Fellow Members,

I have been working in oil and gas Field, our company has set Target for Product Availabilty (Processing ability)for its stake holder, from that it derives and fixes Plant operational Reliaibity as 99.1% and Availabilty as 94% (based on general and past experince of failures, inspection intervention..etc ).finally we have to meet that Target.
if we Drill down further, that Operation Target based on the many Sub-Units Reliabilty and Availabilty up to the concerned Sub units components reliabilty and Availabilty. Management has fixed that based on their Demand and Supply but considering the other factors such as machine aeing factor and life cylce, repeative inspection intervention(gas turbines),degradation of other factors we may not achieve that TARGET. we would like set up a bench mark as this Turbine expected reliabilty is XX% based on its Past and present system and componenet reliabilty.From that we can calculate whether Plant relaibity &Availabilty is feasible target or not. To form that i have been looking for any Best practice/bench marking procedure with respect to the Unit( turbine,pumps,compressors) with respect to the Power( 1 MW, 5 MW, 10 MW) in the same group of Industries apart from chronic problem to that equipment.

Appreciate your suggestions and Time.

This message has been edited. Last edited by: KJAP,

Basically I think your target is 100% availability and 100% reliability or should be. Your numbers of 99.1 and 94% are out of order or it appears to me.

The factors to achieve this are skilled personnel to maintain and install in best of class maintenance techniques and procedures. Without this fundamental your PdM won't take you there - so, plan on training. Lubrication is your #1 program so, it is a good starting place - ensure it's right first and move on the #2 (training) and #3 PdM technologies.

You've asked a not so simple question. Your supervision can demand all they want but until you're given the ____________ to work with, it won't be realized or that's my opinion.

Sam,
From a strictly mathematical and practical sense, 100% reliability is not feasible except at the instant of installation of a new item or an item reconditioned to 'as good as new' condition. All items become less reliable with use, regain some of it after PMs, but the trend is always downward. Availability can approach 100%, but the incremental cost of moving 1% up from say 96% can be 5-10%, and tends to rise exponentially.
Economics and practical considerations demand we pitch our expectations at a lower level. So Jap's question is reasonable.

In practical oil&gas industry has setting availability more than 98%, I think so.

Sam
Achieving 100% plant reliability and Availabilty is possible if you have the following 1- Redundance equipments 2- As Mr.Vee Said if all of your equipments are as good as new.(even recently commissioned equipments were facing some sort of problems)otherwise 100% reliabilty target is not realistic.In oil and gas process Plant they can afford redundance of most of the equipments except some major equipments such as Main compressor doesnot have redundancy (becuase of Capital cost as well as Assured reliabilty) In our plant we have been using these compressors more than 25 years. and we upkeep the machines in each and every scheduled shutdown and maintain reliability above 95% but unable to achive the Management target. To substantiate that we have been trying to find what is the practical Average reliabilty and availability factors based on Power,Apllications, equipment generics..etc(apart from its own problems) with industrial best practice.hope some of the members would have done extensive reserach on that subject.

Mr Vee.
Could you please enlighten me further how to set up this is there any helpful standards or related articles available?

Mr.Aromatics Thailand
Appreciate your reply. it may vary plant to plant depend upon its equipments expected life time , upstream and down stream plant operational factors and scheduled shut down hrs.but hope equipment reliabity has certain bench mark out of those blues.

KJ,

Start with your design pecification and find out the level of reliability and availability the Plant was designed for; in principle we cannot exceed design levels unless we are able to change the design assumptions ( e.g., stated shutdown durations and intervals, stated downtime for Gas Turbine overhauls, water-washes etc).

Reliability improvements will come when you systematically eliminate failures. A good place to start is to make sure your equipment remains clean, dry and properly lubriicated. Moisture i oil can cause many avoidable failures. Make sure also that bolting on rotating machinery remains tight, alignment and balancing are done when required. These steps alone can yield rapid improvement, and are all steps you can recognize in a TPM program. These are relatively simple actions, and you dont need outside consultants to do these. Operating philosophies such as duty/standby, loading roughness, steady operating conditions etc. all impact ont reliability.

The next step is to ensure operations are within the design envelope, that equipment is started up, loaded, operated, and shut down in line with the vendor's procedures.

