All Member Open Forum

Did every successful web-search engine get it wrong? It seems very unlikely, doesn't it?

I am starting this discussion as a place for all of us to explore the merits and shortcomings of using a powerful text-based search engine for intelligence gathering as compared failure codes. 

I am very experienced at using both methods with some of the best companies in the world at documenting failures on work orders. My experience with the shortcomings of failure codes became my idea-seed for designing a text-based search in SQL for analyzing failures in work orders. SQL has its own limitations, however. 

I was extremely lucky to meet exactly the right people, at exactly the right times in my career to not only develop a SQL statement that works to find and record keywords in text, but later on develop VBA programming in Excel that was far more powerful and up to the task. I call my method Contextual Failure Analysis. For those who haven't seen my IMC presentation, I will attach it here.

Let the discussion go where it may. And I will be happy to show anyone interested what I have working and how it compares to Failure Codes. 

Thank you and best regards,
Jeff Wahl
Maintenance Program Manager, CMMS

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Continuing on my discussion - The absolute coolest thing about using a powerful text-based search engine for exposing failure trends is that I am able learn HOW people communicate with one another; THIS IS EXTREMELY IMPORTANT CONCEPT that should not be underestimated. I can say this with confidence because every study I have done, failure codes were used and continue to be used. For every failure trend I've identified I build pivot tables that prove my point. As a matter of fact, I like having failure codes,  if for no other reason than to show how poorly they work.

Here's a perfect example of what I've learned about HOW work orders are REALLY written AND how inconsistent we are at communicating. It's part of the human condition. The following list of action verbs were all used on work orders to convey that large-scale haul truck beds were inspected and replaced. Does anyone what to venture a guess as to how consistently their failure codes were used? The trickiest part of this analysis is the company didn't even bother to track truck beds in their inventory system. I know most of us have lived with stories just like this one.

Truck Bed Change Trend Study

Change

Install

Install bed

R & I

R & R

R &I

R and I

R and R

R& I

R&I

R&R

Remove

Remove & install

Remove and install

Remove bed

Replace

Use bed from

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Original Message:
Sent: 08-15-2019 10:36 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Did every successful web-search engine get it wrong? It seems very unlikely, doesn't it?

I am starting this discussion as a place for all of us to explore the merits and shortcomings of using a powerful text-based search engine for intelligence gathering as compared failure codes.

I am very experienced at using both methods with some of the best companies in the world at documenting failures on work orders. My experience with the shortcomings of failure codes became my idea-seed for designing a text-based search in SQL for analyzing failures in work orders. SQL has its own limitations, however.

I was extremely lucky to meet exactly the right people, at exactly the right times in my career to not only develop a SQL statement that works to find and record keywords in text, but later on develop VBA programming in Excel that was far more powerful and up to the task. I call my method Contextual Failure Analysis. For those who haven't seen my IMC presentation, I will attach it here.

Let the discussion go where it may. And I will be happy to show anyone interested what I have working and how it compares to Failure Codes.

Thank you and best regards,
Jeff Wahl
Maintenance Program Manager, CMMS

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Jeff, I've built an excel tool to do exactly the same thing with our CMMS text fields for operator remarks (what issues were they complaining about) and technician remarks (what actions do the techs have to take).

We too have poorly setup failure codes rendering them useless for analysis. There are way too many which means the technicians just pick the first one on every list. Because of the problems with failure codes you state, we use the word cloud tool.

It tells us what really are the failure modes we are getting which helps us adjust the PM system to attack and mitigate them vs the tribal method for PM content which is not based on FMEA nor failure mode data.


------------------------------
Eddie Lord
CMRP, TEC Methods and Reliability
Michelin
Lexington SC
------------------------------

Original Message:
Sent: 08-16-2019 01:24 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Continuing on my discussion - The absolute coolest thing about using a powerful text-based search engine for exposing failure trends is that I am able learn HOW people communicate with one another; THIS IS EXTREMELY IMPORTANT CONCEPT that should not be underestimated. I can say this with confidence because every study I have done, failure codes were used and continue to be used. For every failure trend I've identified I build pivot tables that prove my point. As a matter of fact, I like having failure codes,  if for no other reason than to show how poorly they work.

Here's a perfect example of what I've learned about HOW work orders are REALLY written AND how inconsistent we are at communicating. It's part of the human condition. The following list of action verbs were all used on work orders to convey that large-scale haul truck beds were inspected and replaced. Does anyone what to venture a guess as to how consistently their failure codes were used? The trickiest part of this analysis is the company didn't even bother to track truck beds in their inventory system. I know most of us have lived with stories just like this one.

Truck Bed Change Trend Study

Change

Install

Install bed

R & I

R & R

R &I

R and I

R and R

R& I

R&I

R&R

Remove

Remove & install

Remove and install

Remove bed

Replace

Use bed from

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO

Original Message:
Sent: 08-15-2019 10:36 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Did every successful web-search engine get it wrong? It seems very unlikely, doesn't it?

I am starting this discussion as a place for all of us to explore the merits and shortcomings of using a powerful text-based search engine for intelligence gathering as compared failure codes.

I am very experienced at using both methods with some of the best companies in the world at documenting failures on work orders. My experience with the shortcomings of failure codes became my idea-seed for designing a text-based search in SQL for analyzing failures in work orders. SQL has its own limitations, however.

I was extremely lucky to meet exactly the right people, at exactly the right times in my career to not only develop a SQL statement that works to find and record keywords in text, but later on develop VBA programming in Excel that was far more powerful and up to the task. I call my method Contextual Failure Analysis. For those who haven't seen my IMC presentation, I will attach it here.

Let the discussion go where it may. And I will be happy to show anyone interested what I have working and how it compares to Failure Codes.

Thank you and best regards,
Jeff Wahl
Maintenance Program Manager, CMMS

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Eddie - I'm looking forward to seeing what you have working. The tool sounds very much like a word-cloud programmed in VBA, by what I understand to be a university professor, for checking to see if his students were copying the work of others or plagiarizing their assignments from published works.  I downloaded a copy and it does work as described.

