Electrical Systems Must Stand on Their Own

Today's competitive market requires optimal decisions regarding facility capital and operating costs. Most manufacturing units are designed based on the economic criteria established for that particular product line. In some cases, the required payback period may be as short as one year.

By DAWN VAN DEE, Senior Electrical Engineer Burns McDonnell Engineering Co., Kansas City, Mo. March 1, 2001

Today’s competitive market requires optimal decisions regarding facility capital and operating costs. Most manufacturing units are designed based on the economic criteria established for that particular product line. In some cases, the required payback period may be as short as one year.

While this makes sense from one point of view, the question is: should decisions regarding a facility’s electrical-distribution system be subject to the same criteria?

Many times the facility’s primary electrical-distribution and emergency-backup systems are initially designed and built to support only those process units included in the original facility design. Over the life of the facility, however, product lines are changed, discontinued or added. As these changes occur, the electrical system is modified or expanded. It is common practice to include the capital cost associated with the electrical system improvements in the overall project estimates.

Why is this a problem? Occasionally, major improvements to the electrical system are required to support a relatively small increase in electrical load. Sometimes, larger projects-already in place-have used up valuable spare capacity in the system-without contributing funds to replace the capacity for future use. In this situation, it is tempting to reduce the capital costs associated with the electrical system-at the expense of reliability, expandability, flexibility and efficiency.

An alternative approach that many facilities are taking is to treat the electrical-distribution system as an infrastructure that is planned and expanded independently of individual projects. This allows the system to evolve over time, with planners using a sound and consistent approach considering technical engineering criteria and facility-wide economics.

No two electrical-distribution systems are exactly alike; requirements vary significantly. Individual facility requirements need to be analyzed and the system planned and designed accordingly. All facilities, however, should develop an approach to system planning that includes these basic considerations:

  • Safety.

  • Reliability.

  • Flexibility.

  • Maintainability.

  • Efficiency.

While each of these-especially safety-is important, this article focuses on the reliability aspects of facility electrical-system design.

Reliability = Profitability

Most decisions in facility operations are based on economic criteria, tied to market expectations for a given product. Market volatility causes decision-makers to conservatively assume short life expectancies for production units, and to require high returns and rapid payback on new projects. Evaluation of the electrical infrastructure is frequently rolled into the economic analysis of production units.

When electrical-system evaluation is embedded in the economic analysis of production systems, power reliability suffers. For example, new projects may be required to have less than a one-year payback period, or an internal rate of return (IRR) of greater than 100 percent. This encourages efforts to reduce initial capital requirements. The result is that the electrical budget often falls victim to cost-cutting, without an analysis of the probable impact that reducing reliability will have during the life of the facility. Budget-cutting measures that compromise reliability include:

  • Selecting less expensive but inferior electrical equipment components and construction materials.

  • Hiring less experienced installation contractors to save costs, reducing design engineering.

  • Limiting maintenance and eliminating redundancy.

Inferior products, poor workmanship, inadequate engineering and low-quality maintenance increase the occurrence of failures, and lack of redundancy increases the time required to restore power. The inability to avoid or respond efficiently to failures can be costly to both the manufacturing process and the bottom line.

Failure Has a Cost!

Costs associated with failures include: lost opportunity from lost production; electrical and process equipment damage; spoiled or off-specification product; penalty for environmental violations; and safety liability.

Reported costs of plant outages due to electrical power interruptions are provided in Table 1, based on Institute of Electrical and Electronics Engineers (IEEE) survey results. The average reported plant downtime resulting from a 1- to 10-cycle (16.7 to 167 milliseconds) electrical interruption is 1.39 hours, and the average downtime resulting from an interruption exceeding 10 cycles is 22.6 hours.

Plant downtime includes time required to restore electrical service and restart the plant. The calculated total cost of failure illustrates that a power interruption lasting fractions of a second can result in substantial economic losses. The large difference between the median and average costs indicates that some plants had exceptionally high outage costs.

Redundancy is Necessary

Electrical-system configuration affects the time required to restore power following a failure event. It also influences the duration of power interruption to the load for planned maintenance of electrical equipment. Configurations commonly used in plant electrical distribution systems include: radial, primary selective and secondary selective.

Figure 1 on page 11 is a typical radial system. Failure of any one component in the series path between the source and the load will result in power interruption to the load until the failed component can be repaired or replaced. Complete shutdown of the load is required for scheduled maintenance of any electrical equipment.

Figure 2 is an example of a primary-selective system. The load is normally fed by Source 1, with Source 2 available for emergency use. The failure of a component upstream from Point A results in power interruption to the load for the duration required to isolate the failed component and perform switching to feed the load from Source 2. Downstream of Point A, the system is arranged radially, and a failed component in the series path between Point A and the load requires repair or replacement to restore power to the load.

