Posted at http://www.platts.com/Magazines/POWER/Power%20News/2004/061004_1.xml
POWER - 06/10/04
After the lights went out for more than 50 million people across the Midwest and Northeast last August, many observers said the push for restructured, competitive wholesale electricity markets may have blacked out as well. Utilities across the country were already embracing a "back-to-basics" business model, while many formerly high-flying energy merchants struggled to stay solvent.
For many, the blackout completed a trilogy that seemed to spell the end of restructured markets, following the disastrous California energy crisis and the sudden demise of Enron, which cast the whole competitive ethos in a bad light.
But for Federal Energy Regulatory Commission Chairman Pat Wood, the Aug. 14 outage occasioned a revelation: Competitive wholesale markets not only have the ability to save customers money, they are also necessary to improve reliability and prevent large-scale blackouts from happening in the first place.
Reliability, more than just economic efficiency, is the real reason to move forward with competitive markets, Wood said in a recent interview. "I think there's a real reason why we need to get there," he said. "It's not just about the economic efficiency and ivory tower [theories], it's reliability and keeping the lights on.
Originally Published at Ask an Expert
Question: RCA Prejudged
What is the secret to avoid leading the ROOT Cause to a pre-judged end.
I may have strong feelings as to what I think the was the cause, but want to be open to others and their opinions
Answer:
The answer is really quite simple - just follow the facts uncovered when collecting data. It should never be one's opinion that drives the development of the logic tree, but the questions that come up when collecting failure data. In essence you are actually building the logic tree during data collection. Every piece of data that you get makes you ask how can this occur. These questions should be reflected on the logic tree. Everyone has an opinion as to how incidents occur. The key is does the data support that opinion. If not, don't be afraid to follow the path clearly marked by the evidence. Start with fact - end with fact and you can't go wrong.
Ron Hughes,
Senior Consultant and Trainer
Reliability Center Inc.
Originally Published at Ask an Expert
Question: PM Optimization
We are just embarking on a PM Optimization here at our Pipe Mill in Australia. We have run a trial and have been impressed with the results.
Are there any pitfalls to be aware of before we proceed?
Answer:
Thanks for your question. Good to see that you are pleased with the results from PMO2000. The most common pitfalls are lack of resourcing (project management and shop floor people) and lack of celebrating the results. Motivating the people who provide their wisdom and effort is worth every ounce of management effort invested. Market to the bean counters as well so they see the work as a very good investment. We suggest that you appoint a project champion who has a duty statement with medium to long term activities. If the champion has also to cope with day to day issues, these issues always seem to take precedence over improvement activities no matter what they are. If you do not have resources internally to support the program we suggest that you contract them in. The next is engaging management to understand the principles of RCM/PMO and accepting that the analysis is done on the basis that every task adds value to the organisation so that not doing the PM will be a sub-optimal outcome for the business. Whilst a company may get away with missing inspections periodically, the risk of doing this is not worth it mathematically speaking. This means that the KPI for On Time PM Completion should be at least 95%. This understanding needs to be shared with the business drivers such as the Operations / Sales / Accounting people. Risks can be taken occasionally. However, in tight situations, one of the benefits you will derive from having done the analysis is that when you do have a tough decision to make, you can easily go to the analysis in the software and see what risks you are taking. You may decide to miss some tasks as a once only and take the punt, but you may decide to shut down for a small period to complete tasks which have consequences of safety, environmental or commercial proportions. Regardless of what decision you take, you are taking them from a position of knowledge rather than the seat of your strides (pants for the non Aussies). The next pitfall is doing too much analysis without implementation. With PMO we know that for the improvements to convert to revenue and reduced maintenance costs, planning and scheduling, defect elimination and performance measurement need to be upgraded to a standard which supports a reliability approach to asset management. If the analysis takes time to implement...don't keep doing analysis before implementation is streamlined and these other processes are happening properly. Next is probably failing to get the plant into a reasonable working condition. No amount of inspection will matter if the defective plant is not fixed. Last but almost most important is that you see PMO as a continuous process of reliability improvement. Understanding how the plant fails and documenting your strategy at failure mode level is just the start. Once you understand this and can access the information easily, there needs to be a review every time the PM fails. This is when you have an unexpected failure. It may be that the failure is random and gives no useful signs of deterioration hence failure is going to happen unexpectedly regardless of PM or you may have decided that the probability and consequences of failure compared with the cost of prevention mean that the cheapest option is to run to failure. In these cases you may sit back and ask yourself is modification a possibility or you may just accept the failure as an inherent part of the design (we all get broken windscreens and punctures in our cars) or you may find that there was an intent to ensure that what happened should have been picked up before the plant stopped. In the latter case, something went wrong and the reliability team should review why.... was the original analysis wrong....did the task actually get done... could the task pick up the symptom using the technique prescribed? Just a few thoughts. Good luck with your initiative. We think that you are on a winner.
