The amount of energy required to operate a compressed air system is based not only on how much air is consumed in demand, but how it is used. The relationship between the supply arrangement and demand usage also will determine energy consumed. In examining demand, the first question is: why do we operate the system the way we do? The ability to break down the demand by issue provides information to take the most informed approach to managing the system efficiently. What are some of the key demand issues that influence supply energy?

Minimum load
Although this is the condition with the lowest energy requirement, it often represents the most hours of operation per year in most systems. It is usually not evaluated and winds up being the stepchild when evaluating compressor sizes. There is usually a significant amount of part load of a larger-than-necessary compressor or compressors sized for peak demand. The waste and the hours of operation create a major opportunity for savings.

Leaks are usually the predominant user at minimum load, making leak benchmarking an excellent tool. On a regular basis, typically every few weeks, maintenance personnel can bring the demand volume back to the previous benchmark period through selective leak management using a powerful ultrasonic scanner and the system's mass flow metering. If demand is managed properly, pressure could be reduced considerably during low load to reduce operating cost as the percentage of unregulated air consumers usually increases as demand diminishes and the system's pressure rises.

Base load demand
Each of the conditions or shifts will have a base load demand that will not vary. The focus for servicing base load should be on the most efficient compressors.

Trim requirements
Each of the shift conditions will have a variable portion of the demand above base load. This is called trim. Service trim demand with compressors that are capable of loading and unloading and turning the motor on and off as required. The speed of bringing the motor and compressor to full load can be critical. The focus on trim is smaller, faster compressors with flexible controls. No larger compressor, no matter how inherently efficient part loaded, is as efficient as two smaller compressors with one off.

Profile of usage
Minimum load, base load demand, and trim requirements provide the profile of usage across the conditions, and there can be from one to five conditions in each system. When the anticipated profile is thoroughly developed, a company may find that it will not generate enough hours in trim to rationalize more expensive and efficient trim units. The base load should be serviced by the most efficient compressors because of the number of hours of service vs the energy consumed.

Do not forget the support of the low load. This probably will be supplied with one or more of the smaller trim compressors. Despite the sensibility of this, normally only the peak is considered. Several large compressors are installed with one extra which is loaded at the first inappropriate distress call from production. Another backup machine is added that gets turned on and left on. When questioned why the last compressor was not turned off, the most popular retort is "they havent complained since." When waste in demand is corrected, the relative size of the compressors to the total requirement increases significantly. With the system being operated with fewer large compressors, the risk of interruption in the system increases in proportion to the size of the largest unit on line.

With no demand control, no waste management plan, and no information, companies should not be surprised that within the first year virtually all of the systems need what they have and the companies are shopping the next prepackaged solution to a poorly defined systems problem.

The next two demand items are the most pervasive of all demand issues and represent the primary reason why the compressed air system is operated in the manner it is.

Peak usage from demand events
All systems have events. These are typically high volume, short cycle air users which create the peak in the system. It is important to know not only the rate of flow, volume per cycle, and the duration of the event, but also the recovery time available between events. If this event is ignored, the pressure would drop when it occurred. In most cases, the system is run at a high enough pressure all of the time so that when the event hits the system, it will stay above the minimum acceptable pressure. It requires additional power on-line all of the time in order to operate in this method. This is obviously not the best operating method, but it is the typical approach.

Most plants have no idea what this event is or what its quantitative value is. Evaluating this and other large events in each condition of the system is an essential part of auditing. The highest demand event may be a single high user such as material transfer or dense phase conveying. In most plants the major source of events is the coincidence of several large events hitting the system simultaneously, such as the start up of a shift at a specific time. Typically this highest peak event occurs at first shift start up. Smaller events occur at the return from breaks.

In one case we shut down a 1500 bhp compressor by disconnecting the shift and break start up horn. Instead of 700 people hitting their equipment at one time, the shift started over a period of 10 minutes. The breaks were split. The power required was significantly reduced. If a company only looked at minimum acceptable pressure and did not trend flow, it probably would miss this opportunity.

Another example of coincidental events would be several solenoid operated or motorized drain valves discharging at the same time. If there are enough of these drains, they are statistically going to overlap each other with some regularity. The more drains, the longer the open duration, and the shorter the intervals between actuation, the larger the event which will occur. Remember that the system sees events in real time at the rate of flow per event in cubic feet per minute (cfm). A 1/2 in. motorized drain valve may be open for only 5 seconds every 10 minutes. During the open time, it may discharge 40 cubic feet of air at line pressure for 5 seconds, but it occurs at 477 cfm rate of flow. The supply system sees this as a requirement for 110 bhp of compressor for this short duration. How much storage capacity there is in the system and the number of drains that actuate will determine how much the pressure will drop during the event. Imagine four or five of these valves opening at one time. Our experience is that this type of event normally occurs in the compressor room and will occur sufficiently to prevent any compressor from unloading even when there is not enough demand to support the supply.

It would be unique if someone operating a compressed air system knew what the events were, how they influence the system, or what to do about them other than operating more compressors all of the time to handle the events once in a while. Events are the single most important issue in designing and operating an air system. The amount of storage, controls philosophy, and size and type of compressors are all relative to event management. There are many ways to manage events besides throwing power at them.

