Maintainability is an important aspect of any system's lifecycle, but process plant engineers typically give it little direct consideration. This is primarily the result of a short-term view of capital project costs that fails to consider lifecycle costs and downstream activities.
There is an old saying among maintenance personnel: "Engineering has it for a year, but maintenance has to live with it for 20 years." This goes with the engineering saying: "Get it out the door; we can always fix it on someone else's work order."
There is considerable truth to these sayings, as many times engineering and maintenance are driven by different metrics, leading to conflicting interests. The result is a rather expensive proposition for a company in the long term, but not this quarter, which makes the bean counters happy.
This article takes a broad view: It discusses some of the qualitative aspects of maintainability of instrument systems, but is also applicable to many other systems.
Before you even start on a system design, here are two principles to keep in mind: Minimize maintenance from the beginning, and get your maintenance people involved.
Reliability is a cornerstone to successful system operation and directly affects system maintainability. The more reliable a system is, the less maintenance it will require. Allocate your costs carefully. For example, maintainability is best improved if the difficult-to-maintain parts of a system are made more reliable. Identifying reliable components, installation types and arrangements, and vendors is a key step to improving the reliability of instrument systems.
Adding fault tolerance can reduce downtime to zero and allow repair off-line. More components, however, will raise the overall maintenance rate.
Getting maintenance involved early in a project can improve maintainability. This seems like a no-brainer but politics, turf wars, personalities, and sometimes even arrogance interfere with what is obviously a good practice. Involvement also brings ownership and an improved relationship between maintenance and engineering. At the very least, maintenance won't be able to complain about the system since they were involved in the design.
The somewhat overworked acronym KISS (keep it simple, Stupid) applies to maintainability. While we live in an increasingly complex and sophisticated world, we should strive to keep it simple wherever we can. You can design something that is really cool, sophisticated, and elegant, but if it's too hard to maintain, it will end up on the junk heap, wasting everyone's time.
Modular design divides the system into physical and functional modules, which can be arranged to facilitate design and maintenance. Easily replaceable modules with logical organization reduce repair time, troubleshooting, training, and engineering. Interconnectability and interoperability should also be considered.
Modular design of software can improve its maintainability.
Some thought must be given to future modifications or expansions. Nothing is typically static in a process plant; improvements are made, additions are added, and things are modified. Engineering must consider what the future may hold and make reasonable accommodations. Failure to do so may make a system that is easily maintained today but difficult to work on tomorrow.
Standardize, Document, and Label
Using recognized national, industry, and company standards and codes is good engineering practice. It improves maintainability by reducing variations of design and installation for maintenance personnel.
Standardization of components reduces inventory and improves reliability and maintainability. The use of commercial off-the-shelf (COTS) components should also be considered, though care should be taken because some commercial components may not meet the necessary industrial requirements.
While many instruments are unique and the technology in some cases changes rapidly, standardization of instrumentation can have the same benefits of reducing inventory and improving reliability. In addition, training and engineering costs can be reduced. Successful standardization requires that reliability engineering techniques be used and adequate maintenance records be kept and analyzed to determine what should be standardized on.
Many of our systems consist of a variety of the same or different manufacturer's instruments. In some cases, these instruments are arranged in proprietary configurations; in others, the systems are more open, leading to some ability for spare part interoperability, which can improve maintainability.
Documentation is extremely important in achieving good maintainability. This is particularly true as the complexity and sophistication of the system increases. Documentation is a matter of discipline, which, unfortunately, many engineering and maintenance systems do not have. It comes in many forms--user, vendor, third-party, and, not surprisingly, in the heads of your maintenance and engineering people. Documentation has different functionalities. Installation documentation, while it may have some overlap, is not the same as maintenance documentation. Make sure that maintenance documentation needs are being met.
Adhering to standard drawing organization, style, formats, symbols, and level and type of information provided can make life easier for the instrument technician. The less time trying to figure out what the drawings are saying, the more time spent on troubleshooting.
Some other considerations for documentation are accessibility, organization, readability, usability, applicability, and comprehensibility. Documentation accessibility is a major concern. Documentation can't help you if you can't find it or you can't understand it or it's wrong once you do find it. You'd think it would be a simple concept to maintain documentation but if you have been in this business very long, I'm sure that you've run across a system or two (or maybe a lot) where you can't find the manuals or the drawings are incorrect or missing. In some companies, this is a costly way of life.
Good software documentation greatly improves its maintainability. System functional requirements, software functional description, flow charts, software annotation, configurations, and I/O and memory mapping are some of the types of documentation that improve software maintainability. Up-to-date digital hardware configuration documentation (both hard and soft) on standard engineering documentation can also improve maintainability.