Regards,
Steve Turner Director
OMCS PMO2000 - Maintenance Analysis for Results
Originally Published at Ask an Expert
Question: Compressor Availability Calculation
What is considered world class compressor availability in the arena of Oil and Gas Production Operations in the Gulf of Mexico (i.e. Offshore locations). Also, how is this availability calculated. Is preventive maintenance subtracted or does it count the same as downtime?
Answer:
Because compressors are used in different configurations, different duty cycles and environments, your question doesn't necessarily have a straightforward answer. My discussions with practicing reliability engineers in that industry has provided the platform for the following information. (World class may be higher than the 'target levels' that I provide below but I believe they are nearly the same.)
Because proper material specifications should be used relative to the environment, no reduction in expectations should be considered for salt water situations.
Reliability = (Operating hours - Downtime hours) divided by (Operating Hours)
Reliability target levels would be:
Redundant (inline spare) systems: Reliability of 99.0%
Unspared systems: Reliability of 99.9%
In most portions of this industry, the availability number is based on the possible use of the system (the compressor train) whether it is used or not.
Availability = (Installed hours - (downtime outage + Maintenance outage) (Installed hours)
The best compressor failure rate listed in the book 'Overall Equipment Effectiveness: A Powerful Production/Maintenance Tool for Increased Profits' is 16.09 per 1,000,000 operating hours. (This rate is not necessarily world class but has been achieved by a significant number of units.) If an average maintenance repair time of 14 days is assumed, then a projected availability can be computed for a redundant (inline spared with automatic switching) compressor system by using the formulas from the book referenced above. The result is a system failure rate of 2.45X10 to the minus 7. This is 99.992% availability.
An analysis of a reliability block diagram of the above system (tested for 10 runs of 100 years each) gave an availability of 99.9936% and a minimum mean time between system downtime events of 50 years. This would be close to world class.
A very useful handbook with formulas, definitions, examples, guidelines and other reliability/productivity tools is 'Overall Equipment Effectiveness: A Powerful Production/Maintenance Tool for Increased Profits'. It can be ordered on-line by going to MaintenanceResources.com
Thank you for your question.
Bob Hansen, PE, CMRP,
Author of 'Overall Equipment Effectiveness: A Powerful Production/Maintenance Tool for Increased Profits'.
Originally Published at Ask an Expert
Question: Plant Reliability Calculations
Please let us know how to calculate the Plant and Equipment Reliability
Answer:
This simple question has provided the basis for many books and cannot be fully covered in a short answer, however I will attempt a brief answer.
Equipment reliability is the probability that an item or system can perform its intended function for a specified interval of time under stated conditions.
For manufacturing systems:
Achieved reliability = (Operating time - Downtime) divided by (Operating time)
Many different approaches are used to provide a Plant reliability measure. This is because plant processes are often disconnected and consist of a mix of product flows. Providing a combined average number for all equipment on the plant is sometimes used but in my opinion is not a very useful number because it doesn't reflect the criticality of various equipment systems.
I would recommend that reliability block diagrams be generated for each major product flow and the achieved reliability of each diagram be weighted or proportioned relative to the amount of time that product was produced during the interval being reviewed. The sum of the weighted results would be the Plant reliability.
This approach would focus on the important production systems and would still reflect the reliability of support equipment by achieving high on-line percentages.
A very useful handbook with formulas, definitions, examples, guidelines and other reliability/productivity tools is 'Overall Equipment Effectiveness: A Powerful Production/Maintenance Tool for Increased Profits'. It can be ordered on-line by going to MaintenanceResources.com
Thank you for your question.
Bob Hansen, PE, CMRP,
Author of 'Overall Equipment Effectiveness: A Powerful Production/Maintenance Tool for Increased Profits'.