Sizing control storage appropriately will help event management (see accompanying section "Control Storage Application and Calculations"). Another method of better managing peak events is the development of higher-than-normal-pressure, large-volume storage which is created off the main air system with typically a very small 150-200 psig, 5-25 hp compressor(s). The air then is reintroduced into the main piping with a valve control system operated by a programmable controller that controls the pressure drop when the rate of change in the system exceeds a preset value normally commensurate with the events which occur. The size of the storage tank is based on the size of the event and differential pressure between the system's demand pressure and the stored pressure.

The energy required is a function of the required rate of flow times the use time divided by the available recovery time before the event recurs. The idea of load shaping is to support events in the system while preventing the larger main compressors in the supply system from seeing the demand increase on a selective basis.

Critical or highest demand pressure required
Most systems operate on an error response basis. If the pressure drops and someone complains, additional compressors are loaded to increase the system's pressure so the complaints will cease. As this is the most common reason for the operating approach, it warrants a little attention. There are always several things that happen:

a. Both the caller and the powerhouse operator assume the supply is insufficient.

b. Neither party defines the problem that provoked the added compressor. There is no trended information.

c. Neither party understands the financial consequences of their four- to six-figure decision.

d. Half of the time, there is no problem. The caller is a gauge watcher. The other half of the time the problem is local to the caller and adding power has no influence one way or the other. (For years we have encouraged compressor operators to ask the caller to call back in a few moments. They seldom call back because the problem was self-correcting.)

e. Neither the caller nor the problem is ever recorded so the real problem can be corrected. The compressor is left on so the operator will never get the call again.

Is there something wrong with this picture? We find that most high, critical pressure users in the system have some rather common maladies. The most predominant is point of use regulators with high differentials. Differentials on regulators show up on the upstream side of the pilot pressure they are trying to hold. If a company has a 20 psig DP on the point of use regulator, which is not unusual, and it is trying to maintain a pilot pressure of 90 psig on this application, it will have to maintain 110 psig in the overhead system and a higher pressure at the compressors. The operator notices a loss of performance which he assumes is insufficient supply pressure.

If the philosophy is to keep calls from coming in, imagine the gyrations you will have to go through or the misdiagnoses in order to satisfy this kind of problem. More than half of the audited plants felt that the main piping was undersized. In all of these cases the rational for this diagnosis was that a user in the system could not hold 15-20 psig less pressure at the point of use on a critical pressure application. The real dilemma is the lack of problem definition.

Another problem which can go hand in hand with the regulator problem is the point of use filter which is loaded with dirt. As the filter gets dirty, the downstream pressure drops. It can run into the differential on the regulator. It will definitely change the way the air user is functioning. Since the user does not know why his air-operated device does not work properly, invariably it is assumed that there is insufficient supply. Nine out of 10 audited plants had never changed the point of use filter cartridges since they had been installed. They also had no replacement cartridges in inventory if they wanted to change the filters. In almost every system, the first two or three highest pressure users had between 15 and 30 psig of differential across the filter, regulator, hose, and disconnects. In most of these plants all point of use installations used the same size installation components regardless of the volume or pressure required by the air-using device. Once the user is installed, if it does not work, companies just increase the regulator until it does. If this does not work, it adds a compressor and elevates the entire system until that one user works. This is how companies typically and mistakenly cope with differential.

Most of the time, the differential at the point of use represents the highest pressure drop in the entire system. Sometimes the problem is another high volume user near the critical pressure application which causes the branch line or sub header pressure to drop into the critical application.

Silly solutions
Make the call. Add more power. Most of the time $100 problems are fixed with $50,000 solutions. It sounds silly, but this is commonplace without investigation, knowledge, or information management tools. Plant engineers and maintenance managers will struggle for months attempting to decide what brand and type of compressor should be used to fix a dirty filter. We hate to think of the tens of thousands of times we repiped systems to a larger size to fix an undersized regulator. In all cases the retrofit of the piping increased the system's storage capacity although no one thought of storage as a solution. There are much less expensive ways to provide point of use storage than increasing the header size.

Knowing that it will cost more to operate a compressor in the first year than it costs to buy and install one, adding another compressor is a serious decision. Even the most uninformed manager will balk at the capital expenditure if there is not a sound definition of the problem. In most cases, the decision to buy another box of compressed air with all of the required accessories will be sitting on the bottom of the to-do pile. Eventually the problem will be bad enough that production will support the purchase. The purchase, installation, and start up will occur. The pressure will be elevated a few psig and the complaints will cease for a period of time until the process starts all over again.

Conclusions
The average demand reduction in the audited plants was 43 percent although this is an on-going process. The average demand pressure requirement has been reduced by 12 psig and many plants feel they can reduce this further. The average savings per year including all costs of compressed air has been more than $400,000; the size of the system and the burdened cost of energy, water, and maintenance will influence the potential savings. The average return on investment—adjusted for tax treatment, cost of capital, and adding depreciation for capital—was 16 months. MT

R. Scot Foss is president of Plant Air Technology, P.O. Box 470467, Charlotte, NC 28247; telephone (704) 844-6666. He is the author of "The Compressed Air Systems Solution Series," 1994, Bantra Publishing; telephone (704) 372-3400

Preventive maintenance of a hydraulic system is basic and simple and if followed properly can eliminate most hydraulic component failure.* However, many hydraulic systems are not designed to facilitate maintenance work. A properly designed hydraulic reservoir and the use of a filtration pump can increase maintenance efficiency and increase equipment reliability.