As-builts are a particularly abused form of documentation. They are many times not done or, if done, are not picked up by engineering, leaving future generations to suffer. Sometimes the engineering system makes it difficult to get as-builts picked up, bean counters won't allocate adequate money to get them done, or the system is not organized and they just don't get done, to the detriment of all.
Out in the plant, system components must be easily identifiable. Labeling must be consistent, standardized, clear, and accessible. While not a substitute for drawings, system identification that provide the capability of tracing wiring, power sources, and identification of components without use of drawings provides for more efficient and safe troubleshooting. Proper identification is also a safety consideration because it can help ensure that the proper things are worked on and that hazards are properly identified.
Access Is Critical
Accessibility means having sufficient workspace and access to perform maintenance safely and efficiently. Adequate workspace is needed not only to repair or maintain the system but also to troubleshoot it. Consideration should be made for opening of doors or removal of panels, removal of parts or modules, required manipulative tasks, needed body and tool positions and movements, duration of the access, and potential exposure to unsafe conditions during access. No instrument tech should need to have a four-foot tentacle with eyes on the end to work on equipment, nor should he need to be one foot tall or eight feet tall. If the access is difficult, short cuts will be taken and undesirable results may occur.
Once we are there, we need to be able to see what we are working on. Lighting level and direction as well as component size, location, orientation, texture, and coloring should be considered.
Component accessibility within an instrument or piece of equipment must also be considered. Low-reliability components should be the most accessible. Components should be replaceable with the least amount of handling: You shouldn't have to move or remove a bunch of stuff to get to what you are working on. Consideration must be given not only to how a component will be removed and replaced in an instrument or system but how the component will be handled once it is outside of the instrument or system.
Physical accessibility of field installations must be considered. This is a balance of functionality, cost, and reliability. Safety, however, must also be a prime consideration. Instruments up in the air lead to the potential for a falling accident, one of the leading causes of accidents in process plants.
Some instruments by their nature or requirements are located in poorly accessible areas: In-situ stack analyzers will be up the stack. A current trend is to close-couple instruments, which tends to put them up in the pipe rack. The answer in some cases is access platforms, which add costs the project engineer doesn't like but otherwise will be installed later on anyway by maintenance. One approach is to look at how often the instrument is accessed and if it is not often, perhaps the access can be done by a scissor lift or manlift or with strict administrative and safety controls. On the other hand, if the instrument is expected to be accessed often, an access platform is probably warranted.
Exposure of the system to weather, other environmental concerns, and stresses generated by other equipment such as heat, vibration, moving parts, etc., must be considered. If an instrument is hard to get to, environmental protections and exposure guards may be bypassed to make maintenance easy in the short term. This may have negative long-term consequences.
Protect and Accommodate Personnel
Whatever maintenance we do must be done safely. No maintenance action should require a person to perform an unsafe act. That being said, there are maintenance activities that are less safe than others. For example, working on a transmitter 15 feet up in the pipe rack is less safe that working on it at ground level. Proper location in the pipe rack to allow the use of a platform, installation of access platforms, or locating where a scissor lift or manlift can be used can improve the safety for such an installation.
Mechanical and electrical hazards must also be considered. Pinches, sharp bends, edges and points, trip hazards, head knockers, and abrasive surfaces should be eliminated or guarded. Adequate electrical and other mechanical guarding also must be provided.
Systems that are hard to work on may encourage unsafe work practices. Remember that the number of short cuts taken is directly proportional the difficulty of a maintenance activity.
Human factors, another neglected area, considers human physical limitations or where human errors can occur because of arrangement, order, color, identification, or other factors that are contrary to human expectation or action. Some physical limitations are height, reach, size, strength, sight, color blindness, and repetitive action limits. Some human factors are culturally based and different cultures will react differently to the same stimulus.
Another human factor is consistency. Humans expect consistency even across different manufacturers' systems and certainly across systems designed by their own company. Consistency makes a maintenance person's job easier and safer.
Morale is also a prime human factor consideration. Maintenance forces with high morale will consistently provide higher-quality maintenance, faster maintenance times, and better safety records.
Streamline Testing, Calibration, and Troubleshooting
Equipment that is difficult to inspect or test will be less likely to be inspected or tested. Short cuts will be taken, potentially leading to not getting the desired benefits of inspection and testing.
A system should be easily tested for troubleshooting. Consider the failure modes and how they will be detected.
Any adjustments should be simple, and the system should be designed to minimize the adjustments necessary to keep it running properly. An instrument's drift specification is an indication of how often adjustments may be required. If you have to keep tinkering with an instrument or instrument system to keep it running, your maintainability will be low and operators are likely not to trust the instrument.