Figure 3 illustrates a secondary-selective system. Half the load is connected to Bus 1 and the other half to Bus 2. Normally, the system is operated with the tie-breaker open, and both sources are in service. However, the electrical equipment is rated to allow 100 percent of the load to be fed from a single source with the tie-breaker closed. The failure of a component upstream of Point B results in power interruption to half the load for the duration required to open Breaker 1, close the tie-breaker and feed the total load from Source 2.

Hot Transfer

Transfer schemes can be implemented to sense the interruption and perform switching automatically. An automatic-transfer scheme will not prevent motor starter contactors at Bus 1 (Figure 3) from dropping out, but remote-start capability can allow operators to quickly restart the motors from a control room.

A “hot” transfer scheme can be used thanks to the secondary selective system. It allows equipment to be isolated for maintenance without interrupting power to the load. For example, the total load shown on Figure 3 can be transferred to Source 1 by first closing the tie-breaker and then opening Breaker 2. De-energization of Source 2 then allows maintenance of equipment in the path between Source 2 and Breaker 2.

Closing Breaker 1, Breaker 2 and the tie-breaker simultaneously increases the available fault current at Buses 1 and 2. The equipment should be rated for the full available fault current. Thus, a hot transfer can be performed without interrupting power to the load.

Economic analysis may not justify full electrical redundancy, but a radial configuration may not provide the desired reliability. “Incomplete” electrical redundancy may be attractive, if the process can still operate profitably under reduced load conditions, or if a cutback in production in lieu of a sustained process outage will mitigate environmental or safety risks.

Lower-rated electrical equipment is a less expensive alternative. A secondary-selective electrical system designed to supply full load when both paths are available, but capable of supplying only 50 percent of the normal load from a single source, may cost only 30 percent more than a system designed to supply the full load from a single source.

Interruption’s Impacts

Understanding how planned and unplanned electrical outages will affect the facility is necessary to define reliability requirements. Total plant downtime is affected by the type of unplanned power interruption in terms of the voltage dip magnitude and duration. The survey presented in Table 1 on page 11 found that an interruption exceeding 10 cycles caused most plants to lose total motor load, while a 1- to 10-cycle interruption caused complete shutdown in about one-third of industrial plants.

Generally, a voltage dip below 29 percent of motor nameplate voltage at the terminals of NEMA design B motors results in motor stalling. Additionally, plant restart time can vary depending on the duration of the shutdown. Factors that contribute to the severity of a shutdown and increase average plant restart time include the time required to:

  • Remove defective products when outage duration results in product defects.

  • Reheat or pressurize the system prior to startup if outage duration allows the production system to cool or lose pressure.

  • Repair production equipment damaged by an uncontrolled shutdown.

  • Remediate for environmental release of hazardous or toxic material due to an uncontrolled shutdown.

Measures can be taken to minimize the effect of power interruptions. For example, capacitors are used to correct voltage dips on the utility supply. Momentary outages lasting a few cycles cause motors to drop out, but remote start capability allows an operator to respond quickly and restore operation in a few minutes. Fast bus transfer schemes applied on high-inertia loads such as large fans or centrifuge keep the load on line in the event of a power interruption. Secondary-selective electrical configurations allow the splitting of redundant process loads between buses so that an outage of one bus does not result in plant shutdown.

Where Maintenance Fits

Planned electrical outages are required to perform preventative maintenance of electrical equipment. Scheduled facility or production-unit shutdowns may provide the opportunity to perform electrical maintenance.

However, production shutdowns may occur infrequently or for insufficient duration to support adequate electrical maintenance. In addition, loads such as fire protection equipment or sump pumps that must remain in service during planned process shutdowns hinder de-energization of electrical equipment.

Redundancy provides the flexibility to de-energize portions of the electrical system to perform electrical maintenance with minimal load interruption. The value of redundancy-compared with shutting down the load-for electrical maintenance must be considered. If planned process outages are expected to support maintenance efforts, then redundancy for the purpose of maintenance alone is difficult to justify. However, if the alternative is to forego maintenance, the effect of poor maintenance on reliability should be considered.

Table 2 shows the effect of maintenance quality on failure rates, again from the IEEE survey. “Inadequate maintenance” was the reported cause of 16.4 percent of all failures. The percent of failures attributable to inadequate maintenance within the classification of the plant’s overall maintenance program is shown on the left side of the table.

For example, of the 1,469 total reported failures, 311 occurred in plants claiming to have excellent maintenance programs. Of the 311 failures, 36-or 11.6 percent-were attributable to inadequate maintenance. The percentages of failures attributable to inadequate maintenance within categories of “months since maintained” are shown on the right side of the table. For example, 310 failures occurred on equipment that had been maintained within the last 12 months. Of these 310 failures, 23-or 7.4 percent-were attributable to inadequate maintenance.

