Heightened reliability and safety concerns have resulted in new maintenance programs and techniques
Every year, additional pressures are placed on operation and maintenance managers to provide a higher level of power system safety and reliability at lower costs. The greatest difficulty for many of these managers is to design, customize, and optimize a comprehensive preventive maintenance (PM) program. Taking a management systems approach first, rather than task specific approach, will provide a strong foundation upon which the elements of a good PM program can be built.
In this article, we'll discuss a broad range of PM programs. We'll also review the various elements of a PM program as well as benefits and common problems.
The vast majority of our industrial facilities have been operating for extended periods, in many cases in a reactive and corrective maintenance mode. The few that have graduated from breakdown maintenance to preventive maintenance have done so for three primary reasons: To prevent failure; to detect the onset of failure; and to discover a hidden failure.
Four task categories, by one name or another, are universally employed in constructing a PM program.
* Time-directed (TD), aimed directly at failure prevention or retardation.
* Condition-directed (CD), aimed at detecting the onset of failure or symptoms.
* Failure-finding (FF), aimed at discovering a hidden failure.
* Run-to-failure (RTF), a deliberate decision to run to failure because the other choices are not possible or the economics are less favorable.
There's still a widespread myth in the maintenance community that all failures can be prevented. This feeling often motivates the use of overhaul tasks without questioning or understanding the fundamentals of the failure mechanism involved. Additionally, maintenance intervals and task performance are generally based on the greatest hodgepodge of ad hoc data and reasoning of why it is done this way! Here, the basic concept of reliability engineering is usually omitted from maintenance program development.
In recent years heightened concern for plant reliability and safety coupled with the availability and quality of standards have impacted the evolution of new maintenance programs and techniques. Enhancements to traditional maintenance programs include predictive maintenance, trending analysis, and condition monitoring.
The ANSI/NFPA 70B Standard, Recommended Practice for Electrical Equipment Maintenance, addresses why a PM program pays dividends. Additionally, NFPA 70B recommends the typical frequency intervals for electrical PM tasks (a valuable base line for the development of a new PM program).
In recent years, insurance companies have also taken a much closer look at the costs associated with electrical failures. The casualty loss analysis shown in Table 1 is part of a risk management program to reduce the number of insurance claims and illustrates the significant cost due to inadequate maintenance.
A Hartford Steam Boiler Insurance Company study on transformer reliability involved several hundred transformer failures, as shown in Table 2, in a recent five year period. Only liquid filled transformers larger than 100kVA (excluding arc furnaces or rectifier transformers) were analyzed. The median age at time of failure from all causes was 6.4 years. From the identified causes of failures, the study concluded two key recommendations.
* Periodically check lightning/surge arrestors and ground resistance to minimize the effects of abnormal voltage.
* Maintenance has to bear the blame for not discovering incipient troubles. The time and expense involved in periodic preventive maintenance is well justified.
Common Maintenance Problems
The industry wide maintenance history of the past three decades shows a trend of common problems associated with maintenance practices.
Insufficient proactive maintenance. Here, the vast majority of plant maintenance staff operate in a reactive mode and, in some instances, the plant management has a deliberate philosophy of corrective maintenance. This approach can also lead to a breakdown-centered maintenance activity.
Frequent problem repetition. When operating in a reactive mode, little time is spent identifying and analyzing repetitive failure problems. This lack of root cause analysis makes the repair and restoration a temporary measure at best.
Erroneous maintenance work. In corrective and preventive maintenance activities, errors will occur. Recent studies have shown that human error is the cause of more than 50% of plant forced outages, and that some form of human error might be occurring in some locations in one of every two maintenance tasks performed.
Sound maintenance practices not institutionalized. The industry has a great deal of knowledge and experience on how equipment should be maintained. Individual plants, however, are usually informed only on a small percentage of this collective knowledge. Additionally, the infrequent commitment to a formalized set of standards, procedures, and training will worsen the effectiveness of the maintenance program.
Unnecessary and conservative PM. Historical evidence strongly suggests that some of our current PM activities are, in fact, not right. In some cases, PM tasks are totally unnecessary because they have little relationship with keeping the plant operative. It's not uncommon to examine a plant [TABULAR DATA FOR TABLE 1 OMITTED] PM program and find that 5% to 10% of the existing tasks could be eliminated. In many cases, the PM may be right, but the frequency is too conservative. This seems to be especially true of major overhaul tasks, where many observations suggest that 50% or more of the PM overhaul actions are performed prematurely.
