How can power monitoring programs identify degrading trends in the electrical system so inspections and preventive maintenance can be carefully targeted?
Because electronic systems are becoming both more prevalent and more sensitive, and because utilities are deregulating, proactive power monitoring programs offer very attractive benefits. If properly executed, they identify degrading trends in the electrical system so that inspections and preventive maintenance can be carefully targeted to specific equipment, allowing you to avoid electrical disruptions altogether.
There are three steps involved in enhancing a preventive maintenance program through power monitoring.
* Determine your objectives.
* Define the monitoring program.
* Choose the appropriate monitoring system.
Let's talk about each in detail.
Power monitoring program objectives can be diagnostic, evaluative, or proactive. These three objectives will determine the equipment choice, method of collecting data, disturbance thresholds, data analysis requirements, and the overall effort required.
Diagnostic program. A diagnostic program monitors to Characterize power quality problems that are affecting a facility or that may affect a planned facility. The following steps are used in the characterization process.
* Characterize the problem. Determine which equipment is not operating properly or failing.
* Correlate the problem with any changes in the system. Match equipment logs with power disturbances. Track down disruptive motor starts, capacitor switching, operation of nonlinear loads, power factor correction, a new industrial neighbor, utility operations, etc.
* Characterize the disturbances. Determine the magnitudes, durations, and frequencies of the disturbances causing the problem.
* Identify possible solutions.
* Install a solution and verify that it works.
Evaluative program. An evaluative program monitors to identify parameters that will improve power quality. The following steps are used in this parameter identification process.
* As in the diagnostic program, monitor as close as possible to the affected equipment at first.
* Monitor for both power disturbances and harmonics. Keep IEEE 519-1992, Recommended Practices for Harmonic Control in Electrical Power Systems, in mind. If you work with utilities, be watchful of series resonance problems and know that transient overvoltages are the most common cause of equipment problems. At plants and facilities, wiring and grounding are the most common source of power quality problems.
Proactive program. A proactive program monitors to identify conditions that could affect power quality and actions that will help avoid the problems. The proactive program distinguishes itself by its treatment of all data over the long term, displaying, analyzing, and archiving the data for trend analysis.
Choosing the monitoring system
What's the current trend? It's toward a distributed, computer-controlled approach because it's convenient and supplies information quickly. The basic configuration includes data-gathering monitors that are usually installed in strategic locations, such as the service entrance, critical-load branch circuits, substations, and other areas. These monitors collect data and then send it back to a central computer on a regular basis. The computer (desktop or laptop PC) controls the monitors, displays the data, exports it to spreadsheets for analysis or to word-processing programs for report writing, builds trends, and stores the data or passes it along for long-term storage.
What should you look for in a monitoring system? There are some general attributes that should be integral to the system. For example, the system as a whole should be very easy, even fun, to use. It should be fast and should efficiently collect and analyze data. (An automatic download capability is very helpful here.) It should collect all harmful power disturbances. This last trait is of particular concern.
There also are some very specific performance characteristics that are important. One is subcycle disturbance detection. If the proposed system isn't capable of detecting subcycle disturbances, it can't supply the information you need to protect critical loads. (See sidebar story "A Sampling Rate Primer.") If the system throws away all but the most damaging disturbances, it can't provide a true picture of your power, build a valid trend, or display the smallest disturbance that affects a sensitive load.
Another is the system's capability of providing data analysis. This is usually accomplished by exporting the data to a spreadsheet program.
Even if you're not interested in a proactive approach, the new power monitoring systems available today save time and personnel costs while still providing you with an easy tool to develop a more aggressive program, should that become necessary.
Data collecting factors
If you want to be proactive, or if you just warn to improve power quality, you have to start by finding out where you are right now. Then, you have to collect and keep your data to see if your situation is improving or degrading. There are certain factors that should be considered when collecting data for preventive maintenance. You should adapt the following plan and parameters for diagnostic and evaluative type programs also.
System plan. A system plan includes the number, placement, and type of power monitors and other equipment you'll deploy to collect data to fulfill your objectives. It also includes a plan for documenting the time, location, and nature of equipment disruptions. (Usually, all employees should be made aware that they can fill out a disruption report for the equipment they use. For some employees, this reporting will be mandatory.) Certainly, reports should cover critical loads, electrical equipment (overcurrent protective devices, capacitors, arrestors), and power protection devices.