I like my method because I learn so much about the way people are communication with one another. The way work orders are written and comments that are made can lead to so many improvements along the way.

I remember bringing up an issue where our building costs seemed strange. Upon deeper exploration we found that some people were writing work orders to the building (facility-level) to get production work done to correct a pump leak (for example).

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Original Message:
Sent: 08-16-2019 07:05 AM
From: Eddie Lord
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Jeff, I've built an excel tool to do exactly the same thing with our CMMS text fields for operator remarks (what issues were they complaining about) and technician remarks (what actions do the techs have to take).

We too have poorly setup failure codes rendering them useless for analysis. There are way too many which means the technicians just pick the first one on every list. Because of the problems with failure codes you state, we use the word cloud tool.

It tells us what really are the failure modes we are getting which helps us adjust the PM system to attack and mitigate them vs the tribal method for PM content which is not based on FMEA nor failure mode data.


------------------------------
Eddie Lord
CMRP, TEC Methods and Reliability
Michelin
Lexington SC

Original Message:
Sent: 08-16-2019 01:24 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Continuing on my discussion - The absolute coolest thing about using a powerful text-based search engine for exposing failure trends is that I am able learn HOW people communicate with one another; THIS IS EXTREMELY IMPORTANT CONCEPT that should not be underestimated. I can say this with confidence because every study I have done, failure codes were used and continue to be used. For every failure trend I've identified I build pivot tables that prove my point. As a matter of fact, I like having failure codes,  if for no other reason than to show how poorly they work.

Here's a perfect example of what I've learned about HOW work orders are REALLY written AND how inconsistent we are at communicating. It's part of the human condition. The following list of action verbs were all used on work orders to convey that large-scale haul truck beds were inspected and replaced. Does anyone what to venture a guess as to how consistently their failure codes were used? The trickiest part of this analysis is the company didn't even bother to track truck beds in their inventory system. I know most of us have lived with stories just like this one.

Truck Bed Change Trend Study

Change

Install

Install bed

R & I

R & R

R &I

R and I

R and R

R& I

R&I

R&R

Remove

Remove & install

Remove and install

Remove bed

Replace

Use bed from

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO

Original Message:
Sent: 08-15-2019 10:36 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Did every successful web-search engine get it wrong? It seems very unlikely, doesn't it?

I am starting this discussion as a place for all of us to explore the merits and shortcomings of using a powerful text-based search engine for intelligence gathering as compared failure codes.

I am very experienced at using both methods with some of the best companies in the world at documenting failures on work orders. My experience with the shortcomings of failure codes became my idea-seed for designing a text-based search in SQL for analyzing failures in work orders. SQL has its own limitations, however.

I was extremely lucky to meet exactly the right people, at exactly the right times in my career to not only develop a SQL statement that works to find and record keywords in text, but later on develop VBA programming in Excel that was far more powerful and up to the task. I call my method Contextual Failure Analysis. For those who haven't seen my IMC presentation, I will attach it here.

Let the discussion go where it may. And I will be happy to show anyone interested what I have working and how it compares to Failure Codes.

Thank you and best regards,
Jeff Wahl
Maintenance Program Manager, CMMS

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

That's right Jeff, the code searches every word in the text fields (each type - operator requests, technician comments) and produces a list of the top ten words encountered for each, ranked in order, along with pareto graphs and all the work orders for each of those top ten words. (I chose ten)
The user selects the date range and scope of the search to be either broad or narrow.

I love it because the tool shows us what the failures and actions are, based on the words people use. It is eye opening and valuable information because it isn't skewed by predefined codes, it is the true picture.

I can't actually share the code itself because the company owns everything we develop.

------------------------------
Eddie Lord
CMRP, TEC Methods and Reliability
Lexington SC
------------------------------

Original Message:
Sent: 08-26-2019 03:37 PM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Eddie - I'm looking forward to seeing what you have working. The tool sounds very much like a word-cloud programmed in VBA, by what I understand to be a university professor, for checking to see if his students were copying the work of others or plagiarizing their assignments from published works.  I downloaded a copy and it does work as described.

I like my method because I learn so much about the way people are communication with one another. The way work orders are written and comments that are made can lead to so many improvements along the way.

I remember bringing up an issue where our building costs seemed strange. Upon deeper exploration we found that some people were writing work orders to the building (facility-level) to get production work done to correct a pump leak (for example).

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO

Original Message:
Sent: 08-16-2019 07:05 AM
From: Eddie Lord
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Jeff, I've built an excel tool to do exactly the same thing with our CMMS text fields for operator remarks (what issues were they complaining about) and technician remarks (what actions do the techs have to take).

We too have poorly setup failure codes rendering them useless for analysis. There are way too many which means the technicians just pick the first one on every list. Because of the problems with failure codes you state, we use the word cloud tool.

It tells us what really are the failure modes we are getting which helps us adjust the PM system to attack and mitigate them vs the tribal method for PM content which is not based on FMEA nor failure mode data.


------------------------------
Eddie Lord
CMRP, TEC Methods and Reliability
Michelin
Lexington SC

Original Message:
Sent: 08-16-2019 01:24 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Continuing on my discussion - The absolute coolest thing about using a powerful text-based search engine for exposing failure trends is that I am able learn HOW people communicate with one another; THIS IS EXTREMELY IMPORTANT CONCEPT that should not be underestimated. I can say this with confidence because every study I have done, failure codes were used and continue to be used. For every failure trend I've identified I build pivot tables that prove my point. As a matter of fact, I like having failure codes,  if for no other reason than to show how poorly they work.

Here's a perfect example of what I've learned about HOW work orders are REALLY written AND how inconsistent we are at communicating. It's part of the human condition. The following list of action verbs were all used on work orders to convey that large-scale haul truck beds were inspected and replaced. Does anyone what to venture a guess as to how consistently their failure codes were used? The trickiest part of this analysis is the company didn't even bother to track truck beds in their inventory system. I know most of us have lived with stories just like this one.

