Uninterruptible Power Supplies: Distributed vs Centralized?

March 1, 1998
If you base your answer to this question on a thorough analysis, you could avoid a major blunder as well as unplanned equipment downtime.Which is a better strategy for your Uninterruptible Power Supply (UPS) system: centralized or distributed? To find out what's best for you and your application, you should do a thorough analysis. First, you have to know what information is important to that analysis,

If you base your answer to this question on a thorough analysis, you could avoid a major blunder as well as unplanned equipment downtime.

Which is a better strategy for your Uninterruptible Power Supply (UPS) system: centralized or distributed? To find out what's best for you and your application, you should do a thorough analysis. First, you have to know what information is important to that analysis, and how best to compare that information against both your needs and constraints. For example, in evaluating the purchase price, do you look at the long list of variables that constitute life cycle costs?

In comparing these two strategies, be careful not to oversimplify. To make an intelligent decision, you must understand exactly what you want your UPS system to do and how large that system is going to be.

To protect your equipment efficiently and completely, consider these issues: * Load requirements * Life cycle cost * Power conditioning and distance * System management * Fire protection * Maintenance * Service contracts and * Advantages/disadvantages of centralized/decentralized systems

How do you determine how much power your equipment requires? One common method is to base load requirements on individual equipment load ratings. A silver label next to the input cord of most computers shows how much current the equipment can draw. For various reasons (i.e., vendors oversize computer power supplies), this method leads you to oversize the UPS.

Specifiers often oversize UPS systems because it's difficult to measure proposed loads. Further, they seek additional run time and try to provide for future loads. These are not bad reasons for oversizing, but sometimes specifiers oversize to a detrimental extent. Consider that a UPS runs most efficiently at or near its rated capacity. So, as your load drops away from the rated capacity of the UPS, its efficiency also drops. The relationship is not linear; the drop is inconsequential when the load is near the rated capacity. However, if the load is far below the rated capacity, then so will be the UPS's efficiency. A large discrepancy can result in lifetime costs, in wasted electricity alone, that exceed the UPS's purchase price.

Another common method is to measure various current draws physically. You probably won't account for inrush current (two to three times the operating load), as it's not likely to exist when you take measurements. Switching loads on and off is generally not a good idea. This method doesn't allow for expansion or upgrades. It leads you to undersize the UPS.

There are two variations of undersizing. The first is where the UPS is too small for the load. This can be disastrous as it can delay bringing your systems on-line. This may lose your company business and cost you your job. Consult all parties who may possibly want UPS coverage, and get their written requests broken down into prioritization categories (essential, preferred, non-essential) or rated on a numerical scale of importance (1 through 10). A numerical scale allows you to weight the various loads in a spreadsheet and develop a menu of possible load ranges to accommodate-a budgeting advantage. Those who depend on you to specify the UPS may see their own needs as more critical than they actually are, but you must accommodate their perceptions.

The second variation of undersizing is where the UPS is too small for future expansion or upgrades. Don't size a centralized UPS exactly to fit your current load unless you're absolutely certain your load won't increase. Your best bet is to call a meeting with every affected party and explain the cost of adding capacity in the future as opposed to adding it now. Have a detailed account of that cost, including any structural (or other) upgrades, labor, and transaction costs you'll incur to install the added capacity. Once you have firm commitments of no load increases, add in a safety factor and consider the next larger UPS. This will help protect you from calculation errors and the ravages of time on the equipment under protection. This will also ensure that, as the batteries age, they will still provide at least the minimum required backup time. To determine how much of a safety factor to add in, you should contact vendors and discuss your particular needs and their equipment.

To determine life cycle costs, you should consider the following: * Initial purchase price (with desired options) * Shipping cost and * Service/maintenance costs

Be careful how you determine the initial purchase price. Some available options are unnecessary, but others will prevent a calamity that may drive your company out of business. Not all vendors offer the same features, so look in-depth at what's out there. Frequently, UPS manufacturers introduce new features that enhance the reliability, manageability, and scalability of their products.

