, scheduling the Tracking Evaluations in the GE Digital APM Schedule Manager, and modifying the existing recommendation or creating a new one based on the discoveries that are made through reevaluation.

For our example, we will create one tracking item.

1. Access the RCA within which you want to create a new Tracking Item.

2. In the left pane, select the Track tab.

   The Track workspace appears.

   

3. In the upper-left corner of the workspace, select .

   The Add/Edit Tracking Item window appears, displaying a blank datasheet.

   

4. Enter the following information in the available fields:

   Field	Description
   Active	True
   Start Date	3/1/2015
   Count	2
   Cost	10,000
   Alert	Williams, Bob Note: Williams, Bob will receive an e-mail that alerts him that the performance of the cooling pump has again fallen below the tolerable limits or that the corrective actions from the first RCA were ineffective.
   Message Subject	Cooling Water Pump Improvement
   Message Text	Cooling water pump 1 has failed and needs to be re-evaluated. The reason for the failure is that the performance of the cooling pump has again fallen below the tolerable limits or that the corrective actions from the first RCA were ineffective.

5. Select .

   The Tracking Item is saved.