Truck Bed Change Trend Study

Change

Install

Install bed

R & I

R & R

R &I

R and I

R and R

R& I

R&I

R&R

Remove

Remove & install

Remove and install

Remove bed

Replace

Use bed from

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO

Original Message:
Sent: 08-15-2019 10:36 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Did every successful web-search engine get it wrong? It seems very unlikely, doesn't it?

I am starting this discussion as a place for all of us to explore the merits and shortcomings of using a powerful text-based search engine for intelligence gathering as compared failure codes.

I am very experienced at using both methods with some of the best companies in the world at documenting failures on work orders. My experience with the shortcomings of failure codes became my idea-seed for designing a text-based search in SQL for analyzing failures in work orders. SQL has its own limitations, however.

I was extremely lucky to meet exactly the right people, at exactly the right times in my career to not only develop a SQL statement that works to find and record keywords in text, but later on develop VBA programming in Excel that was far more powerful and up to the task. I call my method Contextual Failure Analysis. For those who haven't seen my IMC presentation, I will attach it here.

Let the discussion go where it may. And I will be happy to show anyone interested what I have working and how it compares to Failure Codes.

Thank you and best regards,
Jeff Wahl
Maintenance Program Manager, CMMS

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

I understand not being able to show the code - nor would I be inclined to share it. I would be more than happy to show you how well my development works as the user though. I found that I can remove the special sauce for sharing only the results of my analysis. I found that I can add it back in too for additional analysis. How cool is that? 

I didn't make the offer for the purpose of taking any ideas from you or your company. My work is my work and I love it as it is. The only other improvement I could possibly make to make it work better is to add weighting to the keywords. It really isn't need though. I have years worth of keywords that were used to find failures. I catalog them into an examples tab.

I trained a friend in the USAF to use it for his analyses. For one analysis we took the full collective set of manufacturing keywords and removed words they just didn't use, substituted words they did use, and added words that were specific to their operations. It helps to have my starting point for people get getting used to the methodology.

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Original Message:
Sent: 08-26-2019 04:13 PM
From: Eddie Lord
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

That's right Jeff, the code searches every word in the text fields (each type - operator requests, technician comments) and produces a list of the top ten words encountered for each, ranked in order, along with pareto graphs and all the work orders for each of those top ten words. (I chose ten)
The user selects the date range and scope of the search to be either broad or narrow.

I love it because the tool shows us what the failures and actions are, based on the words people use. It is eye opening and valuable information because it isn't skewed by predefined codes, it is the true picture.

I can't actually share the code itself because the company owns everything we develop.

------------------------------
Eddie Lord
CMRP, TEC Methods and Reliability
Lexington SC

Original Message:
Sent: 08-26-2019 03:37 PM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Eddie - I'm looking forward to seeing what you have working. The tool sounds very much like a word-cloud programmed in VBA, by what I understand to be a university professor, for checking to see if his students were copying the work of others or plagiarizing their assignments from published works.  I downloaded a copy and it does work as described.

I like my method because I learn so much about the way people are communication with one another. The way work orders are written and comments that are made can lead to so many improvements along the way.

I remember bringing up an issue where our building costs seemed strange. Upon deeper exploration we found that some people were writing work orders to the building (facility-level) to get production work done to correct a pump leak (for example).

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO

Original Message:
Sent: 08-16-2019 07:05 AM
From: Eddie Lord
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Jeff, I've built an excel tool to do exactly the same thing with our CMMS text fields for operator remarks (what issues were they complaining about) and technician remarks (what actions do the techs have to take).

We too have poorly setup failure codes rendering them useless for analysis. There are way too many which means the technicians just pick the first one on every list. Because of the problems with failure codes you state, we use the word cloud tool.

It tells us what really are the failure modes we are getting which helps us adjust the PM system to attack and mitigate them vs the tribal method for PM content which is not based on FMEA nor failure mode data.


------------------------------
Eddie Lord
CMRP, TEC Methods and Reliability
Michelin
Lexington SC

Original Message:
Sent: 08-16-2019 01:24 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Continuing on my discussion - The absolute coolest thing about using a powerful text-based search engine for exposing failure trends is that I am able learn HOW people communicate with one another; THIS IS EXTREMELY IMPORTANT CONCEPT that should not be underestimated. I can say this with confidence because every study I have done, failure codes were used and continue to be used. For every failure trend I've identified I build pivot tables that prove my point. As a matter of fact, I like having failure codes,  if for no other reason than to show how poorly they work.

Here's a perfect example of what I've learned about HOW work orders are REALLY written AND how inconsistent we are at communicating. It's part of the human condition. The following list of action verbs were all used on work orders to convey that large-scale haul truck beds were inspected and replaced. Does anyone what to venture a guess as to how consistently their failure codes were used? The trickiest part of this analysis is the company didn't even bother to track truck beds in their inventory system. I know most of us have lived with stories just like this one.

Truck Bed Change Trend Study

Change

Install

Install bed

R & I

R & R

R &I

R and I

R and R

R& I

R&I

R&R

Remove

Remove & install

Remove and install

Remove bed

Replace

Use bed from

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO

Original Message:
Sent: 08-15-2019 10:36 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Did every successful web-search engine get it wrong? It seems very unlikely, doesn't it?

I am starting this discussion as a place for all of us to explore the merits and shortcomings of using a powerful text-based search engine for intelligence gathering as compared failure codes.

I am very experienced at using both methods with some of the best companies in the world at documenting failures on work orders. My experience with the shortcomings of failure codes became my idea-seed for designing a text-based search in SQL for analyzing failures in work orders. SQL has its own limitations, however.

I was extremely lucky to meet exactly the right people, at exactly the right times in my career to not only develop a SQL statement that works to find and record keywords in text, but later on develop VBA programming in Excel that was far more powerful and up to the task. I call my method Contextual Failure Analysis. For those who haven't seen my IMC presentation, I will attach it here.

Let the discussion go where it may. And I will be happy to show anyone interested what I have working and how it compares to Failure Codes.

Thank you and best regards,
Jeff Wahl
Maintenance Program Manager, CMMS

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

I never imagined your interest as anything other than 'that's cool man!'. Thank you for sharing and sparking the discussion.

