Manufacturing data systems have come a very long way. In the 21st century they actually save time and money over old manual methods. The way they do this is by actually becoming the tools for getting real work done. Early systems were places where you recorded work after it had been done, creating a second set of work that had to be performed in the model, after the real, productive work was done. That’s why old systems didn’t work well. Nobody kept them up to date.
A modern accounting system, for instance, will always be up to date because a vendor can’t get paid until you enter the data into the system to cut a check. The system does part of the real work, so it has fresh data that come from the actual flow of work.
Sadly, most maintenance systems are not as tightly connected to the work they track as procurement and financial systems are. There are exceptions, but in most factories maintenance work is assigned by handwritten or oral instruction and material is issued when someone asks for it, not when data entry has occurred. The computer maintenance management system (CMMS), if there is one, is usually not where the equipment bills of material (BOM) reside, nor is it usually where the spares inventory is managed.
If spare part purchases and deliveries, maintenance work orders, maintenance staff timekeeping, work instructions, and equipment histories all reside in the same system, the CMMS, there is a chance of integrating all the work that supports maintenance operations.
[pullquote]If spares are requested on screen or checked out from a crib by bar code equipment that requires they be charged to a work order, then material order information can be captured and assigned to the equipment that is being worked on. Similarly, if workers are assigned to maintenance tasks by output from the CMMS, then labor information about work orders can be actual and up to date. Typically, neither of these happens in the CMMS, and the data doesn’t get logged after the fact in the secondary system, so the data isn’t integrated. Today’s maintenance systems usually look very similar to production systems from 1990.
This does not mean that data-driven maintenance is impossible. It just means that anyone who wants data to drive maintenance must roll up his sleeves and get it from files that were intended for other purposes. Here are a few typical data points that maintenance and reliability people would like to develop to guide their work. Each has suggested sources to get a useful chunk of the data. In each case we will be using a Pareto approach, finding the 20% of items that drive 80% of cost, work, downtime, or whatever we are studying. This also typically means we do 20% of the work that an exhaustive survey would require.
Find the assets that consume the highest spare parts cost: Typically, the CMMS does not have spare parts volume connected with the assets on which the spares are used. It also doesn’t usually connect these two with the work orders against which the parts are consumed. Here is the workaround.
The MRO system usually handles 90% of the company’s POs. It usually contains more than 90% of the vendors, but the number of vendors who handle the bulk of maintenance spares is usually relatively small. These vendors can be identified by item codes or purchasing agent assignments noted on the item master. If a printout or screen capture of the maintenance materials vendors can be obtained and ranked by annual dollars purchased, the goal is in sight.
Highlight the vendors who make up 80% of one or two years’ maintenance purchases. The result will usually be a list of a few vendors, up to 20 or so. This will be a manageable number for manual analysis. At this point you may also wish to survey the vendors, usually a larger number, who make up the remaining 20%, and, where it is easy to assign them to specific equipment, add them to the list. Don’t spend a lot of time on this side step, though.
Next, obtain a list for each of the high-volume vendors of the items purchased from them, with quantity and price data columns and material description fields. Obviously, if there is information associating these parts with work order or asset numbers, keep that data. In some cases, this information will not be in the item master system. If that is the case, the vendors may be able to help with this step. When the lists are complete, select the high-volume and high-cost spares. You will probably have a list of no more than 100 or 200 parts.
Finally, take the lists of high-volume and high-cost spares and work with maintenance people, crib personnel, or whoever else you need to associate the parts purchases to specific assets or work orders. This will be a surprisingly easy step, once you have the right people working with you. If there are large numbers of parts being ordered directly by maintenance people, outside the MRO system, this is the time to identify those parts and add them to the list. This information will be available from the same people who are helping you assign parts to assets.
When the job is done, you will have the spare parts consumed by the most costly assets to maintain. It will be a Pareto analysis, but, since it captures the most expensive spares, it won’t miss the assets that consume the most MRO material money. Make a list of assets and create a column listing maintenance material cost for each. Total the maintenance cost column and ensure that an adequate percentage of total material cost is present.
Find the assets that consume the most maintenance labor cost: Gather up a list of maintenance work orders for the past year or two and group them by the assets they serve. It will be best to gather work orders for the same time period you used when listing material purchases, that way you’ll have labor and material for the same repairs.
If maintenance hours are already listed against the work orders, add them up for each asset and put them in a new column on your asset list. If the hours are not available, divide the work orders up by asset and by trade within assets. Then work with trade leaders to estimate the number of hours per year that are devoted to each of the assets generating the most work orders. Put the estimates in the labor column on the asset list. Total the labor column and ensure that it adds up to the bulk of the plant’s maintenance labor cost.
Find the assets that cause the greatest production losses: Production losses are usually well documented, often not in the integrated data system. Daily, weekly, or monthly reports almost always include them, along with information about the assets, or at least work orders, related to the losses. The losses are often listed as costs. If not, it is usually easy for accounting to quantify the cost of the losses. Once the accountants see what you are doing, they will normally be very willing to help.
Working with the appropriate people from production and accounting, make a list of production loss events with the assets responsible and the amount of the losses for each event. Finally, total these by responsible asset and add them in a production loss column to your asset list.
Identify the short list of problem assets for your facility: The list of assets will now have production losses, maintenance material costs, and maintenance labor costs for the period you have chosen. Add the total costs and the percentage of production losses and maintenance cost that are attributable to each asset. Rank the asset list by total cost for the period and ensure that it makes sense to your key maintenance, operations, and financial people.
Once you have agreement to the problem asset list, you will be ready to discuss it with your operations people and management. You will now have the information to begin a list of data-driven maintenance initiatives to correct the non-value-adding expenses that have been gathered by asset. When you know the cost of correcting each problem, you will have the data to identify the financial improvements that the correction will create. This is the basis for a business case to support each maintenance initiative. You will also have demonstrated the value of an integrated CMMS.