Enhancing emergency power
Hospital emergency power systems must be reliable, scaleable, and cost-effective in order to work for extended periods.
In today’s climate, disaster preparedness has entered our social consciousness. Within the healthcare market, hospitals often are being called upon to provide emergency services through disaster situations. Meeting these demands to defend-in-place requires emergency power systems that are designed to support highly reliable operations for extended durations.
Traditionally, emergency systems have been designed based on code requirements defined in NEC 700 and NFPA 110. As the market has migrated to embrace a defend-in-place operational requirement, new codes have emerged to provide increased guidance for designing critical operation power systems (COPS). This new section of the National Electric Code (NEC 708) attempts to focus more effort on designing high-reliability solutions suitable for extended operation.
Urban hospitals and major medical centers rely on high-reliability designs using redundant generator configurations. These configurations rely on multiple large-capacity generators connected together through a switchgear configuration. Though highly reliable, this traditional implementation comes at a complexity and cost that limits its use in smaller hospital applications. These smaller applications often use a single emergency generator, or two isolated generators with a limited cross power capability.
Basic topology choices
When designing backup power systems, multiple power topologies need to be considered. The most basic is a simple single engine-generator configuration. In this configuration, the single engine meets the needs of the site with no redundancy. Though this configuration is conceptually the simplest, it may not be the preferred solution for hospital applications due to reliability limitations. A single engine-generator configuration has hundreds of potential failure modes. The end result, based on third-party studies, is that a single engine-generator typically has reliability of 98% to 99% given a 25-year lifecycle. This includes equipment, system, and human interaction failures.
2N: Redundant configuration
To improve reliability, redundant generator configurations may be implemented. Conceptually, the most basic redundant configuration is to use two large-capacity engine-generators with enough capacity that either could power the entire facility load. This configuration often is referenced as a 2N approach. This approach is used extensively in the data market because it provides a simple topology that matches the typical dual-cord power distribution seen in data applications.
The 2N approach is often used in a modified configuration to enhance reliability in smaller hospital applications. The most basic configuration is a two-generator configuration in which each generator has an allocated life safety, critical, essential equipment, and non-essential equipment circuit. In this configuration, the hospital can continue limited operation through a generator failure. The significant disadvantage of this approach is that during a generator failure, the hospital will experience outages on half of the emergency and essential circuits, resulting in significant impact to hospital operations.
To address the weakness found in completely isolated generator systems, it is fairly common to find provisions to cross-patch power between generator circuits. The simplest approach to this configuration is to use a split bus configuration with Kirk-key interlocks in a main tie configuration. In this configuration, each generator is capable of supplying power to any of the circuits. The disadvantage of this approach is that each generator needs to be sized to carry the full system load as well as extra capacity to support load growth. It is not uncommon for generators in this type of configuration to be loaded to only 20% to 25% capacity. The other option is to size the generator to carry only its own load and then manually shed some essential loads to power a circuit that was connected to the failed generator. These approaches also require on-site staff that is capable of significant human intervention.
The other option in high-reliability generator design is to use paralleled generation. By using multiple smaller generators connected in parallel, the hospital gets the same, or an even greater, level of redundancy offered by the 2N approach while conserving capital and supporting scalability. This type of approach often is referenced as an N+1 or N+2 based on the number of redundant generators running relative to the hospital load levels. Consider the following example:
Initial essential system (800 kW):
%%POINT%% Life safety load: 50 kW
%%POINT%% Critical load: 250 kW
%%POINT%% Essential equipment: 500 kW
Expected essential system load growth of 500 kW
Note: The final essential system size is 1300 kW
Solution using 2N approach:
Two 1500 kW units, total capacity 3000 kW
Initial design load level: 26.6% (note actual load may be significantly less)
Final design load level (after load growth): 43.3%
Estimated fully installed capital cost: $1 million
Solution using integrated paralleled generation:
Three 500 kW units initially, total capacity 1500 kW
Add fourth 500 kW unit at time of load growth
Estimated fully installed capital cost: initial $500,000 + future $150,000
Note: Paralleling switchgear cost included internally with the generators for integrated paralleling solutions.
N+1 implementation limitations
The concept of paralleling multiple gensets to produce greater amounts of power is not new. In fact, this arrangement is common in lager hospital applications. Parallel power solutions have always offered the standby generation marketplace significant advantages; however, the implementation of these solutions has been limited in smaller applications due to the following constraints:
%%POINT%% Issues of single source responsibility
%%POINT%% Issues of local switchgear expertise.
Historically, paralleled power generation was accomplished using third-party vendors that integrated UL891 dead front panelboards into generator paralleling switchgear. Though effective, this approach has its limitations. Cost is the most notable drawback. The installed and commissioned cost for low-voltage traditional generator paralleling switchgear typically is $50,000 to $70,000 per section. The switchgear also needs dedicated floor space inside the building—plan for each section to be 36 in. wide by 48 in. deep and 90 in. tall.
The most troubling issues with the traditional dead front panelboard approach to paralleling relative to acute care hospital applications are complexity and local expertise. Each generator in the system historically included four to six micro-controllers. These controllers were a combination of analog and digital technology from various manufacturers that were hard-wired together into an amalgamated system. Thus, a two-generator paralleling system would require nine to 14 controllers, once the master control section is included in the controller count. Another concern is the expertise required for commissioning and customizing the switchgear lineup. These issues create a significant reason for acute care hospital applications to restrain from embracing this technology.