Poorly defined rationale for PM action. The absence of information on PM task origin or any documentation to clearly trace the basis for plant PM task is generally standard operating procedure. The Federal Aviation Administration for aircraft maintenance and the Nuclear Regulatory Commission for nuclear plant maintenance have recognized the need for a more formalized documentation process.
Maintenance program lacks traceability and visibility. In many cases, there's a lack of a definitive maintenance management information system. Even worse, the performance of root cause analysis on equipment failure and recording the basis of PM actions are often nonexistent. These deficiencies point toward the benefits of base line comparison (bench marking) and computerized PM software programs.
Blind acceptance of OEM input. Original equipment manufacturer (OEM) recommendations are generally conservative and not necessarily applicable to the plant's general profile. The broad range of equipment applications (e.g. cyclic versus steady state, very humid versus dry ambient conditions) will generally impact the frequency of PM tasks rather than the type. Additionally, a certain level of conservatism may be introduced to protect the manufacturer in the area of equipment warranty.
PM variability between like or similar units. Large corporations having multiple plant locations often have different operation and maintenance characteristics at each plant. At the plant level of organization, there appears to be a strong feeling of "I am not like them," "We know what's best for us," and "They don't know what they're doing." These attitudes are driving the lack of commonality and consistency. Obvious cost savings could be achieved through standardized procedures, training, spare parts inventories, etc.
PM Program Development
When developing or upgrading an existing PM program, we need to determine first what we would really like to do in the program and second, take the necessary steps to build that ideal program into the infrastructure. This approach, as shown in Fig. 1, prevents the restrictive alternatives of "second best" choices in the early stage of program development. The next phase is to integrate the program into the existing infrastructure to the maximum extent possible (integrate ideal program into the real world).
PM Program Elements
The previous discussion is a simplistic view of how a PM program is developed. Let's review the management and technical disciplines required to implement the ideal PM program and PM task packaging.
A PM program can be created or upgraded through several key technologies and program management systems. Each must be interactive and complementary. Fig. 2 is an expansion of the "What Task" and "When Done" of an ideal PM program shown in Fig. 1.
There is a multitude of supporting technologies available today. Although a detailed description of these technologies is beyond our scope here, some of the major groupings are as follows.
Failure analysis technology. When a failure occurs, a significant learning opportunity is presented to us. To take advantage of this, it's important to conduct a comprehensive program for failure reporting, root cause analysis, and corrective action feedback. An added benefit is that a good failure analysis program is also a vital element in retaining or increasing the Mean-Time-Between-Failure (MTBF) portion of an availability improvement program.
Incipient failure detection. Behind the ability to utilize condition-directed (CD) tasks aimed at detecting the onset of a failure or failure symptom, an entire diagnostic technology is constantly evolving. This area is generally called Predictive Maintenance Technology and involves technological tools such as thermal imaging, acoustic leak detection, non-intrusive flow measurements, etc. Other terminology including condition monitoring, monitoring and diagnostics, and performance monitoring are all intended to describe the measurement of parameters in a non-intrusive manner in conjunction with trending these results over time.
Information management. The complexity of systems and the availability of computers have developed a powerful synergism often referred to as a Maintenance Management Information System (MMIS). These systems automate the collection, storage, and processing of vital data to obtain the optimum operating efficiency, safety, and reliability of a system.
In some cases, elements of this MMIS include advanced software technologies such as "expert systems," fuzzy logic, and "smart systems" through a neural network. Fig. 3 shows an example of a modular MMIS program. A typical MMIS will incorporate many of the following features:
* On-line for immediate access.
* Database driven.
* Capacity to hold master files (equipment, inventory).
* Store at least one year of history.
* Automatically schedule PMs by calendar and/or usage.
* Allow flexible planning and scheduling periods.
* Permit work order status reporting.
* Forecast trades/craft requirements.
* Produce maintenance history reports with costs.
* Track spare parts and automatically flag reorders.
* Generate purchase requisitions and track purchase orders.
* Transfer data to/from mainframes.
* Provide a multi-user capability.
* Modular to allow phased implementation.
* Predictive maintenance.
* Corrective maintenance requests and orders.
* Bar coding for data capture.
* Performance trend analysis.
* Failure analysis records.
* Condition-directed task measurement and criteria, alerts.
