Lean Maintenance. Joel Levitt
Читать онлайн книгу.
pull out the Labor rate and since all that’s left is multiplication
(2.5 x 440) / 60 = 18.33. Put the labor rate back in and that’s the savings per minute. Pretty easy.
Economic Modeling: How do you know an alternative is Lean?
Economic modeling will help you determine the Leanest alternative (economically speaking). Economic modeling simulates the costs and income from a particular maintenance alternative, given the economic ‘facts’ of the case. Models can be as simple as projecting the costs and income, to sophisticated models that include interest rates, tax policies, and other variables. If the consequences warrant it, alternatives can be analyzed using economic modeling.
Many different strategies can be used in maintaining particular assets. Of course, the first choice is to employ an asset that needs no maintenance! If no maintenance is possible, look at that alternative before ‘settling’ on a PM or another maintenance alternative.
Each choice has both economic and non-economic consequences. Economics is important, but other issues might suggest a particular strategy. For example, PM might be the best strategy for a particular asset but your company has no PM system and a bad track record of allowing downtime for PM, so PM will not work. There are 6 or 7 major strategies, and many more combinations and sub-strategies. A few of them could be:
•Run the unit in breakdown mode–where it gets no attention unless broken. This example is always used for comparisons. In modeling, doing nothing should always be a choice (and sometimes it’s the best choice).
•Redesign the asset to be quick-connected in place, so that you can do a quick switch upon failure and rebuild it off-line for use when the next one fails. Although the failure rate will be the same, each failure would be handled more quickly.
•Design a basic PM program with inspections, basic maintenance, and occasional as-required corrective jobs.
•Just bite the bullet and install a whole Back-up (say a pump) that can be switched over quickly before any chargeable downtime is incurred.
•Planning for component replacement with quick connections is like the second choice above, except that the asset is swapped on a scheduled basis before failure.
Example: Repeated failure of a chemical transfer pump has been occurring for the last several years. Downtime from lost production is valued by the cost accounting department at $500 per hour after the first hour (no cost for the first hour). There is a reservoir that will run for 1 hour before the downstream process shuts down. Labor hours are valued at $40.
Engineering analysis shows that the application is severe and that this performance is the best that can be expected (eliminating the no-maintenance alternative). The skilled mechanics working with engineering have designed a PM task list that will drop the number of failures dramatically.
Economic model for breakdown mode
Currently, in breakdown mode, the pump is failing 4 times a year. Each incident requires 10 labor hours and $2000 of parts to get the pump back on line. Downtime from calling maintenance to full operation is 14 hours (2 hours to respond to the call and 2 hours to get the system filled up and back in operation. A mechanic is required for only 10 of those 14 hours. It is a 1-person job).
In this model, the only thing required from management is to keep the spare parts in stock. A simple min-max system would suffice, with inventory levels depending on the lead time.
Economics of breakdown maintenance
The economics of this choice are clear, but what are the other consequences? There are several kinds of consequences. For one, the customer (operations) is disturbed 4 times a year for over one shift each time, amounting to 52 hours of chargeable downtime a year. This problem is not too bad if you have only one such pump, but imagine if you had 100 pumps, each failing for over one shift, 4 times a year. Of course the pumps would break at the least convenient times, and probably the failures would bunch up (lengthening the time to respond and the downtime).
Running in breakdown mode requires 40 hours of maintenance worker time. Another consequence would be the cost of $35,600 per year (in both above- and below-the-waterline costs). On the plus side, this situation requires little management and no capital expenditures. Although safety and environmental considerations are beyond the scope of this model, it is intuitively obvious that unscheduled and random events such as breakdowns are the least safe choice.
Economic model for hot swapping
Another breakdown alternative is to replace the pump after failure with a rebuilt pump and rebuild the faulty unit off-line. In the computer field, this approach might be called hot-swapping. Components (such as circuit boards) are swapped with the computer running. We can’t actually swap the pump while it is running but we still call the method hot-swapping (maybe we could call it warm swapping). To set up the method we would have to purchase a second full-pump unit and adapt it for quick changes.
The second pump will cost $15,000. After engineering the quick change, swapping to the second pump will take 2 hours. Downtime is down to 6 hours (2 hours to respond, 2 hours to swap and 2 hours to fill the system). The rebuild will cost the same amount as the breakdown model above, $2000 for parts and 10 hours off-line. Only 1 pump core is needed to rotate with the installed pump so we have to be sure to charge for only 1 pump core. The purchased pump is a capital expenditure.
Management needs to ensure that when the damaged pump is removed from service it is rebuilt reliably within 2–4 weeks. That element of management is in addition to a simple stocking strategy like the one described above. If the pump is not rebuilt in time, you run the risk of spending money for the spare and still having the downtime of the breakdown option. Having a breakdown on top of not having the pump rebuilt is the worst of both worlds (and is fat maintenance not Lean Maintenance).
Economics of hot swapping
What are the consequences of this choice? Once again there are several kinds of consequences. The Operations or production process is disturbed 4 times a year but this time it is for less than one shift. Chargeable downtime drops by 24 hours per year, or better than a 50% reduction, which is an improvement. There is no difference in the cost of maintenance labor or parts for this scenario. Annual costs have improved to $19,600 because of the reduction in downtime. The costs above the water line are the same. Each incident will still need 40 hours of maintenance labor and $2000 of parts. There is a capital cost, but it gets paid off within the year. More management is needed than with the previous example, to insure that the pump is rebuilt in a timely way. Safety is not affected because the failure mode is still random and unscheduled.
PM
PM is a common strategy. In fact, it might be tough to run a breakdown-only scenario because most firms do some PM (even if it is only a shot of grease). Frequently we notice that the greatest returns come from the most basic PM activity. There is a diminishing return to increased PM expenditure. Of course, if the subject is a fuel pump on a jet engine, the added benefit is worth while. The effectiveness of the PM is always at issue and is discussed in another section. An ineffective PM will consume resources without providing an adequate Return on Investment. It is important to look deeply at different strategies because the best way to take care of an asset might depend on the downtime cost, not just the breakdown cost.
The PM routine was designed by the mechanics, will take 1 hour a week, and requires downtime to accomplish (but that hour of downtime is free). We will assume that the PM routine as designed is effective. Additional corrective repairs (resulting from the PM inspector finding deterioration and fixing it before failure) take 5 hours per incident (3 incidents per year on average) and $1700 worth of materials per incident. With the new PM program, the system becomes significantly more