You can then refine your maintenance program with RCM, RBI and IPF, eliminate repeat failures with RCA etc.

Availability is strongly affected by shutdown durations and intervals. Good planning scheduling and work preparation can reduce downtime, during normal and shutdown maintenance. You can't do much about installed spares, so dont waste time fretting about lack of redundancy. Instead focus on spares availability and good logistic support to reduce downtime.

Benchmarking can only help you identify where the gaps in performance are, and possibly identify best practices. Right now you work is already cut out, you don't need to spend effort on BM immediately. Of course you will need it once in two years or so to find new gaps in performance.

In the O&G industry we use the term System Effectiveness, which is similar to OEE in the Manufacturing Industry. System Effectiveness is volumetric, not simply uptime. The numbers your Management want are likely to mean System Effectiveness, not Availability on a time basis. This is harder to achieve than simple availability.

K.Jap:

Redundant equipment does not mean 100% reliability. That is a myth. It improves the reliability, but not to 100%. Parallel system availability (redundancies) are determined by Rt = (Ra+Rb)-(Ra)(Rb) where R = e^-lt where l is the failure rate (normally 'lamda' but I do not have that symbol on here) and t is time. (this is the simple normalized curve method and not the Weibull method).

For multiple redundancies in identical systems, it is 1-(1-R)^n where n = the number of identical parallel systems.

I explore that particular subject more fully in the iPresentation: Reliability-Centered Motor Management (http://www.rcm-1.com/forms/sbdra_reg.htm).

As for availability approaching 100%. Vee is absolutely right. The closer you get to 100%, the more the diminishing returns. Take, for instance, Dofasco Steel. With their program, they were able to achieve 96% availability with relative ease. As they go beyond 96%, the costs go up exponentially and many projects that would improve availability are not pursued because their cost does not show a good return.

The key is, with planning production, that the actual availability of the plant is understood such that the actual capacity is determined. This capacity will decrease with age and the reliability of the equipment, which can be deferred through proper maintenance and equipment strategies.

In the meantime, the best approach that I can recommend, and I used to have to do this with students, is to take a step back, take a look at the complete picture, and determine which areas are going to have the greatest impact. Do not try to take on everything at once! That is a sure recipe for failure.

Sincerely,
Howard

Dear ALL,

Thanks for your contribution. hope it will help me to start something intersting.

Vee,

quote:

Originally posted by Vee:
In the O&G industry we use the term System Effectiveness, which is similar to OEE in the Manufacturing Industry. System Effectiveness is volumetric, not simply uptime. The numbers your Management want are likely to mean System Effectiveness, not Availability on a time basis. This is harder to achieve than simple availability.

When you refer to System Effectiveness, are you talking about the ability of the system to achieve a given throughput over a given amount of time? Can you give me an example of a calculation for a refinery? or a pipeline?

Thanks,

Kim Williams

Kim,
Let us start with the Upstream sector first. There is a Production System Capability, largely top do with the Reservoir and Well and Riser limitations. Let us call this X bbls, for a given month (as you know this number can vary depending on a number of factors) The Plant itself can be a bottleneck due to trips, shutdowns,and other reliability issues. Let us say we can process only Y bbls in that period. Then System Effectiveness is Y/X.
For a pipeline. let us say a major transport line, if everything works perfectly (e.g. pumping or compressor stations, no fouling in lines, no ESDs etc., no stoppages due to leaks etc.) it can transport P bbls or scfs. The actual volume is usually less than this for the type of reasons discussed. If this is Q bbls or scfs, the System Effectiveness is Q/P.
A refinery that has an infinite source and an infinite sink can process a certain volume of crude, say M bbls. Actual volume will be affected by Plant unreliability, storage limitations, product disposal issues etc. Let us say this volume is N bbls. System effectiveness is N/M.
Note that in each case the time-availability will be higher than System Effectivenmess. The latter is what makes the $$$.
In the O&G business, the term Availability is often meant in the sense of System Effectiveness.

Vee,

What time period are you using for these calculations?