* Skill requirements vs. skill availability.
PM task packaging. There are three major elements to task packaging.
* Task Specification. The task specification is the formal document by which we assure that a complete technical definition and direction is provided to the implementing maintenance organization.
* Procedure. This document will guide the field execution of a PM task. The procedure must be very detailed on how the PM task is to be precisely achieved. These procedures also expand frequently to Standard Operating Procedures (SOP) and Methods of Procedures (MOP). The risk inherent to PM activities can be controlled and greatly reduced by assuring the development of technically sound and complete task specifications and procedures.
* Logistics. Logistics involves a variety of administrative and production support activities. These are also commonly referred to as Maintenance Planning. Typical considerations include spare parts, training, accessories, test equipment, schedule, drawings, etc. It's evident that logistics play a key role with task specification and procedures.
Logical approach to maintenance (a three-phase approach). Not all PM programs require a high level of management system sophistication. A simple, straight forward "logical" approach to PM utilizes three implementation phases.
* Phase 1: Info. Review and Update.
Inventory existing data (i.e. spare parts, records, fault/coordination study, etc.).
Update/check all data (i.e. record system loading, infrared scan, evaluate Standard Operating Procedure, etc.).
Prove the emergency/standby feeders (i.e. confirm loads connected to emergency generator, etc.).
Prove the single line diagram for the non-emergency load (i.e. systematically shut down power while checking the downstream loads, etc.).
Evaluate the system needs (i.e. verify available fault currents against device capacity, etc.).
* Phase 2: Maintenance Implementation.
Establish division of responsibility (i.e. facility personnel, testing company, etc.).
Establish priority areas (i.e. from infrared survey known trouble areas, etc.).
Ensure emergency/standby systems (i.e. load test generator, test transfer switch).
Establish an overall maintenance plan/schedule (i.e. develop a time line, temporary power needs, backfeeds, etc.).
Perform maintenance (i.e. document as found conditions, take pictures, perform tests, repair minor defects, compare/evaluate test results, etc.).
* Phase 3: Selective Repair and Modification.
Determine repair needs from maintenance reports.
Determine parties to make repairs.
Determine power outage requirement for repairs.
Order parts required for repair.
Schedule outage when parts come in.
This three phase maintenance approach can be easily adapted from a commercial high-rise building to a health care facility. In addition, it fosters a disciplined methodology that will reduce the possibility of "last minute surprises."
Specification of maintenance testing
Electrical maintenance testing is performed in house for routine tasks. More complex tasks generally performed in yearly or longer intervals are often subcontracted to a testing agency. A specification statement such as "circuit breakers shall be thoroughly inspected and tested" is so vague that a well-intentioned maintenance testing program involving primary injection and contact resistance testing is reduced to a secondary injection test and, even worse, to a "button-pushing" walk through.
The surest way to obtain the desired maintenance test is to be very specific when writing the specifications. NETA has developed a specification procedure for electrical maintenance testing of power equipment and systems (NETAMTS 1993). This document will allow you to replace vague phrases, like "Check the circuit breaker for proper operation" (which means to some people that opening and closing the breaker a few times is sufficient), with specification sections like the following.
2. Circuit Breakers: Low-Voltage, Power
1. Visual and Mechanical Inspection
1. Verify that all maintenance devices are available for servicing and operating the breaker.
2. Inspect for physical damage. Clean and lubricate as required.
3. Inspect anchorage, alignment, and grounding. Inspect arc chutes. Inspect moving and stationary contacts for condition, wear, and alignment.
4. Verify that primary and secondary contact wipe and other dimensions vital to satisfactory operation of the breaker are correct.
5. Perform all mechanical operator and contact alignment tests on both the breaker and its operating mechanism.
6. Check tightness of bolted bus connections by calibrated torque- wrench method. Refer to manufacturer's instructions for proper torque levels. In lieu of torquing perform thermographic survey.
7. Check cell fit and element alignment.
8. Check racking mechanism.
9. Lubricate all moving current carrying parts.
2. Electrical Tests
1. Perform a contact resistance test.
2. Perform an insulation-resistance test at 1000 volts DC from pole to pole and from each pole-to-ground with breaker closed and across open contacts of each phase.
3. Perform an insulation-resistance test at 1000 volts DC on all control wiring. Do not perform the test on wiring connected to solid state components.