------------------------------
Eddie Lord
CMRP, TEC Methods and Reliability
Lexington SC
------------------------------

Original Message:
Sent: 08-26-2019 04:39 PM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

I understand not being able to show the code - nor would I be inclined to share it. I would be more than happy to show you how well my development works as the user though. I found that I can remove the special sauce for sharing only the results of my analysis. I found that I can add it back in too for additional analysis. How cool is that?

I didn't make the offer for the purpose of taking any ideas from you or your company. My work is my work and I love it as it is. The only other improvement I could possibly make to make it work better is to add weighting to the keywords. It really isn't need though. I have years worth of keywords that were used to find failures. I catalog them into an examples tab.

I trained a friend in the USAF to use it for his analyses. For one analysis we took the full collective set of manufacturing keywords and removed words they just didn't use, substituted words they did use, and added words that were specific to their operations. It helps to have my starting point for people get getting used to the methodology.

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO

Original Message:
Sent: 08-26-2019 04:13 PM
From: Eddie Lord
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

That's right Jeff, the code searches every word in the text fields (each type - operator requests, technician comments) and produces a list of the top ten words encountered for each, ranked in order, along with pareto graphs and all the work orders for each of those top ten words. (I chose ten)
The user selects the date range and scope of the search to be either broad or narrow.

I love it because the tool shows us what the failures and actions are, based on the words people use. It is eye opening and valuable information because it isn't skewed by predefined codes, it is the true picture.

I can't actually share the code itself because the company owns everything we develop.

------------------------------
Eddie Lord
CMRP, TEC Methods and Reliability
Lexington SC

Original Message:
Sent: 08-26-2019 03:37 PM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Eddie - I'm looking forward to seeing what you have working. The tool sounds very much like a word-cloud programmed in VBA, by what I understand to be a university professor, for checking to see if his students were copying the work of others or plagiarizing their assignments from published works.  I downloaded a copy and it does work as described.

I like my method because I learn so much about the way people are communication with one another. The way work orders are written and comments that are made can lead to so many improvements along the way.

I remember bringing up an issue where our building costs seemed strange. Upon deeper exploration we found that some people were writing work orders to the building (facility-level) to get production work done to correct a pump leak (for example).

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO

Original Message:
Sent: 08-16-2019 07:05 AM
From: Eddie Lord
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Jeff, I've built an excel tool to do exactly the same thing with our CMMS text fields for operator remarks (what issues were they complaining about) and technician remarks (what actions do the techs have to take).

We too have poorly setup failure codes rendering them useless for analysis. There are way too many which means the technicians just pick the first one on every list. Because of the problems with failure codes you state, we use the word cloud tool.

It tells us what really are the failure modes we are getting which helps us adjust the PM system to attack and mitigate them vs the tribal method for PM content which is not based on FMEA nor failure mode data.


------------------------------
Eddie Lord
CMRP, TEC Methods and Reliability
Michelin
Lexington SC

Original Message:
Sent: 08-16-2019 01:24 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Continuing on my discussion - The absolute coolest thing about using a powerful text-based search engine for exposing failure trends is that I am able learn HOW people communicate with one another; THIS IS EXTREMELY IMPORTANT CONCEPT that should not be underestimated. I can say this with confidence because every study I have done, failure codes were used and continue to be used. For every failure trend I've identified I build pivot tables that prove my point. As a matter of fact, I like having failure codes,  if for no other reason than to show how poorly they work.

Here's a perfect example of what I've learned about HOW work orders are REALLY written AND how inconsistent we are at communicating. It's part of the human condition. The following list of action verbs were all used on work orders to convey that large-scale haul truck beds were inspected and replaced. Does anyone what to venture a guess as to how consistently their failure codes were used? The trickiest part of this analysis is the company didn't even bother to track truck beds in their inventory system. I know most of us have lived with stories just like this one.

Truck Bed Change Trend Study

Change

Install

Install bed

R & I

R & R

R &I

R and I

R and R

R& I

R&I

R&R

Remove

Remove & install

Remove and install

Remove bed

Replace

Use bed from

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO

Original Message:
Sent: 08-15-2019 10:36 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Did every successful web-search engine get it wrong? It seems very unlikely, doesn't it?

I am starting this discussion as a place for all of us to explore the merits and shortcomings of using a powerful text-based search engine for intelligence gathering as compared failure codes.

I am very experienced at using both methods with some of the best companies in the world at documenting failures on work orders. My experience with the shortcomings of failure codes became my idea-seed for designing a text-based search in SQL for analyzing failures in work orders. SQL has its own limitations, however.

I was extremely lucky to meet exactly the right people, at exactly the right times in my career to not only develop a SQL statement that works to find and record keywords in text, but later on develop VBA programming in Excel that was far more powerful and up to the task. I call my method Contextual Failure Analysis. For those who haven't seen my IMC presentation, I will attach it here.

Let the discussion go where it may. And I will be happy to show anyone interested what I have working and how it compares to Failure Codes.

Thank you and best regards,
Jeff Wahl
Maintenance Program Manager, CMMS

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Eddie,

Can you explain the word cloud tool your referring to? I'm revising our failure codes in our system and am interested how you used this tool.


------------------------------
Aaron Carter
CMMS Administrator
New Albany OH
------------------------------

Original Message:
Sent: 08-16-2019 07:05 AM
From: Eddie Lord
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Jeff, I've built an excel tool to do exactly the same thing with our CMMS text fields for operator remarks (what issues were they complaining about) and technician remarks (what actions do the techs have to take).

We too have poorly setup failure codes rendering them useless for analysis. There are way too many which means the technicians just pick the first one on every list. Because of the problems with failure codes you state, we use the word cloud tool.

It tells us what really are the failure modes we are getting which helps us adjust the PM system to attack and mitigate them vs the tribal method for PM content which is not based on FMEA nor failure mode data.


------------------------------
Eddie Lord
CMRP, TEC Methods and Reliability
Michelin
Lexington SC

Original Message:
Sent: 08-16-2019 01:24 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Continuing on my discussion - The absolute coolest thing about using a powerful text-based search engine for exposing failure trends is that I am able learn HOW people communicate with one another; THIS IS EXTREMELY IMPORTANT CONCEPT that should not be underestimated. I can say this with confidence because every study I have done, failure codes were used and continue to be used. For every failure trend I've identified I build pivot tables that prove my point. As a matter of fact, I like having failure codes,  if for no other reason than to show how poorly they work.