Evaluate each feature on its own merits. If you find it worthwhile, add it to your list of desired features. Then, prioritize the items on your list. Keep the list accurate and informative to maximize its value. Once you've specified which features you need in your UPS, then compare other makes with those features. In other words, compare apples to apples.

Don't let shipping cost be a major embarrassment. Be sure to ask about delivery times. If you're ordering a custom-designed UPS, make sure to check into the company's record of on-time deliveries. Find out how much you'll pay per pound and total cost of shipping. Suspect any supplier that's much higher or lower than others.

Evaluate warranty programs carefully. Some manufacturers offer pro-rated and extended warranties for little or no additional cost. Some warranties contain exclusions and clauses that make them a poor deal for you. Some exclude batteries, and some even charge for some types of repair while the UPS is still in warranty.

Installation costs can become significant, if you construct special rooms and support systems. Installation costs are almost negligible for decentralized UPS systems. You should develop a detailed estimate of these costs before comparing overall costs. If you find the costs work against the philosophy you had in mind, try calculating costs for a hybrid system. You can install midrange UPSs that protect groups of equipment. So, instead of a truly centralized system serving an entire building or a separate UPS for each piece of equipment, you may be better off with a smaller UPS in an equipment room on each floor. This will cut power losses due to wire length from the UPS. You can take this even further and divide that floor into nodes, circuits, or other groupings to minimize the distance over which you must distribute your expensive conditioned power.

Service/maintenance costs are difficult to calculate and can be the most expensive. Consider the costs of the following. * Service contract (large UPS) * Spare unit(s) (distributed UPS) * Input and output power connectors * Oversizing for future expansion * Battery replacement * Installation, if part of service contract. (See previous discussion.)

And don't forget to check with the UPS vendors about battery exchange programs. You run a sealed battery for, say, three years. At the end of that three years, your vendor sends you a replacement package. You send your old battery back. They give you credit for your old battery and bill you at a reduced price for the new one. These arrangements are typically a win-win situation for all parties.

There are two main controllable variables when it comes to operation costs: * Personnel training/staffing and * Power distribution costs (efficiency losses)

Personnel training and staffing costs depend on the complexity of the UPS system and whether you have a service contract. If your system has a backup generator, chances are you already have a trained maintenance staff. A UPS battery room would be just one more item to add to its list of responsibilities and would not have a significant effect on labor costs. However, the staff would need specialized training, both for battery safety and the maintenancesteps involved. Poor battery maintenance practices can defeat the purpose of having a UPS. On the other hand, you may avoid specialized training (and/or some service contracts) if certain battery technologies will work for you. You may, for example, be able to extend battery replacement intervals to two decades: this would reduce maintenance to mostly routine inspections.

Power distribution costs are a function of your distribution voltages. As the voltage goes up, the current goes down for the same kW/kVA load. The lower the current, the smaller the voltage drop. Three-phase distribution is far more efficient than single phase. Distribute at the higher voltages and three-phase whenever possible. When you distribute conditioned power, your losses are more expensive than for ordinary power, so you further maximize efficiency as you minimize distribution of conditioned power.

Essentially, you can affect power quality only after the power enters your property. A centralized UPS has ramifications related to the distance between it and the protected equipment. A decentralized UPS has ramifications related to maintenance, control, and fire safety.

Power conditioning and centralized UPSs: As the distance between the UPS and protected equipment increases, the chances of a power problem on the AC line increases. Therefore, a UPS's effectiveness tends to decrease as the distance to the application increases. This isn't a problem, if you can ensure the environment through which the output AC must travel won't induce noise into the clean power. While not impossible, creation of such an environment is a formidable, and often expensive, task.

Power conditioning and decentralized UPSs: The distance between a decentralized UPS and the protected equipment is usually the same as the power cord's length (i.e., 6 ft). Consequently, the chances of power problems injecting into the AC lines between the UPS and application are minuscule.

Advances in network management systems now allow you to manage distributed power protection systems from a single management station. This station can be the same one that manages the other network components. This network control is important because hundreds of UPSs in a decentralized environment can be a logistical and managerial headache. If you can't control your UPS(s) from a central location, consider a hybrid system or centralized UPS. If the UPS itself prevents networking, consider replacing it with a different model.