4. Make adjustments for final settings in accordance with the coordination study.
5. Determine minimum pickup current by primary current injection.
6. Determine longtime delay by primary current injection.
7. Determine short-time pickup and delay by primary current injection.
8. Determine ground-fault pickup and delay by primary current injection.
9. Determine instantaneous pickup value by primary current injection.
10. Verify the calibration of all functions of the trip unit by means of secondary injection.
11. Activate auxiliary protective devices, such as ground-fault or undervoltage relays, to insure operation of shunt trip devices. Check the operation of electrically-operated breakers in their cubicles.
12. Verify correct operation of any auxiliary features such as trip and pickup indicators, zone interlocking, electrical close and trip operation, trip-free, and antipump function.
13. Check charging mechanism.
3. Test Values
1. Bolt-torque levels shall be as specified by manufacturer.
2. Compare microhm or millivolt drop values to adjacent poles and similar breakers. Investigate deviations of more than fifty percent (50%).
3. Insulation resistance shall not be less than 100 megohms. Investigate values less than 100 megohms.
4. Trip characteristics of breakers when adjusted to setting sheet parameters shall fall within manufacturer's published time-current tolerance band.
Categories of equipment to be tested
Electrical power equipment are broken down into major categories. These categories are used to make specific references to the maintenance testing requirements. The following are suggested categories.
* Power transformers.
* Distribution transformers ([greater than or equal to] 75kVA).
* LV circuit breakers ([greater than or equal to]100A).
* LV switches ([greater than or equal to] 100A).
* MV circuit breakers.
* MV switches.
* Metal enclosed switchgear.
* Ground fault protective systems.
* Protective relays and associated instrument transformers.
* Meters and associated instrument transformers.
* MV cables.
* Bus ducts.
* Grounding systems.
* Motor control centers.
* Automatic transfer switches.
* Emergency/standby systems.
* Others (specify)
Determination to witness tests versus certified test reports
Since maintenance testing is often performed by a testing agency, you will sometimes elect to witness the field tests being performed. This approach will assure you that field tests are performed in accordance with the maintenance testing specification.
Additionally, in the event that some equipment does not successfully pass the test and discrepancies, it's valuable to have you witness the problem first hand. This will help in coordinating remedial repairs and assigning responsibilities.
A cost effective approach to the more complex maintenance testing activities is to involve your in-house maintenance personnel in the project. Their familiarity with equipment and layout will assure a smooth project. Simple tasks such as torquing, cleaning, and racking of devices can be performed by these people. This team approach to the maintenance program provides a great on-the-job training opportunity and increases the sense of ownership and accountability in the safety and reliability of the electrical power system.
The testing agency will provide a certified test report detailing the equipment tested, procedures, results, recommendations, and data sheets. This document should reflect the entire maintenance testing project. Then, the data may be input into a MMIS program as part of the testing agency's maintenance management service, unless you have the in-house capability. Also, you may supplement the field inspection and testing by witnessing a random sample of each category of equipment being tested.
The determination will be based on the following priorities:
* Criticality of the equipment. Critical and high reliability equipment may justify the test witnessing of the entire line-up.
* Availability of inspection personnel. This limitation will dictate how much to rely on the certified test report.
* Need for verification of equipment performance. Much of today's electrical equipment have performance standards that result in an additional price premium. Efficiency (e.g., motors/drives) and power availability (e.g., generators/UPS) are some examples. This type of equipment may also justify test witnessing.
System function test
A system function test is performed upon completion of the equipment field maintenance test. It's intent test is to prove the proper interaction of all sensing processing and action devices. Witnessing the system function test is probably the most encompassing and crucial activity of the entire testing project.
Qualification of testing agency
The term "testing agency" generally implies a testing contractor, testing laboratory, or a third party. There are five major considerations to qualify the testing agency, each of which are critical to a meaningful testing program.
* Agency qualification. The testing agency should be a corporately independent testing organization that can function as an unbiased testing authority, professionally independent of manufacturers, suppliers, and installers of equipment of the systems evaluated. The testing agency should have been engaged in such practices for a minimum of five years; it should submit proof of this qualification when requested.
* Personnel qualifications. The testing agency should utilize only full-time test technicians and engineers who are regularly employed by the firm for testing services. Electrically unskilled employees should not be permitted to perform testing or assistance of any kind. Electricians and/or linemen may assist, but not perform testing and inspection services.
