|Home Page||Public Training Course Schedules||Over 150 Best Practice Articles||Expert Systems / Tools||This Month's Features / News||About Us||Your Question / Contact Us|
Links to related training, expert systems / development tools and further reading below:
Relevant Training Course / In-house Workshop Highlights:
Expert Systems / Tools:
Checklist of functional requirements (available on request)
Relevant Further Reading: The following further articles were mentioned in this paper:
a. Permanently Maintained Website Articles:
b. Previously Featured Articles from our Archives (Up to 2 per organisation available on request):
B005: Level Scheduling
B006: Scarce Skills Management
B014: Effective Bill of Material Design
B021: Fault Recording and Corrective Action Systems or Operations Management for Continuous Improvement
B025: Version Control
B027: Killing Old and Unprofitable Products
T002: Commonality Trees
T006: Pareto Analysis
T010: 2 Bin Systems
T014: Re-Order Point Systems
T016: Forecasting in Low Volume Environments
T017: Loading Boards
T018: Lot Traceability
Maintenance, Repair and Overhaul (R & O) Planning & Control Systems
The following article describes the control systems requirements for maintenance, repair and overhaul (R&0) environments. It discusses the two types of maintenance: ("garage" and "re-manufacturing"), the characteristics of each, how to plan and control them with particular reference to capacity and Materials Management, Key Performance Indicators (measures of performance), Preventive Maintenance, Total Productive Maintenance (TPM) and the IT requirements to support the maintenance activity.
Links on the left to related training and (for topics highlighted in red) further reading
Repairs / replacements may be scheduled based on estimated life, or "fix on fail" or a mixture of the two.
There are two basic types of repair and overhaul:
The Garage Type
Garage businesses may do routine servicing and repairs / replacements on an individual basis, or in vehicle or aircraft sets (Equivalent to make to order). Units may be returned for repair, fixed in situ, or remotely. Diagnosing faults are a key feature of this type of business particularly in computer maintenance. Remote diagnostics may be performed.
Maintenance may take on an air of repetition or small batch production in higher volume applications. In these higher volume applications help desk and problem management are key features. Frequently version control (see Previous Best Practice B025 Version Control) is necessary to control the maintenance process and to provide for lot traceability. (See Previous Technique T018: Lot Traceability.) Examples of garage maintenance include vehicle repairs and servicing, computer maintenance, aircraft repairs and servicing, works engineering / maintenance.
The Re-manufacturing Type (Supply of reconditioned units)
Remanufacturing businesses are really two businesses in one:
These re-manufactured products are used (or may be then sold) as replacement products (usually in exchange for the old unit). (It is the equivalent to make for stock involving small to medium batch production). Examples of this are alternators for motor vehicles, printer ink-jet cartridges, and aircraft spares.
In order to try to identify the control needs for this sort of business we need to try to characterise the environment as follows:
Any one business may also encounter both types. In this case a hybrid control system may be required, when aspects of both of the following sections are applicable.
For any activity some aspects of materials planning and control and some aspects of capacity planning and control are necessary. The articles on Capacity Management and Materials Management provide further reading on these topics.
Let us now consider this in 3 dimensions and then draw some analogies with previous examples in other articles on the site:
The above cube breaks down into 8 segments, four for Garages and four for Re-manufacturing. In addition there are also a few other important aspects of the environment, which are important. Volume and mix have a profound effect on the control systems needs.
1. Materials Planning
This is generally a low volume or repair to order environment, where materials are ordered when required and some common long lead-time items may be forecast and scheduled. Some difficulties may be encountered in forecasting due to the low volume and intermittent demands. Control systems techniques that are appropriate include:
Safety stocks may be dictated by the need to support AOG (Aircraft On Ground) or VOR (Vehicle Off Road) requirements or support for other critical plant or computer systems.
Why not to use sophistication and MRP1
Some sophisticated modelling systems have been developed to try to forecast when units are likely to arrive for repair or need repair, coupled with a second model to determine which parts are likely to be replaced in each unit. This is used to create a Bill of Material (See Previous Best Practice B014: Effective Bill of Material Design) to drive an MRP1 provisioning system.
We would not use a provisioning model in this way except to establish initial stock holding requirements at new equipment acquisition or product launch. This technique requires two forecasts:
It is a lot easier to forecast demand (See Demand Management) based on historical usage at component level by recording issues from stores, (and probably just as accurate). The last time we took out an MRP1 system which required product forecasts and replaced it with a simple reorder-point system at component level it reduced stock by 20%, and stock-outs by a factor of 10!
However there is a trade off between service level verses the cost of holding stock. This balance can be tipped in your favour in several ways:
Then use the analysis as follows:
Beware of seasonal demand. For example more aircraft fly in the summer, so you need to watch out for that. Build stock in the spring, and reduce stock before the autumn.