Here's a perfect example of what I've learned about HOW work orders are REALLY written AND how inconsistent we are at communicating. It's part of the human condition. The following list of action verbs were all used on work orders to convey that large-scale haul truck beds were inspected and replaced. Does anyone what to venture a guess as to how consistently their failure codes were used? The trickiest part of this analysis is the company didn't even bother to track truck beds in their inventory system. I know most of us have lived with stories just like this one.

Truck Bed Change Trend Study

Change

Install

Install bed

R & I

R & R

R &I

R and I

R and R

R& I

R&I

R&R

Remove

Remove & install

Remove and install

Remove bed

Replace

Use bed from

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO

Original Message:
Sent: 08-15-2019 10:36 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Did every successful web-search engine get it wrong? It seems very unlikely, doesn't it?

I am starting this discussion as a place for all of us to explore the merits and shortcomings of using a powerful text-based search engine for intelligence gathering as compared failure codes.

I am very experienced at using both methods with some of the best companies in the world at documenting failures on work orders. My experience with the shortcomings of failure codes became my idea-seed for designing a text-based search in SQL for analyzing failures in work orders. SQL has its own limitations, however.

I was extremely lucky to meet exactly the right people, at exactly the right times in my career to not only develop a SQL statement that works to find and record keywords in text, but later on develop VBA programming in Excel that was far more powerful and up to the task. I call my method Contextual Failure Analysis. For those who haven't seen my IMC presentation, I will attach it here.

Let the discussion go where it may. And I will be happy to show anyone interested what I have working and how it compares to Failure Codes.

Thank you and best regards,
Jeff Wahl
Maintenance Program Manager, CMMS

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Original Message------

Eddie,

Can you explain the word cloud tool your referring to? I'm revising our failure codes in our system and am interested how you used this tool.


------------------------------
Aaron Carter
CMMS Administrator
New Albany OH
------------------------------

Good day Richard - I am looking forward to seeing how many people try following your slides and how many people run R on their actual work order data. Running R on your sample data is one thing, but in order to be truly useful, R must prove itself with meaningful implemented improvements. Anyone? I'm looking forward to hearing your feedback.

Cheers,

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Original Message:
Sent: 09-23-2019 11:37 AM
From: Richard Lamb
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Aaron, you seem adventurous. Thus, if you want to go down the power road I point you to working the problem with machine learning and AI upon the learning rather than stop at the visualization clouds are. 

You will need to work with the following:

• Machine Learning with R, Brett Lantz, Chapter 4.
• Slide set explaining how to work with "R" such that you can jump into the chapter. Download at https://analytics4strategy.com/trn-rhandson 

Enjoy

Original Message------

Eddie,

Can you explain the word cloud tool your referring to? I'm revising our failure codes in our system and am interested how you used this tool.


------------------------------
Aaron Carter
CMMS Administrator
New Albany OH
------------------------------

Original Message------

Good day Richard - I am looking forward to seeing how many people try following your slides and how many people run R on their actual work order data. Running R on your sample data is one thing, but in order to be truly useful, R must prove itself with meaningful implemented improvements. Anyone? I'm looking forward to hearing your feedback.

Cheers,

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Yes, but... SQL and Excel, in combination, have more than enough horsepower to run data through for optimization programs. I ran a relational database software call "Brio Intelligence" that was the most flexible SQL software out there. Brio was far beyond anything CMMSes offer in the box as their toolset.

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Original Message:
Sent: 09-23-2019 02:25 PM
From: Richard Lamb
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Jeff, it is not a matter of R proving itself anymore than my tool box having to prove itself. R is part of the tool box for bringing the management of our assets and processes to be fully data-driven because by doing so, we finally can accomplish what we've wanted to dating back to the 1950s when the field was developed.

Recall, that the webinar on building super tables with our CMMS data was for the purpose of building the super tables to import to the analytics that "R" allows us to get at. Slides at https://analytics4strategy.com/train-builddatatables

Original Message------

Good day Richard - I am looking forward to seeing how many people try following your slides and how many people run R on their actual work order data. Running R on your sample data is one thing, but in order to be truly useful, R must prove itself with meaningful implemented improvements. Anyone? I'm looking forward to hearing your feedback.

Cheers,

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Jeff, this is fantastic.

You did a great job articulating many points, particularly on failure code designs, the time spent coding data as the data is getting entered, filling in missing codes using the free text, and in this thread how you learned the added benefit of gaining insight in how people communicated, all pointing to the observation "text-based analytics is a more natural way to find failures".  

I've been working with data from the CMMS/EAM for almost 5 years now, developing ways that it can be useful in maintenance and reliability work processes, and have long ago come to the conclusion that you need to use the text field.  For every reason Jeff articulated.  One thing I do is try to develop ways to apply a lot of current research to actual applications in practical ways.  I think it's worth mentioning a current research effort out there about: due to all the observations Jeff had, why don't we actually revisit how we capture data in the maintenance management system?

Their first paper goes about that by essentially taking an FMEA approach to measure sources of human unreliability in the maintenance management work process itself: 
https://ws680.nist.gov/publication/get_pdf.cfm?pub_id=927498

There's lots of references to SMRP in there too :)




------------------------------
Sarah Lukens
GE Digital
Roanoke VA
------------------------------

Original Message:
Sent: 08-15-2019 10:36 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Did every successful web-search engine get it wrong? It seems very unlikely, doesn't it?

I am starting this discussion as a place for all of us to explore the merits and shortcomings of using a powerful text-based search engine for intelligence gathering as compared failure codes.

I am very experienced at using both methods with some of the best companies in the world at documenting failures on work orders. My experience with the shortcomings of failure codes became my idea-seed for designing a text-based search in SQL for analyzing failures in work orders. SQL has its own limitations, however.

