By Dominique Dieken, HSB Professional Loss Control
Power plant owners and operators are regularly bombarded with advertisements for a gamut of products and services that have one thing in common: a claim that their use will increase MW output or offer another form of cost savings. Fire protection systems, on the other hand, will not increase output by a single Watt and can even take a bite out of a power company’s profits if no fires occur. So why should anyone spend money on fire protection?
As an engineer with 13 years’ specialized experience in fire protection for electric generating plants, the author has surveyed hundreds of power plants. Amazingly, professionals at power plants still occasionally make statements like, “We don’t have much to burn,” implying that fire protection is not important at their plant. Just a few minutes spent thinking about possible fire hazards at any given power plant will quickly dispel this misconception. Virtually every type of power plant has thousands of gallons of lubricating oil, which is a Class IIIB combustible liquid in accordance with the fire codes. Many generators are cooled with hydrogen, a highly flammable gas with a very wide explosive range. In addition, every fossil fuel is by definition either combustible or flammable (coal, oil, natural gas). One is also likely to find electrical cables covered with combustible insulation, oil-filled transformers, combustible construction (wood cooling towers and PVC fill), and combustible loading within support buildings (warehouses, shops, offices).
Granted, power plants are not fire magnets in the league of apartment buildings and mobile home parks, which make the evening news on a regular basis. However, the relatively rare occurrence of a major fire at a power plant almost always leaves a significant impact on the plant, its employees and the bottom line. The insurance industry calls this the 80/20 rule: 80% of fires collectively cause 20% of the monetary losses and the remaining 20% of fires cause 80% of the monetary losses. In other words, it is a relatively small percentage of fires that have a detrimental effect.
Industry standards, such as NFPA 850, Recommended Practice for Fire Protection for Electric Generating Plants and High Voltage Direct Current Converter Stations, provide guidance on where and how to install fire protection systems, but they do not prioritize the importance of such installations. These publications are based on conventional wisdom and consensus standard-making, but they do little in terms of helping a designer or plant owner to decide whether the budget will be better spent on providing sprinkler protection for the steam turbine or for the cooling tower.
Fires have caused losses in excess of $30 million at power plants where fire protection systems were absent or failed. A fire loss of such magnitude, or even a few million dollars, is something a power company will remember and “feel” for many years to come. Fire is the largest cause of loss at fossil power plants, responsible for approximately 50% of the losses in the fossil fuel power plant category.1
One problem in both the power generation and the insurance industry is that fire loss information is not readily available. Power companies are reluctant to publicize such information due to the public image concern, and insurance companies will not share such information even after the claim is closed because they consider it proprietary. This leaves industry user groups, such as the Edison Electric Institute (EEI), to gather as much information as possible and share the information with the group. Such user groups also do not publicize their information openly but rather use the information to identify trends and develop guidelines. Other organizations, such as the Electric Power Research Institute (EPRI), publish studies on a limited basis. One of these studies, Turbine Generator Fire Protection by Sprinkler System,2 although published 18 years ago, is considered one of the most exhaustive and comprehensive studies in the industry and is also cited as a reference by the industry standard, NFPA 850. The bottom line is that if it can burn, it already has, more than once, and will burn again unless it is properly protected.
Some utilities and power companies base their fire protection priorities on one or more actual fires that occurred within the company. For example, one utility experienced a fire on a coal conveyor that caused several fatalities. As a result, the utility has provided automatic sprinkler protection on all coal conveyors throughout their plant fleet. While such decisions may seem reasonable, especially when the event is still fresh on everyone’s mind, they are knee-jerk reactions to one event rather than a scientific approach to risk evaluation. To risk analysis experts, likelihood is a key factor in assessing risk and risk reduction strategies, as likelihood is directly proportional to risk.
While consulting firms such as HSB Professional Loss Control perform risk-based analyses on a regular basis, a discussion of such an analysis is outside the scope of this article. Reference 3 provides detailed information on the methodology and steps involved in risk and performance-based fire protection evaluation. The benefit of risk-based analysis is that it provides a decision support tool based on fire scenario likelihood AND expected consequences (i.e. people exposure, equipment and structure damage, production down time, etc.). Fire prevention and fire protection alternatives can be assessed in terms of risk-reduction benefit versus cost, providing decision-makers an opportunity to prioritize fire safety improvements and optimize budget allocations. In essence, a risk-based analysis is the scientific study of any particular fire loss scenario and evaluation of cost-benefit when considering installing a fire protection system.
Fire likelihood data for electric generating plans is not widely published, but analysts who specialize in this field have their sources. As can be seen from Table 1, fires involving major power plant equipment are rare, but the likelihood that a power plant will experience a major fire within its life span is cumulative when adding equipment population. For example, a probability of one fire event in 2,222 unit years (or 0.00045 fire events/unit-year) for a catastrophic transformer fire may not sound like anything of great concern. However, when calculating the probability of a catastrophic fire involving one of many transformers over a 30-year plant life, the probability becomes significant.