Modifications to an existing hydraulic system need to be accomplished professionally. Here are some recommendations on what should be included.

Filtration pump with accessories
The use of a properly designed filtration pump will reduce contamination introduced into the hydraulic system when fluid level is topped off or when new fluid is added.

Hydraulic fluid from the distributor is usually not filtered to the requirements of an operating hydraulic system. This oil is typically strained to a mesh rating but not filtered to a micron rating. However, hydraulic fluid must be filtered to 10 microns absolute or less for most hydraulic system (25 microns is the size of a white blood cell, and 40 microns is the lower limit of visibility with the unaided eye).

Many maintenance organizations add hydraulic fluid to a system through a contaminated funnel and may even use a bucket that has had other types of fluids and lubricants in it previously, without cleaning them.

Recommended equipment and parts:

• Portable filter pump with a filter rating of 3 microns absolute.
• Quick disconnects that meet or exceed the flow rating of the portable filter pump.
• A ¾-in. pipe long enough to reach the bottom of the type of container the distributor uses to deliver fluids.
• A 2-in. reducer bushing to ¾-in. npt to fit into the 55 gal drum, if you receive your fluid by the drum. If you receive fluid in larger quantities, mount the filter pump assembly to the supports of the double wall tote tank.
• Reservoir vent screens should be replaced with 3/10-micron filters and the openings around piping entering the reservoir should be sealed.

Designing a frame that will allow the filter pump and fluid drum to be transported by fork truck could further enhance the fluid handling operation. Regulations require that secondary containment be addressed. The assembly should include a catch pan so that any fluid any spilled fluid would "leak" into the pan.

Hydraulic reservoir features
A well-designed hydraulic reservoir will minimize the risk of introducing contamination when oil added to the system or contaminates being allowed to enter through the air intake of the reservoir. A valve should also be installed for oil sampling.

The air breather strainer should be replaced with a 10-micron filter if the hydraulic reservoir cycles. (The breather should be sized to the output of the reservoir.) A quick disconnect should be installed on the bottom of the hydraulic unit and at the ¾ level point on the reservoir with valves to isolate the quick disconnects in case of failure. This allows the oil to be added from a filter pump as previously discussed and would allow for external filtering of the hydraulic reservoir oil if needed. Install a petcock valve on the front of the reservoir that will be used for consistent oil sampling.

Recommended equipment and parts:

• Quick disconnects that meet or exceeds the flow rating of the portable filter pump.
• Two gate valves with pipe nipples.
• One 10-micron filter breather.

Do not weld on a hydraulic reservoir to install the quick disconnects or air filter.

Maintenance of a hydraulic system is the first line of defense to prevent component failure and thus improves equipment reliability. These equipment modifications can enhance that effort. MT

Ricky Smith is president of Technical Training Div., Life Cycle Engineering, Inc., 4360 Corporate Rd, Ste. 100, North Charleston, SC 29405; (843) 744-7110

*Preventive maintenance issues were discussed by the author in a previous article "Developing PMs for Hydraulic Systems".

Is it because many maintenance managers, supervisors, and technical staff fail to recognize the important role they play in the strategic goals of the organization?

Having been involved with hundreds of maintenance organizations and their respective employees over the years, I have come to recognize a fundamental difference among them.

There are Leaders, Fast Followers, Slow Followers, and Laggards.

The Leaders are very open minded and willing to take risks. They are entrepreneurial in making things happen. They recognize that their role as a maintenance manager or supervisor is to constantly challenge the status quo and look for ways to improve their contributions to the balanced strategic objectives of their company.

The Fast Followers are the ones that do not want to take a lot of risk on their own, but look to those they regard as Leaders and follow them. This reduces their risk because they can learn from the mistakes the Leaders made, but there is still risk involved because they often do this before all the results are in.

The Slow Followers will wait until the Leaders and Fast Followers have adopted something and proven that following suit will give them a competitive advantage. They are risk adverse and want to have lots of information about how to accomplish a project successfully and the results they can expect. Often, by the time they are ready to adopt it there are many companies able to help them, unlike the Leaders who probably had to figure it out on their own.

The Laggards are the ones that do not accept change very well and, even if they do finally decide to follow suit, are bound to not be highly successful in implementing it. This is because they often follow suit reluctantly and therefore do not give the project the resources required to make it a success.

I have found that 5 percent of maintenance organizations are Leaders, 20 percent are Fast Followers, 50 percent are Slow Followers, and 25 percent are Laggards.

It is time to change this; it is time for maintenance to stand up and get bold on how it is very important to the strategic objectives of the organization.

Start by identifying an area of improvement. It could be a new technology that will allow you to enhance your current computerized maintenance management system (CMMS) without replacing it. Management may find it more palatable to add a new system that will provide additional value rather than replacing an existing system with a similar system that does basically the same thing.

Perhaps implementing a new advanced maintenance methodology such as RCM will help insure your company meets objectives. Maybe you can identify a recent, high-profile issue in the organization. For instance, maybe you have experienced poor quality during production recently so your yield rate has dropped. Determine how maintenance may be able to impact this problem.