I was extremely lucky to meet exactly the right people, at exactly the right times in my career to not only develop a SQL statement that works to find and record keywords in text, but later on develop VBA programming in Excel that was far more powerful and up to the task. I call my method Contextual Failure Analysis. For those who haven't seen my IMC presentation, I will attach it here.

Let the discussion go where it may. And I will be happy to show anyone interested what I have working and how it compares to Failure Codes.

Thank you and best regards,
Jeff Wahl
Maintenance Program Manager, CMMS

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Thank you for your contributions Sarah. I believe that defining problems and resolutions are best accomplished using communicated text between individuals. When I find time I will list some of the ways I've used my search engine to find, measure and improve systems site-wide.

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Original Message:
Sent: 08-23-2019 04:53 PM
From: Sarah Lukens
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Jeff, this is fantastic.

You did a great job articulating many points, particularly on failure code designs, the time spent coding data as the data is getting entered, filling in missing codes using the free text, and in this thread how you learned the added benefit of gaining insight in how people communicated, all pointing to the observation "text-based analytics is a more natural way to find failures".

I've been working with data from the CMMS/EAM for almost 5 years now, developing ways that it can be useful in maintenance and reliability work processes, and have long ago come to the conclusion that you need to use the text field.  For every reason Jeff articulated.  One thing I do is try to develop ways to apply a lot of current research to actual applications in practical ways.  I think it's worth mentioning a current research effort out there about: due to all the observations Jeff had, why don't we actually revisit how we capture data in the maintenance management system?

Their first paper goes about that by essentially taking an FMEA approach to measure sources of human unreliability in the maintenance management work process itself:
https://ws680.nist.gov/publication/get_pdf.cfm?pub_id=927498

There's lots of references to SMRP in there too :)




------------------------------
Sarah Lukens
GE Digital
Roanoke VA

Original Message:
Sent: 08-15-2019 10:36 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Did every successful web-search engine get it wrong? It seems very unlikely, doesn't it?

I am starting this discussion as a place for all of us to explore the merits and shortcomings of using a powerful text-based search engine for intelligence gathering as compared failure codes.

I am very experienced at using both methods with some of the best companies in the world at documenting failures on work orders. My experience with the shortcomings of failure codes became my idea-seed for designing a text-based search in SQL for analyzing failures in work orders. SQL has its own limitations, however.

I was extremely lucky to meet exactly the right people, at exactly the right times in my career to not only develop a SQL statement that works to find and record keywords in text, but later on develop VBA programming in Excel that was far more powerful and up to the task. I call my method Contextual Failure Analysis. For those who haven't seen my IMC presentation, I will attach it here.

Let the discussion go where it may. And I will be happy to show anyone interested what I have working and how it compares to Failure Codes.

Thank you and best regards,
Jeff Wahl
Maintenance Program Manager, CMMS

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

I had the joy of seeing Jeff's demo.  Jeff has developed a toolset for leveraging CMMS data in practical ways developed over years of experience.  He showed me umpteen case studies of how the results of his tool have been applicable in business settings - for measuring successes to quantifying problems in a business case for getting the needed resources to solve the problems.  

To me, that's the real value of data-driven approaches - when the data and analytical methods are integrated in a work process from data generation to being used to achieve successes in your business.

I believe that as more emphasis occurs on emerging technologies and utilizing data in intelligent ways in maintenance and reliability applications, this is the direction if we want to utilize work order data from the CMMS in a meaningful way.  But as Jeff has well figured out and showed me, it's not about looking at the data itself, it's about implementing best practices in how descriptions are captured and how to capture the knowledge of the practitioner to be used in conjunction with the data that gives truly valuable results.

------------------------------
Sarah Lukens
GE Digital
Roanoke VA
------------------------------

Original Message:
Sent: 08-26-2019 03:29 PM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Thank you for your contributions Sarah. I believe that defining problems and resolutions are best accomplished using communicated text between individuals. When I find time I will list some of the ways I've used my search engine to find, measure and improve systems site-wide.

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO

Original Message:
Sent: 08-23-2019 04:53 PM
From: Sarah Lukens
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Jeff, this is fantastic.

You did a great job articulating many points, particularly on failure code designs, the time spent coding data as the data is getting entered, filling in missing codes using the free text, and in this thread how you learned the added benefit of gaining insight in how people communicated, all pointing to the observation "text-based analytics is a more natural way to find failures".

I've been working with data from the CMMS/EAM for almost 5 years now, developing ways that it can be useful in maintenance and reliability work processes, and have long ago come to the conclusion that you need to use the text field.  For every reason Jeff articulated.  One thing I do is try to develop ways to apply a lot of current research to actual applications in practical ways.  I think it's worth mentioning a current research effort out there about: due to all the observations Jeff had, why don't we actually revisit how we capture data in the maintenance management system?

Their first paper goes about that by essentially taking an FMEA approach to measure sources of human unreliability in the maintenance management work process itself:
https://ws680.nist.gov/publication/get_pdf.cfm?pub_id=927498

There's lots of references to SMRP in there too :)




------------------------------
Sarah Lukens
GE Digital
Roanoke VA

Original Message:
Sent: 08-15-2019 10:36 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Did every successful web-search engine get it wrong? It seems very unlikely, doesn't it?

I am starting this discussion as a place for all of us to explore the merits and shortcomings of using a powerful text-based search engine for intelligence gathering as compared failure codes.

I am very experienced at using both methods with some of the best companies in the world at documenting failures on work orders. My experience with the shortcomings of failure codes became my idea-seed for designing a text-based search in SQL for analyzing failures in work orders. SQL has its own limitations, however.

I was extremely lucky to meet exactly the right people, at exactly the right times in my career to not only develop a SQL statement that works to find and record keywords in text, but later on develop VBA programming in Excel that was far more powerful and up to the task. I call my method Contextual Failure Analysis. For those who haven't seen my IMC presentation, I will attach it here.

Let the discussion go where it may. And I will be happy to show anyone interested what I have working and how it compares to Failure Codes.

Thank you and best regards,
Jeff Wahl
Maintenance Program Manager, CMMS

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

, let's review a few tools that return to our custom, individualized performance rules.