We call the period just before reorder "the point of vulnerability". It is particularly relevant in assembly type work where the number of shortages of raw materials can be dramatically reduced by increasing batch sizes of the "C" items without any real penalties in inventory stock holding. The mathematics is also very simple. If an assembly contains 100 parts and each is replenished weekly, the risk of shortage is on average 100 per week. If the "C" items are replenished annually (typically 50% of the part numbers) the risk reduces to 51 per week, i.e. 50 part numbers are still vulnerable every week but 50 part numbers are now only vulnerable once every 50 weeks (one per week on average). If you then increase the batch size of the "B" items (typically 30% of the items) to say 2 weeks, you reduce the risk of shortage further to 36 per week (20 "A" risks + 15 "B" risks + 1"C" risk). If you then halve the batch size of the "A" items to twice per week replenishment, you have taken out about 50% of the inventory carrying costs while simultaneously reducing the risk of shortages by about 45%. In practice there is little extra risk simply because these "A" items are usually produced on a flow type basis. Because the cost of material shortage is significantly higher than any other costs in terms of lost output and dissatisfied customers (who may go elsewhere), you get the benefits of Pareto from both ends (reduced inventory costs and reduced shortages).
In practice we have used this technique many times now and have produced benefits at one end of the spectrum of a 40% reduction in stock holding without reducing a very high service level. At the other end of the spectrum we have reduced shortages by a factor of 10 whilst simultaneously reducing stockholding by 10%. The last time we did this in an aerospace company AOG disruption to production virtually disappeared, and service levels doubled (They were bad to start with). Stock levels did not increase.
It depends on your current situation as to which benefit you will see.
Four final cautionary notes:
2. Materials Control
The call-offs above may be via Kanban systems. Otherwise stores stock recording and control is essential, particularly for attractive items with a black market value. Strenuous efforts must be made to keep the shop clean and unused materials returned to stores routinely. A technique very valuable in this environment is "5S's", (See Previous Best Practice B026: Housekeeping), which aims to improve housekeeping. The last time we did this we filled many skips with rubbish and racking for the rubbish! We reduced the floor space utilised by about one third, we started to be able to find things in stores and we stopped ordering things we already had. This is particularly important if lot traceability (see Previous Technique T018: Lot Traceability) is required.
Shop floor control and tracking of jobs, features significantly in all but the smallest shops. (See Previous Technique T017: Loading Boards).
Version control is important in all safety critical situations, to plan and implement changes and to isolate problems. In particular suspect batches may require to be recalled (See Previous Best Practice B025 Version Control).
It was once the proud boast (and it made the works magazine) that our repair shop could service a 30 year old vehicle from floor stock. I was horrified that we could do this! And for three reasons:
3. Capacity planning
There is a classic dilemma in maintenance work. If the maintenance people are busy the place is not earning money. If they are not busy they are usually first on the redundancy list. Scheduling of maintenance work exists against a background of unusual breakdowns, which have to be accommodated in a hurry. The only 100% reliable way of managing this situation is to have spare capacity either through sub-contracting or through re-deploying maintenance personnel to other duties when not busy. This is very difficult unless routine scheduled maintenance predominates. Another problem is the lack of routine scheduling information (standard methods and times) for non-routine operations. A typical problem of this type of work measurement is the establishment of "loose standards", which if used to drive incentive schemes gives rise to serious problems. As an aside: incentive schemes are no substitute for good supervision. However "rule of thumb" time estimates and Rough Cut Capacity Planning (see Participative Sales and Operations Planning) is possible. Skills are the usual resources that need to be scheduled, not plant. If Total Productive Maintenance (see below) is being utilised scheduling becomes simpler because a higher proportion of the work is scheduled rather than breakdown dominated.
4. Capacity Control
This again is a classic dilemma. Do we do the urgent first or the very urgent? Frequently a job will be shelved to accommodate a more urgent one. This process can degenerate into very cluttered workshops and high work-in-process stock holding. Running a strict good housekeeping regime of operations control can alleviate this. (See Previous Best Practice T026: Housekeeping). The only satisfactory way to avoid building unwanted work-in-process is to use a simple form of input/output control. I.e. do not issue another job until the last is out of the way. (See Capacity Management: Scheduling ) One way which we have used to control work in process is to restrict the number of work or kitting trolleys to one per individual so that they can only be working on one job at a time. The trolley is used as a Kanban to request the next job from stores, when the previous job has been started. Also it is common to hold some sub-assemblies in work-in-process. Unless these require significant lead-time to assemble it is hard to justify holding sub-assemblies and this situation often leads to cannibalising one job to make a more urgent job. Our advice is do not do it unless you really have to, and draw a "Commonality Tree" to assess the need (see Previous Technique T002: Commonality Trees).
The use of loading boards (See Previous Technique T017: Loading Boards) is common in this environment. More recently electronic loading boards with pick and place facilities are being used.
Skill management may be very important to maintain. (See Previous Best Practice B006: Scarce Skills Management)
5. Other Important aspects
Tools management is essential with "Shadow Boards" (See Previous Best Practice B026: Housekeeping) used to ensure tools can be located when needed and in safety critical situations such as aircraft assembly it ensures that they are not lost.