For fire protection, the NEC (in Secs. 645-10 and -11) requires computer rooms to have special labeling and a means of disconnecting the UPS systems from their loads-from the principal exit doors.

What about distributed UPSs that aren't in a computer room? Sec. 645-11, Exception 2, specifically excludes UPSs rated under 750 VA. Most individual units for workstations are well under this limit. You'll find further information under Sec. 645-11, Exception 2 and under Art. 685, Integrated Electrical Systems.

Generally, distributed UPS systems do not require special attention for fire protection. It's a good design practice, however, to review your proposed system with the local authority having jurisdiction (AHJ) and with the technical liaison of your fire department.

Every UPS will eventually fail without maintenance, and some maintenance requires downtime. You must have a plan that will minimize the impact of UPS downtime on the users. Consider the following.

* Advance notification systems: Some UPS systems periodically test the battery system and warn the user if battery condition is poor or failure is imminent. They allow for scheduled maintenance, thus allowing users to work around system downtime.

* Off-line capability: Some UPS systems can go off-line for maintenance, repair, or replacement without interrupting the load. This capability is essential when there are no "after hours" situations.

* Service contract: Your system configuration may require you to engage a service contract. The service provider normally guarantees a maximum service response time. Determine what "response" means and exactly what the contract guarantees. For example, will the service technician simply bypass the UPS, or will he or she make every reasonable attempt to restore it to full operation as soon as possible? What is the ratio of trained service people to installed base? What factors affect you due to your geographical area or particular configuration? What factors could overload the service system just when you need it most; as in the case of very bad weather? Your contract should provide a financial penalty to the provider for performance failure.

Failure minimization: That's the cornerstone of the centralized/decentralized debate. In other words, you can either put all your eggs in one basket (centralized UPS) or you can diversify (decentralized UPS). Having all your eggs in one basket isn't necessarily a bad idea, as long as you watch that basket. You can have a backup UPS, so your total UPS installation 5 n32. The other option, diversification, means your total UPS installation 5 n12 (or 1, or whatever number you deem is sufficient backup).

Immediate business concerns may force you to run equipment whether your UPS is working or not. You can reduce or eliminate risk with a redundant UPS that backs up the first unit, or with a modular UPS that has user-replaceable modules (which you keep onsite). For distributed UPS systems, you can either size the units such that neighboring units have enough capacity to assume the load of the failed unit or keep spare units onsite. Spare units need not go to waste, but could be used on workstations until called into service, or simply rotated in and out of service on a monthly or quarterly basis.

So, what's your best choice? That depends. You have to weigh the advantages and disadvantages of each strategy. For centralized UPS systems, the main advantage of centralizing your UPS to a single location is that by doing so, you also isolate all your maintenance, repair, replacement, and security to that same single point. The main disadvantages to a single location are the distribution-related power-quality problems, the higher cost of operation, and (in most cases) installation.

For decentralizing your UPS across multiple locations, the main advantages are the lack of power distribution-related power quality problems and a cure to single-point failure. The main disadvantages are the added managerial complexity and a possible decrease in security.

You can centralize individual-equipment-sized UPS to one location (group small ones together), but you cannot decentralize a whole-building-sized UPS (break a big one into smaller pieces) across individual equipment. Whichever approach you choose, the capital involved means you must make your decision carefully after proper analysis.

About the Author

Shawn Smith

Voice your opinion!

To join the conversation, and become an exclusive member of EC&M, create an account today!

Sponsored Recommendations

Electrical Conduit Comparison Chart

CHAMPION FIBERGLASS electrical conduit is a lightweight, durable option that provides lasting savings when compared to other materials. Compare electrical conduit types including...

Fiberglass Electrical Conduit Chemical Resistance Chart

This information is provided solely as a guide since it is impossible to anticipate all individual site conditions. For specific applications which are not covered in this guide...

Considerations for Direct Burial Conduit

Installation type plays a key role in the type of conduit selected for electrical systems in industrial construction projects. Above ground, below ground, direct buried, encased...

How to Calculate Labor Costs

Most important to accurately estimating labor costs is knowing the approximate hours required for project completion. Learn how to calculate electrical labor cost.