Once you have identified the project you want to implement you need to create a clear business case and a maintenance strategy to achieve it. This requires research. Talk to your peers. See who else had a similar problem and find out what they did to resolve it. Talk to suppliers who may have products or services that can help you because they can put you in contact with customers who have solved this problem and give you an idea of the effort required and the results you can expect.

Remember, this is a business case that is going to require approval potentially from several levels above you. You need to speak their language. You need to show them that by doing this they will gain something that they want (such as increased revenue, more production output, or higher quality) or they will avoid something that they don't want (such as safety problems or environmental issues). You also need to show them that you have a clear strategy to ensure the success of the project.

It is time for maintenance management to recognize that we are business people, too. We need to take an active role in helping shape our organization and improving it. We need to show by example how we can make a difference. Senior management will never take maintenance seriously until we ourselves do. MT

Employing condition monitoring equipment for vibration analysis and operational deflection shape, he first tested Machine No. 1 and quickly discovered a resonance problem at the current machine operating speed. Then he detected that Machine No. 2 was out of alignment by a quarter of an inch. He arranged for repairs to be made on both machines. Over the next few months, he diagnosed serious problems in the eight remaining machines and initiated repairs on them as well. Within six months, the technician had saved the lumber plant $1.5 million, based on previous maintenance expenditures and downtime costs associated with the 10 problem machines.

Although dramatic short-term savings are unusual, reliability programs can have a major financial impact over the long term. The most effective programs emphasize training, precision repair skills, and a proactive approach on the part of maintenance and other departments.

For companies planning or implementing a reliability program, there are six important steps for improving the reliability of production equipment.

1. Train top managers, maintenance technicians, and other personnel regarding reliability issues.

Ideally, reliability training should involve every department level in a plant, including top management. The plant manager and key department heads should receive training on basic reliability concepts and on how reliability programs affect the bottom line. Managers and supervisors need to understand how to implement reliability programs, and how to monitor and document results. Also, maintenance managers and technicians need to acquire the specific skills necessary to perform precision maintenance. See accompanying section "A Wealth of Training Options."

Early in the training process, participants from the various departments must reach agreement on program goals and on the meaning of reliability. Broadly defined, reliability consists of maximum functionality from bearings and/or machines with a low incidence of failure at the lowest possible cost. But terms like "maximum functionality" and "lowest possible cost" have different meanings in different plants and even within different departments of the same plant. Program participants should formulate their own definition of reliability reflecting the realities of a particular industry and plant environment.

Equally important is defining failure. The meaning of failure can vary from industry to industry, and even from department to department within the same plant. For example, consider a pump that fails and is quickly replaced with a standby unit without interrupting production. Maintenance might regard this event as a failure because it involves an unanticipated pump repair. The production department, on the other hand, might not term it a failure because there was no loss of production. It is important, therefore, to reach a plant-wide consensus on what constitutes failure. A sample definition might be: Failure occurs when a bearing or machine fails to meet 100 percent of its functional requirements during regularly scheduled operating times. Under this definition, the pump replacement would constitute a machine failure. Agreed-upon definitions and program goals should be documented and made available to key managers, maintenance supervisors and staff, machine operators, and other interested parties.

2. Employ condition monitoring equipment to track bearing and overall machine health.

Condition-based maintenance is a critical component of most reliability programs. Condition monitoring technologies include temperature and vibration monitoring, infrared thermography, ultrasonic noise detection, and operational deflection shape. Employed properly, these technologies can identify bearings or machines on the verge of failure and enable the scheduling of necessary repairs.

The use of vibration monitoring in particular has increased dramatically over the past 10 to 15 years. Vibration devices, linked with trending software, can quickly pinpoint changes in bearing or machine conditions. Current advancements include wireless systems and industrial decision support software. Moreover, vibration amplitudes can be a good indicator of overall maintenance effectiveness; high amplitudes are often a sign of poor maintenance practices.

Although an important tool, condition monitoring alone cannot guarantee precision maintenance. In fact, some condition monitoring practices can even be counterproductive. For example, consider a plant that sets a vibration alarm limit of 0.3 in/sec for its industrial fans. A fan that registers 0.29 in/sec will pass the plant's vibration test, but it still may run roughly and fail prematurely. The alarm limit in this case is a minimum requirement for operation, not a precision specification. To avoid this problem, always establish two sets of vibration criteria—an upper alarm limit that signals imminent trouble and a lower precision specification. In this case, the precision specification might be as low as 0.06 in/sec. Fans and other machines maintained to precision specifications have a much longer life expectancy than those maintained to alarm-limit specifications.

3. Conduct failure analysis by inspecting failed machine components.

To an alert maintenance staff, failed machine components can provide valuable information. For example, small indentations in bearing raceways and rolling elements usually indicate an ineffective sealing arrangement, whereas an OD surface that has a worn, mirror-like appearance indicates that the bearing was spinning in the housing. Other telltale evidence includes flaking, smearing, and abnormal wear patterns.

With proper training and experience, maintenance technicians can learn to recognize these signs and diagnose machine problems. In more difficult cases, machine components can be returned to the manufacturer for expert analysis.