* Test equipment and calibration requirements. The test equipment should reflect a new generation/technology of test equipment. This requirement assures that the owner's "state-of-the-art equipment will be tested by state-of-the-art test equipment." It testing agency should have a calibration program with accuracy directly traceable in an unbroken chain to the National Institute of Standards and Technology (NIST). Dated calibration labels should be visible on all test equipment.
* Safety requirements. The tests should be performed with apparatus de-energized except where otherwise specifically required. The testing agency should have a designated safety representative on the project. As a minimum, the safety standards should include applicable Occupational Safety and Health Administration (OSHA) requirements, such as 1910.269, 1910.332 through 1910.335, and NFPA 70E, Electrical Safety Requirements For Employee Workplaces.
* Report preparation capabilities. The final report should include a summary of the project: a description of the equipment tested; a description of the test conducted; the test results; conclusions and recommendations; an appendix, including appropriate test forms; identification of the test equipment used; and a signature of the responsible testing agency representative.
Requirement for inspection and testing
You should obtain several vital pieces of information prior to the beginning of the field testing program to assure a smooth coordination of the testing effort. The following is a partial checklist.
* Copy of the latest drawings. The single line diagram is the most valuable tool to the testing agency.
* Vendor data. Include manuals, spare parts list, and equipment accessories, including extension handles, relay test plugs, racking mechanisms, etc.
* Factory inspection and test reports.
* Copy of contracts/supplements.
* Blank copies of test/inspection forms.
* Copy of short circuit and protective device coordination study.
Types of equipment testing
There are five basic types of testing: dielectric, stimulation and response, signature, special tests, and a system function test. Each of these has one common goal: to identify abnormalities or weaknesses such that remedial action can be taken prior to equipment operation. In addition, maintenance testing establishes a base line for future maintenance test comparison. This base line data will become very important in the future when performing trending and predictive maintenance analysis.
Dielectric. Dielectric testing usually requires the equipment to be isolated from potential and consists of the following.
* Insulation resistance (IR) tests.
* Polarization index tests.
* Dielectric absorption tests.
* Hypotential tests (AC/DC).
* Double insulation power factor tests.
* Slot discharge tests.
* Surge impedance tests.
* Impulse tests.
Stimulation and response. Stimulation and response testing may or may not require de-energizing equipment, depending upon the test to be performed. Examples include the following.
* Primary injection tests for CT's and breakers.
* Secondary injection tests for relays and breakers.
* Ratio burden and saturation tests for CTs and PTs.
* Transformer turns ratio and winding resistance tests.
* Breaker time travel analysis.
* Contact resistance tests or ductor tests.
Signature. Signature or real-time tests must be conducted with the system under normal load and voltage conditions. These tests include the following.
* Vibration analysis.
* Thermographic inspection.
* Ultrasonic inspection.
* Transient/harmonic signatures.
* Acoustic emission tests.
* Optical time domain reflectometry.
Special tests. These special tests are non-electrical tests that provide quantitative data for analysis. These tests include the following.
* Insulating fluid tests (surface tension, acidity, power factor, color, etc.).
* Dissolved gas analysis.
* Wear metal analysis.
* Heat rise test.
System function test. This final test follows component and equipment testing using the various techniques listed above. This test proves the interaction of sensing and action devices that must "communicate" to make the overall system result occur. The testing agency should develop and propose the test parameters. These would include considerations such as interlock safety devices for fail safe functions, alarms and indicating devices, initiation of sensing devices, and evaluation of overall system performance.
EC&M ArtiFax Articles:
"Hipot Tests Assure Power Cable Reliability" Jan. 1993. "Nine Ideas to Improve Your Electrical Maintenance" June 1994. "How Can Frequency of Preventive Maintenance Be Extended?" July 1994 $14.95 for a set of articles. Order No. 2200. To order by fax dial 800-234-5709. (Have a credit card and your fax number ready when you call.)
ANSI/NFPA 70B, Recommended Practice for Electrical Equipment Maintenance. NFPA 70E, Electrical Safety Requirement for Employee Workplaces. For ordering information, call 800-344-3555. NETAMTS-1993, Maintenance Testing Specification for Electrical Power Distribution Equipment and Systems. For more information, call 303-697-8441.
Jean-Pierre Wolff is Vice President, Electro-Test, Inc., Pleasanton, Cal., a NETA (interNational Electrical Testing Assoc.) member.