The use of housekeeping techniques such as "5S's" is appropriate. (See Previous Best Practice B026: Housekeeping)
Diagnostic skills and possibly tools are required. These may be required to support remote diagnostics. Considerable effort may be required to establish this infrastructure.
1. Materials Planning
This breaks down into 3 parts:
2. Materials Control
Call offs are more likely to be via Kanban control because of increased repetition. It is vital to monitor yields in this situation to ensure that the correct numbers of un-salvaged stocks are sufficient to satisfy demand. Re-manufacturing creates a special problem for lot traceability. (See Previous Best Practice B026: Version Control). If a part has been recycled and it fails, what is the cause of the failure? Is it the original manufacture or the recycling process?
3. Capacity Planning
Because more time standards on work content are available (however informally) estimating jobs is easier. Because processes are more predictable Routes (Routings) (See MRP2) can be established to use in shop loading. Because there is repetition, demand is also smoother. The combined effect of these factors makes capacity planning easier. The use of "Level Scheduling" is recommended. (See Previous Best Practice B005: Level Scheduling.)
4. Capacity Control
Because demand is smoother and more repetition is present, skills management is less important and in fact more deskilling or automation may be possible. Switching effort to stripping rather than rebuilding can accommodate troughs and conversely reducing stripping to satisfy immediate demands can accommodate peaks.
5. Other important aspects
Sometimes the organisation may be slightly schizophrenic, flipping from job shop to volume producer. At this point it is worth considering some method of segmentation along resource utilisation lines. (See Organisation Redesign)
Effect of volume and mix
Volume increases are much easier to manage than mix increases. When volume increases, segmentation of the product is possible and automation of the ring-fenced product implemented. Kanban systems are most appropriate in this situation. Forecasting of demand is also easier.
As mix increases the overall business complexity increases. Considerable thought needs to be given to the proliferation of variety (see Previous Best Practice B027: Killing Old and Unprofitable Products).
Quality is a given these days, however faults measured in Parts per million is less applicable to this situation because volumes are generally lower. It is more common to measure the utilisation of the equipment that is the result of the maintenance process rather than the process itself. However a useful technique is "FRACAS" (See Previous Best Practice B021 Fault Recording and Corrective Action Systems or Operations Management for Continuous Improvement) which aims to follow up all faults to prevent a recurrence.
Often maintenance has a primary goal of meeting demand, which can often be measured directly in up-time, (of computers), down time (of critical plant), on the ground time (of aircraft), or off the road time of vehicles). Contributory measures to this include response time, and mean time between failures. This may need categorisation by critical plant or critical components.
The cost of maintenance can be significant so the productivity/costs are important measures. However productivity presupposes a standard output such as vehicles, repairs, etc. Which as we have said before may be difficult to establish because of the variable work content involved. This often causes up-time etc. (above) to be used as a substitute. One key feature of productivity is cash utilisation, which can be significantly influenced by good materials and Capacity Management above.
Further information on performance measurement may be found in "Focused Improvement Systems".
Preventative maintenance (PM)
Although the measurement of mean time between failures has been understood for decades, preventive maintenance has been slow to gain acceptance, as opposed to "fix on fail", except in the transport industry and safety critical applications. This is barely acceptable for non-bottleneck processes but is totally unforgivable for bottleneck processes. Because a scheduled maintenance event is less disruptive to output than a breakdown it is worth considering monitoring mean time between faults and the type of fault from the outset. Often patterns emerge where preventive measures can be taken to avoid a breakdown. This requires a degree of recording by item of plant and fault.
The most common technique for forecasting service intervals is "mean time between breakdowns". This is the arithmetic mean average of the intervals between failures. In predicting life expectancy it is often possible to avoid the breakdown by repairing just before the expected failure. Statistically an interval can be set which gives a high confidence of avoiding failures. The difficulty is gathering the data to make this prediction with confidence.
Total Productive Maintenance (TPM)
A more recent development and based on TQM principles is "Total Productive Maintenance", the main themes of which include continuous improvement achieved through allocating ownership (see Previous Best Practice B001: Ownership), of equipment, and routine maintenance to the "owner". The maintenance function's role is then elevated to major work, and prevention.
It is a fact that the cost of maintenance can become prohibitive as plant, vehicles, aircraft or software ages. Replacement theory dictates that this is monitored and that there is a planned event to replace the item before it actually dies but also before the cost of maintenance becomes uneconomic. It is not the intention to explore the mathematics of this here but simply to point out the fact that it can be cheaper to replace than maintain or continue to maintain, so you should not automatically choose the repair option, and you should keep records of repair costs.
IT support for the maintenance function
A checklist of functional requirements for the maintenance function is available on request. This service is not available to consultants. Below is a summary of typical packaged software support for the maintenance function:
Links to related training and further reading below:
To discuss your consulting or training needs with one of our independent consultants or trainers please Contact Us
Whilst great care has been taken to provide relevant, accurate, practical, advice based on our considerable process design and development experience, this will almost certainly require interpretation into the context of your unique business. Please be careful in doing so and if in doubt seek expert advice. We would welcome your feedback!
© SM Thacker & Associates 2012