The percentage of failures increases with poor maintenance and with longer durations between maintenance. In addition to facilitating maintenance, redundancy improves the ability to respond effectively to failures. The degree of redundancy desired depends on the level of reliability the electrical system should provide. The following points will assist in selecting a system:

  • If the goals are to maximize availability and avoid shutting down all loads, the applicable configuration is secondary selective with an automatic transfer scheme.

  • If benefit is obtained from limiting the duration of downtime caused by the loss of the source, a primary selective configuration applies.

  • If the risk of failure is acceptable, a radial configuration is appropriate.

Justifying and Analyzing Reliability

Justification of reliability improvements should be based on economic, environmental or safety criteria. Economic considerations have been discussed above. Environmental criteria refer to the potential release of hazardous or toxic materials due to an uncontrolled facility or production-unit shutdown. Safety relates to potential hazards that exist when certain loads are lost such as fire protection equipment or ventilation equipment used to control the concentration of combustible gases, vapors or dust.

In evaluating reliability improvements, the following steps may prove of use:

  • Establish a “base case” electrical-system design that meets sound technical criteria.

  • Quantify the reliability of the base-case system.

  • Identify the reliability improvement.

  • Quantify the reliability of the modified system.

  • Determine the incremental change in reliability between the base-case and modified systems.

  • Determine the incremental investment and life-cycle costs of the improvement.

  • Determine the benefit derived from the change in reliability.

  • Evaluate the value of the improvement.

System reliability can be quantitatively described using analytical methods, such as cut-set methods. The cut-set method involves the following general steps:

  • Inventory the system components.

  • Assign reliability indexes to each component or minimal cut-set.

  • Identify the failure modes and minimal cut-sets.

  • Calculate the overall system reliability measures.

The inventory of system components consists of the equipment and connections shown on the one-line diagram. Reliability indexes are parameters that describe failure probabilities such as failure rate, time to repair the component or restore power by switching procedures, planned outage frequency for maintenance and outage duration for planned maintenance. Assignment of values for reliability indexes can be obtained from field surveys of actual equipment failures, manufacturer’s published data and theoretical parts-count method based on the military standard known as MIL-HDBK 217D.

A failure mode describes a failure that renders a component inoperable, including failure to function continuously and failure to respond acceptably. A cable fault is an example of a failure to function continuously, and the failure of a relay to clear the fault is a failure to respond. A minimal cut-set consists of the smallest combination of component failures that will result in interruption of power to the load. When components are connected in series, a failure of any one component will result in system failure. When components are connected in parallel, a simultaneous failure of components in each parallel path will result in system failure.

How To Calculate Electrical System Reliability

Basic calculations in Figures 4 and 5 are used in the cut-set method to determine overall system reliability for simple series- and parallel-connected components. The following variables are considered:

f = cut-set failure rate (per unit-year);

? = component failure rate (per unit-year);

r = average downtime per failure (“restorability,” in hours);

and, 1 / ? = mean time between failures (MTBF)

? r = hours of downtime per year.

Complex System Analysis

Among the methods applied to analyze complex systems are fault-tree and Monte Carlo analyses. Computer software packages are available that aid in these analyses. A fault-tree graphically models the combinations and relationships of minimal cut-sets using gates. The gates represent logic functions used in Boolean algebra such as AND, OR and NOT, and behave like switches in an electrical circuit, or valves in a piping system, to describe the flow of events that result in interruption of power to the load. The fault tree makes complex systems more manageable by graphically showing where to apply calculations to obtain overall system reliability. For example, an AND gate connecting two independent events means both events must occur to interrupt power to the load.

Monte Carlo analysis generates a system probability distribution. Individual probability distributions-defined by the mean standard deviation and distribution function-are first assigned to describe the failure probability of each component in the electrical system.

In step 2, a random failure rate is chosen for each component, based on the specified probability distribution. In step 3, the overall system failure rate is calculated from the randomly selected values for each system-component or device.

Steps 2 and 3 are repeated-say, for 1,000 times-until a probability distribution describing the overall system reliability is generated. Monte Carlo analysis lends itself to performing sensitivity analysis, which is instrumental in identifying key components or system blocks that significantly impact system reliability.

The electrical infrastructure will probably outlive many production-based systems-which should make the electrical system less sensitive to market volatility. Consequently, the electrical system should not be subjected to the same standards as production systems. Instead, an appropriate evaluation carefully considers the impact that electrical reliability will have during the facility’s life-with the expectation that the electrical infrastructure will survive many product cycles.

From Pure Power, Spring 2001