For example, using Table 1 data, a plant with 16 transformers represents an annualized fire likelihood of 0.0072 fire events/year (0.00045 fire events/unit-year x 16 units). Over a 30-year period, this represents a cumulative likelihood of 0.216 fires/year (0.0072 x 30 years). Assuming a transformer fire aggregate consequence level of $2,500,000 (for unprotected transformers), the annualized risk potential is $18,000 (0.0072 fire events/year x $2.5 million). As a rough first-order cost/benefit estimate, if the cost of providing fire protection for these transformers is $250,000, then the payback period is approximately 14 years. Whether such a payback is acceptable will be at the discretion of risk and asset managers. These individuals should keep in mind that this timeframe is well within the employment and career duration of most managers.
In the modern world of availability clauses, penalties, and merchant plants, it is obvious that a plant incapacitated by a major fire will not only stop selling electric power, but will also lose significant amounts of money while the plant is in a forced outage for repairs. The unavailability of any prime mover such as a boiler or turbine-generator will shut at least one entire unit down. The unavailability of some auxiliary equipment, such as a cooling tower or step-up transformer, will have the same effect on operations as a prime mover.
Depending on the fire and the equipment that is involved, it could take anywhere from a few weeks to over a year to get back in business. For example, a major steam turbine-generator fire could take 6-9 months to repair, depending on availability of spare parts. A 200 MVA transformer has an estimated replacement time of 42 weeks according to HSB Professional Loss Control proprietary data – including lead time, shipping, disconnect & reconnect, and rigging – unless a suitable replacement transformer is readily available. Relatively minor cable fires have resulted in forced outages of a unit for an entire month due to the difficulty in routing and reconnecting the cables. The loss of a cooling tower could shut a unit down for 3-6 months while the old tower is demolished, repairs are made to the basin and the new tower is constructed.
While properly designed fire protection systems installed in accordance with nationally recognized standards will not reduce the likelihood of a fire nor completely eliminate any fire damage, they can significantly reduce the magnitude of damage when a fire occurs. For example, while a lube oil fire under the operating deck of a 200 MW steam turbine-generator without a properly designed fire protection system is expected to cause property damage in the $10-13 million range with a six-month downtime, the same fire with a properly designed fire protection system would not be expected to exceed $2 million in damage, with less than four weeks of downtime. While that still sounds like a considerable loss, it is much better than the first scenario.
As previously mentioned, the risk-based analysis takes into account the monetary issues of whether the expenditures for any given fire protection system are justified or whether they outweigh the risk. Obviously, a significant variable in the analysis is the life expectancy of the equipment or the plant. However, current industry trends indicate that there are few plans to shut down many plants in the near future. To the contrary, the trend is to rehab aging plants with current technology to avoid building new plants. While the cost-benefit ratio will vary from plant to plant, some key equipment fire protection systems could benefit most plants (based on risk analysis, loss experience and conventional wisdom):
- Steam turbine-generator bearing and below-deck sprinkler protection
- Coal hazards sprinkler protection (conveyors, transfer structures, tripper decks)
- Boiler burner front sprinkler protection (boilers using fuel oil for start-up or as a main fuel only)
- Cooling tower sprinkler protection (unless the tower is Factory Mutual approved as less combustible)
- Gas turbine fire protection by gaseous agent or preaction waterspray
- Outdoor oil-filled transformer deluge protection (for transformers over 300 MVA or when transformers have inadequate spatial separation distance)
- Electrical cable firestopping and waterspray/sprinkler protection (cables passing through building fire divisions, high cable concentration or where cables are exposed to an external fire)
As higher insurance self-retention limits, up to $10 million, are assumed by some owners and operators, interest in fire protection and loss control appears to be increasing in the industry. These higher “deductibles” mean that owners and operators are assuming more out-of-pocket risk while insurance underwriters may be less concerned about the level and quality of fire protection than before. This change in risk position is increasingly attracting management’s attention to fire protection.
While an investment in fire protection will not increase a plant’s megawatt output, it can save a plant and/or power company from financial hardship, negative media coverage, employee casualties and high insurance premiums. As with any capital investment, a return on the investment should be expected in the form of the above items. The question of “why should I spend money on fire protection” now becomes: “where is my fire protection money spent most efficiently?” Before committing a budget for any improvement, management needs to study all probable fire scenarios and start with the item that has the greatest risk reduction factor. Needless to say, the provision of fire protection systems should never be viewed as a substitute for proper equipment care, maintenance and housekeeping.
1. Marsh, Inc., A 30-Year Review of Large Losses in the Power Industry 1968-1998.
2. Turbine Generator Fire Protection by Sprinkler System, EPRI Report, 1985.
3. Barry, Thomas F., Risk-Informed, Performance-Based Industrial Fire Protection, TFBarry Publications and Tennessee Valley Publishing, Knoxville, Tenn., 2002. Publication is available at www.fireriskforum.com.
* Simplified Fire Hazard and Risk Calculations, Fire Protection Handbook, 18th Edition, National Fire Protection Association.
* Electric Generating Plants, Fire Protection Handbook, 18th Edition, National Fire Protection Association.
* A 30-Year Review of Large Losses in the Power Industry, Marsh, Inc.
Dominique Dieken, P.E., CFPS, is a senior engineer with HSB Professional Loss Control (www.hsbplc.com). Mr. Dieken has 13 years’ experience in industrial fire protection, specializing in electric power generation. He holds a B.S. from Cal Poly San Luis Obispo and is a registered fire protection engineer. He has served as a volunteer fire fighter in Greenwich, Ct.