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Original Message:
Sent: 08-28-2019 04:38 PM
From: Sarah Lukens
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

I had the joy of seeing Jeff's demo.  Jeff has developed a toolset for leveraging CMMS data in practical ways developed over years of experience.  He showed me umpteen case studies of how the results of his tool have been applicable in business settings - for measuring successes to quantifying problems in a business case for getting the needed resources to solve the problems.

To me, that's the real value of data-driven approaches - when the data and analytical methods are integrated in a work process from data generation to being used to achieve successes in your business.

I believe that as more emphasis occurs on emerging technologies and utilizing data in intelligent ways in maintenance and reliability applications, this is the direction if we want to utilize work order data from the CMMS in a meaningful way.  But as Jeff has well figured out and showed me, it's not about looking at the data itself, it's about implementing best practices in how descriptions are captured and how to capture the knowledge of the practitioner to be used in conjunction with the data that gives truly valuable results.

------------------------------
Sarah Lukens
GE Digital
Roanoke VA

Original Message:
Sent: 08-26-2019 03:29 PM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Thank you for your contributions Sarah. I believe that defining problems and resolutions are best accomplished using communicated text between individuals. When I find time I will list some of the ways I've used my search engine to find, measure and improve systems site-wide.

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO

Original Message:
Sent: 08-23-2019 04:53 PM
From: Sarah Lukens
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Jeff, this is fantastic.

You did a great job articulating many points, particularly on failure code designs, the time spent coding data as the data is getting entered, filling in missing codes using the free text, and in this thread how you learned the added benefit of gaining insight in how people communicated, all pointing to the observation "text-based analytics is a more natural way to find failures".

I've been working with data from the CMMS/EAM for almost 5 years now, developing ways that it can be useful in maintenance and reliability work processes, and have long ago come to the conclusion that you need to use the text field.  For every reason Jeff articulated.  One thing I do is try to develop ways to apply a lot of current research to actual applications in practical ways.  I think it's worth mentioning a current research effort out there about: due to all the observations Jeff had, why don't we actually revisit how we capture data in the maintenance management system?

Their first paper goes about that by essentially taking an FMEA approach to measure sources of human unreliability in the maintenance management work process itself:
https://ws680.nist.gov/publication/get_pdf.cfm?pub_id=927498

There's lots of references to SMRP in there too :)




------------------------------
Sarah Lukens
GE Digital
Roanoke VA

Original Message:
Sent: 08-15-2019 10:36 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Did every successful web-search engine get it wrong? It seems very unlikely, doesn't it?

I am starting this discussion as a place for all of us to explore the merits and shortcomings of using a powerful text-based search engine for intelligence gathering as compared failure codes.

I am very experienced at using both methods with some of the best companies in the world at documenting failures on work orders. My experience with the shortcomings of failure codes became my idea-seed for designing a text-based search in SQL for analyzing failures in work orders. SQL has its own limitations, however.

I was extremely lucky to meet exactly the right people, at exactly the right times in my career to not only develop a SQL statement that works to find and record keywords in text, but later on develop VBA programming in Excel that was far more powerful and up to the task. I call my method Contextual Failure Analysis. For those who haven't seen my IMC presentation, I will attach it here.

Let the discussion go where it may. And I will be happy to show anyone interested what I have working and how it compares to Failure Codes.

Thank you and best regards,
Jeff Wahl
Maintenance Program Manager, CMMS

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Original Message------

, let's review a few tools that return to our custom, individualized performance rules.

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Not really, Richard. I've shown it several times on an intimate level, which is the way I prefer to do it. Thanks for asking though.

One day, maybe. Or even present at an SMRP Conference. It's been hard sometimes, overcoming some of the closed-minded attitudes I've encountered. Everyone who's seen what I can do with it has been impressed, even the previously skeptic people. 

Best regards,

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Original Message:
Sent: 09-18-2019 06:47 AM
From: Richard Lamb
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Jeff, have you given thought to offering to explaining your tool in an SMRP webinar?

Original Message------

, let's review a few tools that return to our custom, individualized performance rules.

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

I echo Sarah's comments regarding Jeff's methodology.  Jeff took the time to walk me through text driven failure coding and the information that could be derived from his approach. I was thoroughly impressed and can see a number of practical applications particularly from my perspective.

Well done Jeff and thanks for taking the time to share.

Jason Wright
University of British Columbia
Vancouver, BC

------------------------------
Jason Wright MMP CMRP
Manager of Facilities
Vancouver BC
------------------------------

Original Message:
Sent: 08-28-2019 04:38 PM
From: Sarah Lukens
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

I had the joy of seeing Jeff's demo.  Jeff has developed a toolset for leveraging CMMS data in practical ways developed over years of experience.  He showed me umpteen case studies of how the results of his tool have been applicable in business settings - for measuring successes to quantifying problems in a business case for getting the needed resources to solve the problems.

To me, that's the real value of data-driven approaches - when the data and analytical methods are integrated in a work process from data generation to being used to achieve successes in your business.

I believe that as more emphasis occurs on emerging technologies and utilizing data in intelligent ways in maintenance and reliability applications, this is the direction if we want to utilize work order data from the CMMS in a meaningful way.  But as Jeff has well figured out and showed me, it's not about looking at the data itself, it's about implementing best practices in how descriptions are captured and how to capture the knowledge of the practitioner to be used in conjunction with the data that gives truly valuable results.

------------------------------
Sarah Lukens
GE Digital
Roanoke VA

Original Message:
Sent: 08-26-2019 03:29 PM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Thank you for your contributions Sarah. I believe that defining problems and resolutions are best accomplished using communicated text between individuals. When I find time I will list some of the ways I've used my search engine to find, measure and improve systems site-wide.

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO

Original Message:
Sent: 08-23-2019 04:53 PM
From: Sarah Lukens
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Jeff, this is fantastic.

You did a great job articulating many points, particularly on failure code designs, the time spent coding data as the data is getting entered, filling in missing codes using the free text, and in this thread how you learned the added benefit of gaining insight in how people communicated, all pointing to the observation "text-based analytics is a more natural way to find failures".