Monitoring program itself. The monitoring program includes how much data you will collect (how thresholds will be set); how often you will collect this data; how and how often this data will be analyzed; how you will use the collected data (how often reports will be generated, what they include, and who receives them); and how and how often this data will be archived.
For the scope of EPRI's Distribution Power Quality (DPQ) Project (See sidebar story "Utility Deregulation and Power Quality" on page 76), steady-state trends such as watts, vars, and power factor were satisfactorily characterized after just a few weeks of periodic sampling. Based on this demonstrated data gathering performance, this procedure may also apply to various applications, such as industrial sites, depending on their power demand and consumption, their utility contract, and what they stand to gain by improving their energy rate or efficiency.
How important is downloaded data? "Important enough that every download should be saved," says Henry Diaz, a proactive power monitoring advocate and Manager of Power and Protection Engineering for MCI. Why? "Because you might be able to discern a trend only after a minimum of 90 to 180 days of data." "With two years of data, you can build really strong trends that point to specific areas of concern. The point is that you have to save every download; otherwise, you can't have a true picture of any decrease in power quality or power factor, or see the detrimental effects of harmonic distortion, or an increase in utility effects."
"If we see a degrading trend, we send an engineering team out to find out why." "Is it load- or utility-related ?" If it's utility-related, we confront the utility with the data. "Nine times out of ten, if it's a utility problem, it's because a new subscriber with big demand has been added to the grid, or the utility is overloading the feed. "Regardless of the situation, you'll have a brand new level of respect from the utility with which to increase your success rate, simply because you are armed with the trend data."
RELATED ARTICLE: A SAMPLING RATE PRIMER: RMS AND WAVESHAPE FAULTS
To provide an accurate rms measurement of the voltage and current waveshape, an instrument must sample the waveform at bast twice per cycle, assuming the cycle is periodic. To accurately measure the rms value of a waveshape with harmonic content, however, the sampling rate must be twice the frequency of each of the harmonic frequencies you want to measure. For example, to measure 3 kHz, you must sample 6000 times per second.
But accurate rms measurements are not enough. Why? Let's consider an extreme example to answer this. Suppose we consider a square wave and a sine wave. Now, it's possible to sample the square wave, with all of its harmonic content, often enough to reveal its true rms value. However, this information may be of little value because it's completely possible that the rms value of the sine wave can be identical to that of the square wave. This scenario points out the importance of sampling rate of waveshape faults. Without a technique for sampling waveshape faults, an instrument can not capture damaging subcycle distortion caused by power source switching, loose wiring, and utility recloser and power factor correction capacitor operation, for example. [ILLUSTRATION OMITTED]
RELATED ARTICLE: UTILITY DEREGULATION AND POWER QUALITY
One development that will greatly impact power quality in the near future is electric utility deregulation. This will certainly open up free market competition and will eliminate public utilities commissions (PUCs), those bodies that serve as utility watch dogs.
Suddenly, it will be every person (company) for themselves. Utilities will need to demonstrate and market their ability to supply quality power. Plants and facilities will need to verify utility-supplied power, and most will need to attend to their own internal needs.
How did the idea get started? In 1992, possibly to begin preparing for deregulation, the utility-member institution EPRI (Electric Power Research Institute) began a 3-year nationwide survey of utility power quality called the Distribution Power Quality (DPQ) Project. To analyze the project's large amount of data, a prime service contractor applied standard statistical techniques that, until that time, had not commonly been used on power quality data. (The techniques helped find correlations between some types of utility feeders and some types of disturbances, for example.)
One of the techniques involved the use of time trends. Long-term trends that came out of the DPQ project proved particularly useful. For DPQ, they supplied enough data points to make comparisons between different types of feeders valid. They helped show how distribution system changes affected power quality and if harmonics were increasing.
At nearly the same time, several large industrial customers with critical power needs began using similar techniques to avoid power disturbances and to monitor their utility-supplied power. If harmonics on a particular part of the distribution system were increasing, or disturbances were becoming more severe, an inspection could be called for and changes implemented before any real problems occurred.
Companies with large programs today have critical power needs. For them, any loss in productivity or down times is a large expense, so they can easily justify an all-out effort to avoid problems. However, new power monitoring systems are inexpensive, easy to use, and easily expanded as needs change. Anyone responsible for power quality should be prepared for an inevitable change in power service in the coming months and take steps now to avoid serious problems.
Susan Owen is Product Manager for BMI (Basic Measuring Instruments), Santa Clara, Calif.