4. Pinpoint the root cause of machine problems.

Distinguishing the root cause from various side effects and symptoms often requires careful analysis. For example, a bearing installed on an oversized shaft eventually will run hot, causing the bearing lubricant to degrade. A cursory examination of the bearing arrangement might point to the degraded lubricant as the cause of failure. More careful analysis, however, would determine the actual cause the oversized shaft.

Once the source of a problem is identified, the maintenance staff can target its efforts effectively and begin corrective action. After corrective action is taken, the results should be verified to confirm the findings of the root cause analysis.

5. Practice proactive maintenance by making precision repairs.

A cement producer used a proactive approach recently in repairing a large induced-draft fan with a history of problems. The fan, driven by a 2600 hp dc motor, evacuates hot gases from a cement kiln. Temperatures in the application typically reach 650 F. Although the fan had a speed rating of 1200 rpm, for many years it operated at 1100 rpm because of a serious vibration problem. The reduced fan speed limited the kiln to only 80 percent of capacity. During planned maintenance every March, the fans bearings were replaced regardless of condition and the fan was balanced, but the vibration problem persisted.

Finally, as part of a new proactive maintenance program, the plant arranged for a series of highly accurate bearing position measurements to be taken using infrared technology. Readings were taken in both the cold and hot-running state. Analysis showed that thermal growth during fan operation caused misalignment of as much as 0.040 in. between the fan and the motor. Based on these measurements, the fan and motor were realigned to account for thermal growth. Subsequently, the fan was brought up to its full operating speed of 1200 rpm without an increase in vibration. Kiln capacity and throughput also increased. In addition, during the next shutdown period, the fan's bearings were inspected and did not require replacement, saving labor and bearing costs.

6. Establish key performance indicators to monitor program success.

Key performance indicators are the hard data used to gauge a reliability program's effectiveness. There is no single, all-purpose indicator that works in every situation. Instead, program supervisors should select one or more key indicators that are well suited to a particular industry or plant environment.

Petrochemical plants, for example, typically have standby machines and redundant equipment, which are brought online quickly in the event of failure. Machine availability levels of 99 percent or higher are the norm. Here, gauging the effectiveness of reliability programs simply on the basis of machine availability data would be difficult. Annual repair costs or mean time between failures would be better indicators of a program's effectiveness.

Paper manufacturers, on the other hand, rarely have redundant paper machines. If a machine fails, production halts until the machine is repaired and returned to service. When failures occur, availability levels drop. Here, machine availability would be an excellent measure of reliability. It also might be appropriate to have a variety of key performance indicators for auxiliary and ancillary equipment. MT

Information supplied by SKF USA Inc., 1510 Gehman Rd., Kulpsville, PA 19443; (888) SKF-2000; fax (215) 513-4736. For information on training programs, contact SKF Reliability Maintenance Institute, SKF Industrial Services Center, 4392 Run Way, York, PA 17406; (717) 751-2900; fax (717) 751-2901

If someone claims he can implement a computerized maintenance management system (CMMS) in six months or less, there is a good chance something key to the implementation is missing. Most well-planned CMMS implementations require 18 months or more of hard work and can cost the organization more than two to three times the purchase price of the application.

This article will not explain how to implement a CMMS in six months or how to do it on a shoestring budget. Rather, it will identify steps required to successfully integrate a system into an organization''s unique technical and business environment. It will identify an expeditious process that eliminates unnecessary steps and leads to more complete system use. It also will identify common problems that can derail implementation and explain how to avoid them.

Data preparation
Preparing the data to include in a CMMS is the first and most important part of implementation. Data can originate from one of two sources: an existing manual system or an existing application. In either case, data preparation is a huge effort that relates directly to the amount of time required for implementation and its success.

The time required for preparing data from a manual system depends partly on the number of assets (or equipment) to be managed. In addition, organizations without a CMMS may not have a numbering system for assets, parts, and materials. That means information about every equipment/asset, part, and material must be gathered and organized. The equipment/assets should be classified in terms of criticality and then aligned into a hierarchical structure for reporting and ease of use. At the same time, information should be collected about job plans, personnel, and other areas the system will log or track. Waiting to start this organization until the beginning of the implementation phase will automatically extend the time frame.

If starting with an existing system, the data organization effort will focus on ensuring complete and accurate data. If massive amounts of information must be standardized, the data preparation phase can easily surpass the time required to start from a manual system. In addition, systems more than 15 years old may not contain major pieces of data required in the new system, or the data may not be in a form that is conducive to conversion. These factors can extend the implementation time frame. Comprehensive and accurate data gathering is the foundation of a successful CMMS implementation and cannot be cut short. So, what can be done to shorten the time frame while maintaining, or even improving, the integrity of the data?

Pre-design. One way to shorten and streamline implementation is to start data collection early. In fact, once the request for proposal (RFP) and demonstration steps of selection are complete, the format and content to be used will be known. To get a tremendous jump on implementation, collect data into an electronic form that identifies the specific data and layouts needed. The data layouts also should be used during contract negotiations to provide the selected vendor with an idea of the volume of data and its initial format. This will help the vendor map data to the new system for electronic data transfer. An experienced CMMS implementation consultant can help start the data collection effort even before vendor selection by providing general guidelines for data formatting as well as his insight and guidance.