For example, let's assume a refinery can process 100,000 bbl/day of crude (name plate). Let's also say it's unadvertised maximum demonstrated rate is 105,000 bbl/day. Let's also say that the monthly plan rate is 90,000 bbl/day. Now, on day 2 of the month, we have a process upset that causes us to cut rate to 80,000 bbl/day for 5 days. So we've lost 100,000 bbl per name plate capacity, or 50,000 bbl per monthly plan. Remember, we have a higher MDR. So we run at 105,000 bbl/day for the next 10 days. At this point, we've "caught up" to our monthly plan rate and if we don't have any more hiccups, we'll make the total volume planned based on a 90,000 bbl/day rate.

What's our System Efficiency?

Do you allow for makeup capacity, or do you calculate losses on a daily basis?

Thanks,

Kim Williams

Kim,
The time period can be a year, month or day, depending on what brings the revenue.
For example, if you were selling gas to a Power Plant or into a gas grid, the contracts may be based on a 24 hour time period. In which case, a day is the unit of time. If you have a monthly or quarterly production target, then that is the period to use. If your Company files quarterly P&L statements to the NYSE, then a quarter is the relevant time period.
Normally, we use a year as the time period for refineries. We would deduct downtime barrels for planned and unplanned maintenance over two years, annualise that, add the Turnaround downtime barrels annualized to match the Turnaround cycle time and gives us the total annualized 'lost barrels'. Annual capacity less lost barrels divided by the annual capacity gives us the System Effectiveness.

Management's interest in Availability is focused on revenue $$$. The Maintenance manager and Reliability Engineer have to focus on 'Time Availability' as well. For example protective and detective systems' availability affect Safety and Environmental performance. The latter is a 'given' for top management. At a lower level, we can only manage 'time-availability' of individual items; that is a practical reason for us to be concerned with this metric.

Your comment, quoted below, reminds me that some managers manage performance indicators, not performance. What matters to tham is to 'look good', not necessarily bring in more revenue. You say

quote:

a refinery can process 100,000 bbl/day of crude (name plate). Let's also say it's unadvertised maximum demonstrated rate is 105,000 bbl/day. Let's also say that the monthly plan rate is 90,000 bbl/day.

Several points to consider, and these are my views:
1. Refineries often do periodic test runs where they run at maximum sustainable throughput for 2-3 days, mainly to identify bottlenecks. Even if they dont do test runs, if in any month they can run 3 days in a row at higher than nameplate and still get the products out to the market, that should be their new nameplate capacity. Some ancient figure carved in stone when the plant was first built, ignoring all the de-bottlenecking and upgrade projects completed since then is not what matters.
2. The 105000 bbl/d is not a sustainable level for long periods; storage, handling and dispatch issues will come in; reliability issues will crop up with equipment not being available for PM on time and CM work being rushed through 'at any cost' to keep production going. If we can run for 5 days at this level and still manage the logistics and equipment reliability, this should be the nameplate capacity. If we cannot sustain for 3 days in a row, let us not consider this distraction at all.
3. The plan reflects the reality of the situation. It includes limitations on crude availability, ability to dispatch products and planned equipment downtime for the maintenance program. These are necessary to ensure sustainable operation. That is then the reality, so if we produce 80000bbl/d, the loss is 10000 bbl/d not 20000bbl/d. With your RCM hat, you can appreciate that the expectation is only 90000bbl, so with a delivery of 80000bbl, the gap is only 10000 bbl.

4. If we lose 10000bbl/d for 5 days and make it up by operating the plant at 100000 or 105000 bbl/d for any 10 days(or 7d at 105k) thereafter, within the scheduling window of 1m or 3 m as appropriate, we have not 'lost' any throughput. There is no loss in revenue. It does not matter whether the 10d are continuous or broken, as long as it is within the scheduling window. The important thing is that the scheduling window is tied firmly to the form of contracts we have with the customers. This is the essence of good production scheduling, and for this reason the maintenance planner must talk to the production scheduler regularly.
5. The 'time-availability' of individual equipment is another matter; if the process upset was caused by equipment unreliability, that should be of concern to the maintainers. Next time around they may not be lucky, and the lost production may be lost forever. That is real loss of $$$, and the shareholders wont't like it!

This message has been edited. Last edited by: Vee,