I've been working with data from the CMMS/EAM for almost 5 years now, developing ways that it can be useful in maintenance and reliability work processes, and have long ago come to the conclusion that you need to use the text field.  For every reason Jeff articulated.  One thing I do is try to develop ways to apply a lot of current research to actual applications in practical ways.  I think it's worth mentioning a current research effort out there about: due to all the observations Jeff had, why don't we actually revisit how we capture data in the maintenance management system?

Their first paper goes about that by essentially taking an FMEA approach to measure sources of human unreliability in the maintenance management work process itself:
https://ws680.nist.gov/publication/get_pdf.cfm?pub_id=927498

There's lots of references to SMRP in there too :)




------------------------------
Sarah Lukens
GE Digital
Roanoke VA

Original Message:
Sent: 08-15-2019 10:36 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Did every successful web-search engine get it wrong? It seems very unlikely, doesn't it?

I am starting this discussion as a place for all of us to explore the merits and shortcomings of using a powerful text-based search engine for intelligence gathering as compared failure codes.

I am very experienced at using both methods with some of the best companies in the world at documenting failures on work orders. My experience with the shortcomings of failure codes became my idea-seed for designing a text-based search in SQL for analyzing failures in work orders. SQL has its own limitations, however.

I was extremely lucky to meet exactly the right people, at exactly the right times in my career to not only develop a SQL statement that works to find and record keywords in text, but later on develop VBA programming in Excel that was far more powerful and up to the task. I call my method Contextual Failure Analysis. For those who haven't seen my IMC presentation, I will attach it here.

Let the discussion go where it may. And I will be happy to show anyone interested what I have working and how it compares to Failure Codes.

Thank you and best regards,
Jeff Wahl
Maintenance Program Manager, CMMS

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Thank you for your feedback Jason. Your group, people in the field working with the data, are the opinions that matter to me. I am one of many reliability people in that group. Doing it from the inside. There is so much more to talk about. Each improvement initiate deserves its own discussion thread.

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------

Original Message:
Sent: 09-22-2019 03:16 PM
From: Jason Wright
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

I echo Sarah's comments regarding Jeff's methodology.  Jeff took the time to walk me through text driven failure coding and the information that could be derived from his approach. I was thoroughly impressed and can see a number of practical applications particularly from my perspective.

Well done Jeff and thanks for taking the time to share.

Jason Wright
University of British Columbia
Vancouver, BC

------------------------------
Jason Wright MMP CMRP
Manager of Facilities
Vancouver BC

Original Message:
Sent: 08-28-2019 04:38 PM
From: Sarah Lukens
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

I had the joy of seeing Jeff's demo.  Jeff has developed a toolset for leveraging CMMS data in practical ways developed over years of experience.  He showed me umpteen case studies of how the results of his tool have been applicable in business settings - for measuring successes to quantifying problems in a business case for getting the needed resources to solve the problems.

To me, that's the real value of data-driven approaches - when the data and analytical methods are integrated in a work process from data generation to being used to achieve successes in your business.

I believe that as more emphasis occurs on emerging technologies and utilizing data in intelligent ways in maintenance and reliability applications, this is the direction if we want to utilize work order data from the CMMS in a meaningful way.  But as Jeff has well figured out and showed me, it's not about looking at the data itself, it's about implementing best practices in how descriptions are captured and how to capture the knowledge of the practitioner to be used in conjunction with the data that gives truly valuable results.

------------------------------
Sarah Lukens
GE Digital
Roanoke VA

Original Message:
Sent: 08-26-2019 03:29 PM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Thank you for your contributions Sarah. I believe that defining problems and resolutions are best accomplished using communicated text between individuals. When I find time I will list some of the ways I've used my search engine to find, measure and improve systems site-wide.

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO

Original Message:
Sent: 08-23-2019 04:53 PM
From: Sarah Lukens
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Jeff, this is fantastic.

You did a great job articulating many points, particularly on failure code designs, the time spent coding data as the data is getting entered, filling in missing codes using the free text, and in this thread how you learned the added benefit of gaining insight in how people communicated, all pointing to the observation "text-based analytics is a more natural way to find failures".

I've been working with data from the CMMS/EAM for almost 5 years now, developing ways that it can be useful in maintenance and reliability work processes, and have long ago come to the conclusion that you need to use the text field.  For every reason Jeff articulated.  One thing I do is try to develop ways to apply a lot of current research to actual applications in practical ways.  I think it's worth mentioning a current research effort out there about: due to all the observations Jeff had, why don't we actually revisit how we capture data in the maintenance management system?

Their first paper goes about that by essentially taking an FMEA approach to measure sources of human unreliability in the maintenance management work process itself:
https://ws680.nist.gov/publication/get_pdf.cfm?pub_id=927498

There's lots of references to SMRP in there too :)




------------------------------
Sarah Lukens
GE Digital
Roanoke VA

Original Message:
Sent: 08-15-2019 10:36 AM
From: Jeff Wahl
Subject: Contextual Failure Analysis: A Text-based Search Engine for Failure Analysis

Did every successful web-search engine get it wrong? It seems very unlikely, doesn't it?

I am starting this discussion as a place for all of us to explore the merits and shortcomings of using a powerful text-based search engine for intelligence gathering as compared failure codes.

I am very experienced at using both methods with some of the best companies in the world at documenting failures on work orders. My experience with the shortcomings of failure codes became my idea-seed for designing a text-based search in SQL for analyzing failures in work orders. SQL has its own limitations, however.

I was extremely lucky to meet exactly the right people, at exactly the right times in my career to not only develop a SQL statement that works to find and record keywords in text, but later on develop VBA programming in Excel that was far more powerful and up to the task. I call my method Contextual Failure Analysis. For those who haven't seen my IMC presentation, I will attach it here.

Let the discussion go where it may. And I will be happy to show anyone interested what I have working and how it compares to Failure Codes.

Thank you and best regards,
Jeff Wahl
Maintenance Program Manager, CMMS

------------------------------
Jeff Wahl
Maintenance Manager, CMMS
Loveland CO
------------------------------