In many cases, problems that users blame on the system are actually problems with data. Data requires verification and validation to avoid many of these problems. Verification is the act of checking the data. Does the data represent every asset? Is it accurate? Is the data worth the time and money to prepare it for the new system? If not, a substantial data collection and preparation process may be necessary. Validation is the process of cross checking referential data when it enters the system. There are tools that assist in advanced validation. Remember that data must follow the integrity requirements of the new system to qualify for vendor support.

Another potential pitfall is the coding structure used to identify equipment/assets, parts, and hierarchical relationships. Avoid coding structures that use ranges. Eventually, range-based systems become saturated╛usually when the organization can least afford it. Workarounds, such as adding a new range of numbers, will eventually require another workaround and the problem will snowball. The best coding structures provide the flexibility to add large amounts of data in the future.

Before starting data collection, be sure that data standards are in place and that they will be adhered to. Data standards should consider and include:

• System identification coding structures
• Abbreviations
• Descriptive fields
• Units of measure
• Equipment criticality
• Prioritization

The forethought put into standards development and enforcement will impact system usability. If users cannot easily find accurate information in the system, they will stop using it. Ultimately, poorly constructed standards or lack of standards can derail the implementation.

System and process integration
Software itself is not a solution. Organizations that try this approach almost never achieve the maintenance benefits they set out to obtain. Because their business processes never change, this approach simply automates existing problems so that they occur faster than before. In these cases, the implementation will lose momentum or stall completely. When implementations fail, people remember for a long time. Implementing any new system in the future will be even more difficult.

The implementation success rate is improved substantially when the maintenance organization integrates processes, personnel, and the CMMS. This approach enables organizations to plan and schedule more than 60 percent of all maintenance activities instead of constantly reacting to problems. This approach enables more work to be accomplished with less effort, making complex systems more manageable and reliable.

Achieving this proactive state requires a concerted effort to integrate the functional capabilities of the selected system with the way the maintenance organization works. It also means that a reactive organization must reassess and redesign its processes and procedures. Whether the organization is reactive or proactive, this is the time to take a hard look at existing processes and procedures to redesign them for maximum effectiveness.

The best approach is to redesign practices, processes, and procedures prior to, or during, system selection. This means prospective systems can be evaluated in context of the redesigned processes. On the other hand, knowing the system that has been selected will allow the organization to design its business processes to maximize the new systems capabilities. If implementation has already begun, spend the extra time to integrate the system and maintenance processes before going live.

Change management
Many system implementations fail because of human factors such as changed responsibilities and the natural resistance to change. This makes managing expectations an essential part of implementation.

Organizations go through three distinct periods of transition from the way business has been conducted to a new manner of organization:

• The Frozen period describes the existing business processes, "the same old grind."
• The Fluid period is the time when the organization is in transition from old processes to new, streamlined, standardized processes. This period officially starts by outlining processes and changes and describing how the new software and systems will be integrated. During this period, people hear that their jobs may be changed or eliminated or that new people may be joining the organization.

Staff reactions vary. Generally, 20 percent accept change, 60 percent wait and see, and 20 percent strongly resist. During this period, managers and supervisors will hear from the vocal minority who resist change. Reactions usually start during the selection process and become louder as the implementation progresses. These people will be unhappy with the changes and will need coaching and assistance from their managers.

The success of the implementation is directly impacted by how well change is managed throughout the period, so managers must be prepared and know how to manage through the change. It is essential to continue change management activities and user support throughout the implementation to maintain continuity and momentum. During this stage, staff and managers realize the full impact of change as they experience new tasks and duties. It may result in losing people who will not accept the new way of doing business.

• The Re-Frozen period starts after the new software and business processes have been used for a period of time and have been adopted as the new way of doing business.

System installation and acceptance testing
Most implementation efforts focus on installation and testing. For the most part, these steps are well planned and managed using proven methodologies. However, there are some areas that require special attention to ensure success.

Integrated testing. The software and any new hardware should be installed and tested for all aspects of the organizations technical environment. Testing must include networking, remote-site communication, Intranet and/or Internet connections, and any other technical component the system will impact or require. Most of these tests should be conducted by the vendor or by internal technical staff with vendor oversight and assistance.

Interface testing and data conversion. Once the technical hardware and software are working properly, technical testing should address the interfaces between the CMMS and other business systems. This will ensure that data being transferred back and forth is properly processed and validated. Typically, a vendor will develop interface processing that affects its system because it knows the system and its needs. Data conversion from an existing system and/or data collected electronically needs to be tested to ensure the integrity of data relationships. Errors must be corrected and re-testing must occur.

Vendor assistance during data conversion and loading will ensure that validation and integrity are in place from the system point of view. Internal technical and user staff (members of the core team and selected assistants) share responsibilities for data conversion and loading. They are responsible for verifying data relationships and integrity from the organization's point of view.

Functional testing. Thorough and complete functional system testing should be finished satisfactorily before the system is accepted as final. Each department and functional position that will be affected by the system needs to perform its own specific testing. The best approach (though often missed) is to test system functions and business processes as integrated units. For example, the system's material management functions need to be tested using each material management business process for each functional position. If the business processes were redesigned, the new process must be tested.