Modern equipment typically comes with self-test diagnostics, some more than others. The required level of diagnostics is a function of the "invisibility," or level of abstraction, of the internal functionality of the instrumentation. For example, the level of abstraction for something you can measure with a meter is different from what is happening inside a microprocessor, hence the level of diagnostics is different.
Diagnostics are not just the domain of the equipment supplier. The addition of low-tech blown fuse indicators, process variable indicators, and pilot lights can go a long way toward speeding up your maintenance/repair time. The use of internal status or alarm bits in programmable devices and user-designed diagnostics are another fertile area. System capabilities today make possible computer access to internal diagnostics, and the resulting displays can enhance maintenance capabilities. Look for opportunities to add cost-efficient diagnostics to your instrument systems.
For diagnostics to be successful, they must have comprehensive coverage and provide speedy localization of the problem. They must also be understandable. Cryptic or generalized error messages or poorly documented diagnostics can limit their effectiveness--if you can't understand what your diagnostic system is saying, how can you identify your problem?
Are Your Logistics Logical?
Logistics is the ability to have available, at the needed time, the resources and parts necessary to make the repair.
Tools and training are fundamental. The proper tool for the job is a key to safe and efficient maintenance. Some companies "save" money by not providing the proper tools, when they could buy the tools 10 times over with the cost of one process outage prolonged by not having them. Penny wise, pound foolish!
Today's systems can be complex and sophisticated, making maintenance difficult for the untrained. It is becoming increasingly harder to work on systems without some form of formal training and, in some cases, refresher training. Consideration of training requirements must be done up front in a project. But remember training alone is not a good substitute for expertise (knowledge + capability + experience + application).
Obviously, if you install a new machine for which you don't already stock the needed or recommended spare parts, then you will have to make arrangements for parts. This could include stocking them either by purchase or distributor/manufacturer on-site stocking agreements. Since it may be impossible to stock all possible spare parts, the availability of parts locally, regionally, or at the factory must be factored in.
Maintenance resource availability is also a key to good maintainability. In this day of downsizing and an older workforce retiring without adequate replacement, significant resource shortages may manifest themselves in some companies. Resource planning based on the true resource needs rather than artificial cost or structural constraints can significantly improve maintainability on a system basis. Contracting is an often proposed solution but on a general basis; out-of-house expertise is seldom an equal substitute for in-house expertise.
Informational and expertise resource access must also be considered. It seems like a no-brainer that a company would make available those expertise resources within the company, but many companies have organizational silos that tend to prevent this. Vendors should also be considered, but beware of promised resources as they have a habit of disappearing after you buy the equipment or let the contract.
Prevent and Predict
Equipment that is not easily accessed or is difficult to work on will be less likely to receive preventive maintenance. Maintainability is a function of the amount of preventive maintenance, but it is a complex function. The more work required to maintain a piece of equipment, the less maintainable it is. But on the other hand, the preventive maintenance also reduces the potential for failures, which improves maintainability.
Predictive maintenance, on the other hand, allows one to detect problems early and schedule maintenance when resources are available and operational needs are more easily met, which can improve maintainability.
New technologies such as web-enabled monitoring, maintenance-based artificial intelligence, multimedia data access, wireless sensors, improved diagnostics, and self-validating devices will allow improved monitoring of equipment and systems, which will enhance instrument maintainability.
William L. (Bill) Mostia Jr., PE, of WLM Engineering, League City, Texas, has more than 25 years experience applying instrumentation and control systems in process facilities.
1. Reliability, Maintainability and Risk, 5th Ed, David J. Smith, Butterworth Heinemann, Woburn, MA, 1997, ISBN 0 7506 3752 8
2. Maintainability & Maintenance Management, 3rd Ed, Joseph D. Patton Jr., ISA, Research Triangle Park, NC, 1994, ISBN 1-55617-510-8
3. Assurance Technologies - Principles and Practices, Dev G. Raheja, McGraw- Hill, New York, NY, 1991, ISBN 0-07-051212-4
Understand Your Constraints
It would be nice if there were adequate resources available and maintenance systems were perfect so work could get done in an efficient and timely manner. But that's generally not the case.
Many plants operate under constraints that limit the ability to maintain systems efficiently. Maintenance system constraint analysis can point to areas of potential improvement. Some of the common constraints are:
* Conflicting interests: politics, vested interests.
* Resource limitations: staffing, capability, quality, tools.
* Logistic limitations: access to spare parts and resources.
* Administrative or procedural inefficiencies: complex or lack of procedures, poor or no training.
* Operational constraints: marketing, production pressure.
* Organizational inefficiencies: structural deficiencies, poor resource allocation, work order-itis (concentration on work orders rather than the work), lack of ownership, poor morale.
* Lack of external support: organizational, vendor local and factory support.