The best metrics I have used are to count and track failures and ratio this with failures prevented.  Observations (inspections) should not only look to detect potential failures but also situations and conditions that could lead to a potential failure.  This is a readily acquired ability to recognize divergent operations and also adverse human behavior that result in 80%+ of equipment failures.

Thanks for your reply barry!

On which basis (daily, weekly, monthly) should one apply this metrics( depending on the number of machines you have).  Also should I follow one machine at the time or can I mix all the machines at the same time to estimate the metrics ? (An example is more explicit)

Barry from Texas posted:

The best metrics I have used are to count and track failures and ratio this with failures prevented.  Observations (inspections) should not only look to detect potential failures but also situations and conditions that could lead to a potential failure.  This is a readily acquired ability to recognize divergent operations and also adverse human behavior that result in 80%+ of equipment failures.

A team responsible for equipment performance inspects equipment, one at a time, for 3 to 4 hours every workday.  Failures and failure prevention are logged when found and reported weekly.  The inspection team is multi-discipline and autonomous in making decisions.  Highly reactive to ever changing conditions and varying situations.  During the 3 to 4 hrs., no meetings, no emails, no distractions.  Total focus on taking care of the plant.

Barry from Texas posted:

A team responsible for equipment performance inspects equipment, one at a time, for 3 to 4 hours every workday.  Failures and failure prevention are logged when found and reported weekly.  The inspection team is multi-discipline and autonomous in making decisions.  Highly reactive to ever changing conditions and varying situations.  During the 3 to 4 hrs., no meetings, no emails, no distractions.  Total focus on taking care of the plant.

 

Ok ! How to count and track failures and ratio this with failures prevented.  Let’s say in one week an experienced team counts  10 failures on a set of machines,  and prevents 20 other failures by adequate decisions. The ratio would be then 10/20 = 50%.  Did I get the math ? If yes, is there a world class value, a target value to reach , or  should I be looking for an increase or decrease of this ratio ?

Forget world class.  Just prevent as many failures as possible.  I turn the ratio around so in your example it is 2:1.  Keep tabs on the ratio and compare it to other teams in the plant.  Set up a competition.  Exchange ideas.  Build momentum.  The idea is to get everyone to prevent failures.  Change the language, change the culture, change behavior.  I have developed an entire system for preventing failures called          (9 to 1).  You can guess the meaning.

 (9 to 1)... Well I must admit that I did not get the subtlety.  I may need more clarification on this point, I am interested in learning more....

(9 to 1) (preventions to failures) is a fully designed system for preventing failures.  It was developed in a very large oil refinery to change the general approach to operations, maintenance, and inspection from failure detection, condition monitoring, and predictive maintenance to recognizing and eliminatung the precursors to failure.  It is not an engineering approach but rather uses organizational psychology and behavior change management.  It was estimated that the program saved the refinery $200 million in avoided losses PER YEAR.

Can this method (9 to 1) be transposed to the mining industry in the heavy equipment maintenance field?

I yes, where can I afford the book explaining that approach?

I am newbie in my field and in my organization regarding the maintenance activity of heavy equipment there are two parts: the daily inspection and reparation (level 1 to 2) activities on one hand and on the other hand there are the level 3 fixing/reparation activities ( plus other level which are not in my concern).

First to estimate my daily inspection and reparation (level 1 and 2) maintenance effectiveness, I will first focus on the ratio failure prevented to failures you talked about. I will report on a spreadsheet for each inspected machine the failures corrected and the prevented failure and calculate the ratio. So I need to establish a correct form to achieve that.

Secondly to estimate the effectiveness of the level 3 maintenance activities (where the machine need to be moved on a specific area), I can use the MDT (Mean down Time) metric which is equipment specific.

Do you think this is a good approach? (I may  get confused with several different forms,  or if a machine goes from daily inspection to a severe reparation  if I need to estimate the scheduled and unscheduled down time).  Any improvement suggestion or guidance will be appreciated.

You are asking the right questions.  First, there is no book outlining 9 to 1 but the good news is you don't need one.  The primary action for you to consider and eventually convince yourself is the correct path is to abandon what you are doing now.  I know this sounds drastic but looking over your response you are focused on failure detection (inspection) and repair (maintenance).  9 to 1 changes the behavior, the mindset, the focus to failure prevention.  The concept is certainly transferrable to the mining industry but I will admit, I have almost no experience in mining.  Step one, challenge the onsite team of operators, inspectors, maintenance techs, and engineers to work on the premise that no equipment will fail during my watch.  Let them work together to come up with a strategy to KEEP THE EQUIPMENT FROM FAILING.  Every time they perform an action to prevent a failure, reward the behavior and tabulate the result.  Allow this reinforcement to continue the good behavior of preventing failures, step back, and watch out.  The culture of your entire site will change and the overall results will be astounding.  I have seen this happen and know it works. 

Method to detect problems could be ultrasound. Compressed air/gas/vacuum/steam/electrical/mechanical applications. Turbo leaks, hydraulic leaks/valve bypassing, tire air leaks, A/C leaks, air filter housing leaks and more http://www.uesystems.com/applications

How much do you really want to know about your machines? What about the ability to collect wireless vibration/temperature/pressure/flow/levels or motor current monitoring would this also add value to your mobile/stationary equipment?  Provide you with the right mix of information to track and trend changes without a large output of man hours?

https://www.waiteswireless.com/

https://www.panpwr.com/

How much information is needed to make decisions early enough to prevent secondary or catastrophic failures, only you can decide but hope this information may shed new light on what is available to assist in getting you information

Dave

Hello Nicktruck,

My last work was in a mining industry which is an underground mining.  Sad to say that I believe that there is no 1 KPI, metric or maintenance indicator that will tell you the whole story of your maintenance function.  You need at least a minimum of 5 indicators.  For mining tell me if I am wrong but, this type of industry is one of the most problematic when it comes to lubrication which means that mining industry are heavy on lubrication costs.  I would recommend that your indicators should include:

- Cost of maintenance and repair - to be updated weekly or monthly

- downtime in hours - weekly or monthly

- ratio of Preventive Maintenance Compliance vs Emergency Repairs

- MTBF/MTTR

- Lubrication cost - monthly report

Hope this helps.

Hello,

Thanks for your reply! The method team estimate the MTBF, Availability, ratio of Preventive Maintenance Compliance vs Emergency Repairs ( hours) already.

As I am on the field ( and still in learning process), I have to estimate effectiveness of operations I specified earlier (see post above). . So Mean Down time is  a good indicator as well as the ratio of prevented failures/ failures tracked. But I am still opened to other suggestions.