Pilot testing. Pilot project testing is a good method of quality assurance for organizations with large, complex, multifaceted, or multi-plant situations. A single department, plant, or facility that represents the organization or an entity with unique requirements should participate. The selected entity should use the system in full production for a period of time. This approach isolates problems to the pilot entity so they can be identified, corrected, and retested without affecting the larger organization or causing a poor first impression. Once the testing is complete, the system can be rolled out to the remainder of the organization.

To be effective, pilot tests must be planned carefully. There are two approaches: complete software environment testing or restricted software environment testing. Both approaches impact other business systems and the test data. A complete environment test verifies the functionality of interfaces with other systems. This approach is limited by the coordination required to integrate data affected during the pilot with the remaining organizational entities after testing is complete. Before testing begins, consider the following issues: Will test data be discarded after testing (requiring system data to be reloaded from scratch when the system goes live)? Will the pilot entity keep using the system and test data and have other entities join in? How will data sent to other systems be handled? How will other systems that interface with the new CMMS continue to process data for other organizational entities during the pilot?

The restricted environment approach isolates the test environment and data to the system and the test entity. In this approach only the CMMS is tested for the pilot area. While easier, this approach will not test the functionality of the interface with other systems.

Reporting needs. The final step of system installation and acceptance testing is to address reporting requirements. Most CMMS vendors provide a core set of reports. Because vendors are not in the business of report writing, their offerings are usually quite basic because every organization wants variations to the standard reports. It is best to use a good ad hoc report writing/ query tool. New reports can then easily be developed to match specific needs. Only reports absolutely required at startup should be identified. This is a good time to eliminate reports that are never read. Always determine whether required reports will be provided by the vendor or by internal staff. A subset of reporting is performance measures and key performance indicators. These should be identified early in the project and provided at startup to document and show progress.

Organizational roles and training
There are essentially four groups of individuals in an organization who need training: server system administrators, client system administrators, organization managers, and system end users. Resource requirements and roles need to be considered early in the project for budgeting and implementation planning.

Server system administrators. Server system administrators maintain technical system functionality related to networks and network performance, system servers and other hardware, system software, e-mail and other office automation software, PC hardware and software, database, security, report writing, data storage, and disaster recovery. These administrators provide internal technical support before involving the vendor.

Their training should cover all technical aspects of the selected CMMS. They should learn where their responsibilities start and end, what they can and cannot do related to system customization, and how to escalate problems through the vendor's technical support center.

Server system administrators will be used nearly full time approximately six months prior to startup and one to two months after implementation. Their time can generally be reduced to quarter time after startup. Exceptions will be for report writing, system support, and new software version testing and rollout.

If the system must be tailored or customized (refer to prior articles if you are tempted), or additional interfaces between the CMMS and other business systems are required, the number of resources and required skills will expand accordingly. Customization requires expert software development skills with the specific CMMS application through advanced training and/or certifications. This is not the best use of these limited resources that usually support other important systems.

Client system administrators. The client system administrators provide the first line of support to system users as well as user training and problem resolution, standards enforcement, and security administration. If there are multiple sites, there should be a client system administrator at each facility. The administrator should learn to operate every aspect of the system to its full potential. These administrators also provide staffing backup for emergencies and other absences. Administrators should have in-depth knowledge of:

• The CMMS selected
• Data use processes
• System security policy
• Software administration processes
• Organizational standards and procedures

Client system administrators should be trained close to system start up. Provide adequate time for them to be involved in system acceptance testing and to prepare internal user training. They will be required approximately full time for the first two years of system use and half time or less after that.

Organization managers. Throughout the organization, managers influence users attitudes toward a new system. To help these managers buy in to the new system, explain why the organization needs the new system and processes, why this particular system meets the organizational objectives, and what business and personal benefits they can expect. Keep in mind that personal benefits may be different from business benefits. It is equally important to train managers on ways to manage change including gaining staff buy in and overcoming objections.

System end users. System users are the final, largest group to be trained. They also have the greatest impact on system use and its ultimate success. Training should be based on functional modules that address the business processes and system functionality needed to perform specific tasks for a functional position. Functional position training should address general system operations, navigation methods, and the basic requirements for the functional position. This provides end users the ability to learn system capabilities related to their work. Unnecessary information will result in information overload and less-than-effective system use. Training should occur immediately before users go on line to keep information fresh in their minds. It is also the foundation for advanced training and followup remedial training after startup.

Support and startup
There are three areas of system support to consider for implementation and ongoing production.

Application software support includes the activities surrounding the installation of software (whether for initial installation or future releases). The CMMS vendor should be used as a resource for general application software support, providing backup for issues or problems that the organization's system administrators cannot handle. Most vendors try to issue at least one major product update each year with new features and technological advances. Using the vendor for system support (under the software support agreement) will free up internal technical resources to concentrate on other areas.

Technical support helps ensure effective, reliable interfaces between the CMMS and other systems after implementation. Other technical support includes system operations (backup, etc.), database administration (performance analysis and fine tuning), system security at the system and database levels, and server software administration. Server system administrators deliver technical support.

User support and training provides general end user application support and helps system users conduct specific activities efficiently and within the organization's standards and practices. This support should consist of a person or persons with in depth knowledge of the CMMS, usage standards and practices/processes that are in place, system security, software administration, and data management. Their activities include training, individual support, and problem resolution for users. Client system administrators deliver application and desktop support and training.

Inevitably, problems arise when the system and processes go live. Often, hardware vendors blame the software and software vendors point to the hardware or the data. Meanwhile, the organization loses valuable uptime and momentum. It is also easy to lose track of problems during the heat of implementation activities or to anticipate that someone else is handling a problem. Keeping a log of problems throughout implementation is the foundation for clear communication to move all parties toward resolution instead of wasting time assigning blame. The log captures information about the circumstances and helps assign responsibility for the correction. Again, it is important to ensure that user data is not the culprit. The log should identify:

• Problems encountered (what was attempted, what occurred, when, where, under what conditions, how many times, who has had the problem)
• Resolution (how each problem will be corrected)
• Seriousness (showstoppers need immediate identification and resolution)
• Responsibility for resolving and correcting each problem
• Ongoing status

The log and outstanding problems need to be reviewed regularly and frequently.

Post implementation review
This is the last part of implementation and the final step in the transitioning from the Fluid period to the Re-Frozen period. The review should be conducted after three to six months of full operation. The exact timing should be determined by how smoothly the new system operates. If there are many problems, an earlier review may be necessary. The purpose of the review is to identify and correct problems with both system and business processes.

System problems should be directed to the software vendor. These should be documented thoroughly in the problem log. Serious problems are problems that stop the system from functioning and/or negatively impact other systems. These need the vendor's immediate attention and resolution. Be sure to verify that the problem is not caused by erroneous data. If the vendor recommends a workaround, be sure to obtain the schedule for a final correction. The vendor may have to make system changes to correct the problem. Acceptance testing may be necessary depending on the nature of the changes made. Typically, these types of problems are more likely to occur after the installation of a new release or update.

Business process problems are generally caused by conflicts between the way the system functions in a specific area and the way business is conducted, or the need for remedial training.

Process/system conflicts must be resolved by changing the way the organization functions or altering the way the system operates. The latter can be achieved by changing system master data (not likely) or changing the configuration of the system. It is best to work with the vendor to reach a resolution. There may be cases where tailoring or customizing the system is necessary. However, this should be a last resort. If customization is required, have the vendor incorporate the changes into the system. You may need to pay for the changes. But having the vendor change the system should ensure that the changes do not invalidate the support agreement╛making you responsible for system maintenance and hastening obsolescence.

CMMS consultants may be helpful during implementation. They should provide guidance, direction, and professional input to the organization through the entire effort including problem logging and resolution. Consultants also should be responsible for facilitating the post implementation review due to their expertise and independent perspective.

Be sure to choose a consultant who is versed in maintenance and material management in addition to software selection and implementation. Many consultants are excellent technicians but lack experience in business process and change management or vice versa. They should anticipate pitfalls and barriers to success and guide the organization around obstacles.

This implementation plan, along with a formal selection process, will help ensure:

• The avoidance of common implementation pitfalls
• The earliest possible startup
• The complete usage of the new system
• Properly loaded, verified, and validated system data
• Data conformance to standards
• Achievement of expected benefits and ROI for the organization

The organization is now ready, set, and prepared for successful CMMS implementation. MT

Previous articles in this series include "Successful CMMS Implementation: Getting Your House In Order" (MT, 9/01, pg 14) and "Avoiding Pitfalls in CMMS Implementation" (MT, 7/01, pg 15).

Derold Davis and Joe Mikes are senior consultants at Westin Engineering; (916) 852-2111. They both have more than 15 years of experience in providing system selection and implementation methodologies, proven maintenance practices, productivity improvement practices, and methods and strategies for increasing operational reliability and reducing maintenance overhead.

Robert C. Baldwin, CMRP, Editor

Early this month I was able to hear all three futures discussed at a workshop on Tether-free Technologies for e-Manufacturing, e-Maintenance, and e-Service at the University of Wisconsin–Milwaukee. The event was produced by the Center for Intelligent Maintenance Systems, a National Science Foundation Industry/University Collaborative Research Center.

The attendees covered the spectrum: academia, government, industrial maintenance, research, equipment manufacturing, maintenance services, and software developers.

Three agendas were represented: people interested in learning what may be possible in the future, people looking to gauge the consensus of attendees on how things are likely to turn out, and people interested in boosting their preferred view of the future. Most attendees were working all three.

Workshop presentations covered the benefits of tether-free or wireless technologies in various industries, emerging technologies, and standards.

I was fascinated by various viewpoints represented in the breakout session on Emerging Technologies: Needs in e-Maintenance. All revolved around using technology to increase maintenance effectiveness while reducing its cost, with each constituency plotting a way to add the savings to its own bottom line, possibly at the expense of the others.

But all agreed that the formula for success is based on knowledge—knowledge of business processes, manufacturing processes, and reliability and maintenance processes. The trick, they said, is educating enough people in the enterprise about plant asset management so they see the benefits of investing in the technology needed to make it work.

In other words, technology by itself is not worth very much. Its value is derived from its ability to drive the enterprise business model toward an agreed upon preferred future.

I came away from the meeting vowing to remember that although people may agree that certain technologies are beneficial, their reasons for investing it may be vastly different. This suggests that you have to get to know the people you deal with and learn their interests before you can expect to appeal to them to support investment in the processes and technologies that are important to you. MT