Implementation and drawbacks
To access the benefits of parallel generation while removing the cost and complexity limitations, generator manufacturers have integrated generator paralleling into the genset package. This integrated approach to generator paralleling uses a single digital controller for each generator, connects the generator controllers together through simple two-wire digital communications, and integrates the power switching onto the generator. Integrated generator paralleling technology maintains all the benefits of traditional paralleling while removing its constraints.
Using integrated generator paralleling and combining smaller generator sets to power an acute care hospital is possible and highly desirable. This arrangement provides several advantages over large, single engine, and 2N configurations.
While the integrated generator paralleling approach has several benefits, the constraints of this approach should also be considered when designing, expanding, or replacing an existing emergency generator system. One of the constraint that must be considered is the location of the emergency distribution switchboard. Eliminating the central paralleling switchgear sections does not eliminate the requirement for the centralized distribution equipment. The intregrated approach still requires distribution equipment to house the necessary system feeder breakers and to provide a single point of generator connection.
According to the facility requirements, many hospitals require that the normal/emergency (N/E) power distribution feeders be protected by power draw-out circuit breakers so that any failed breaker may be quickly and easily replaced by the in-house maintenance staff. The option of waiting for an off-site electrician and having the N/E power on the distribution panel totally shut down while a failed breaker is replaced is not acceptable to many hospitals. The use of draw-out feeder breakers is one of the primary reasons that paralleling switchgear is so large and expensive. An integrated paralleling approach must also bear this expense while providing this functionality.
Another constraint to be considered is the level of customized load control desired. Traditional paralleling systems use custom PLC programming and full interaction with distribution breakers to enable sophisticated load control. As integrated paralleling approaches modularize the paralleling functionality from the distribution functions the load management tends to become more basic. An example of this is deferring from automatically drop generators off line at low load levels because of the lower level of facility load management integration. The integrated paralleling approach also tends to rely on permissive and load shed functionality that is implemented at the transfer switch level versus internal to the distribution equipment. The traditional paralleling approach, through custom PLC logic, has the internal capability of adding or dropping generators, adding or shedding load, and anticipating possible N/E load additions such as a fire pump or elevator. Customizing load control logic for each site allows for adequate generation capacity on-line while not running more generators than are required for the load.
When sizing generators in the healthcare industry, adequately planning for anticipated load growth is difficult, or the need for all the anticipated growth may not be immediate. If growth projections are too aggressive, precious project capital dollars are expended before the need is necessary. If growth projections are too low, the facility may be left without reliable standby power, implement unnecessary load shedding, or require expensive generator upgrades.
An integrated approach to paralleling generators would allow generators to be added as needed.
Reliability and redundancy
NEC Article 708 is driving a greater focus on reliability of healthcare services. Within the healthcare industry, especially at acute care hospital sites, backup power system design and failure mode operating philosophies are critical. One healthcare company has used reliability as a differentiator in the marketplace. With multiple generators installed in the N+1 configuration, redundancy is built in and reliability is increased because each generator backs up the other. The resulting gains in reliability for the critical loads are significant. For example, if a standby generator has a reliability of 98%, an N+1 configuration has a reliability of 99.96%, and an N+2 configuration has a reliability of five nines (99.999%).
Though catastrophic failures of standby generators are not common, multiple generator solutions significantly mitigate the effects of such an event. The inherent redundancy of the system ensures backup power even during equipment failure. When implementing integrated paralleling solutions with small generators, reliability is secured through implementation of:
%%POINT%% Proven high-volume engine-generators
%%POINT%% Power switching devices optimized for repetitive switching
%%POINT%% Integrated system components designed and manufactured to work together
%%POINT%% Integrated automatic backup systems
%%POINT%% Consistent hardware, firmware, and wiring
%%POINT%% Available coverage during maintenance.
Availability and serviceability
Anyone who has tried to purchase a large engine-generator in the past five years knows that delivery times can be as long as 18 months. Manufacturers generally build large engine-generators as they are ordered and ship these units to the site after manufacturing is complete. Smaller units (750 kW and below) are manufactured in high volumes with some inventory being established at the factory as well as the dealership. In the event a failure occurs and a unit needs to be replaced, the delivery of the new unit can take days rather than months. Furthermore, if the unit fails during a natural disaster, delivering a smaller unit to a site is more feasible when roads and highways may be difficult to navigate.
Rather than use one large industrial class engine, the integrated paralleling approach uses mass-produced engines. Consequently, maintenance and replacement parts are less expensive. Multiple generator solutions also provide flexibility during service operations. With multiple generators available, unit(s) can be taken out of service for repair or scheduled maintenance without complete loss of a site’s standby power. Remaining in-service units can still serve critical site loads should the need arise. The capital cost to replace a smaller unit is a fraction of the large, single generator expense.
Using multiple generators in the N+1 configuration instead of a large single-unit solution offers greater application flexibility. This fact can be a significant advantage in meeting many site-specific logistical constraints. Multiple smaller generators offer greater weight distribution, making rooftop installations more feasible.
Smaller generators are shorter and lower, providing flexibility in applications with height or depth constraints. Experience has shown that using this concept requires less space than the traditional 2N large engine-generator configuration, providing cost savings with building construction or additions. The generators do not need to be located side by side or even together, thus providing significant installation flexibility for retrofit projects.
Finally, the units do not have to be of the same kW rating supporting flexible growth plans. Engine fuel choice provides additional flexibility. An integrated paralleling solution may offer the application the choice for any combination of diesel, natural gas, or bi-fuel powered engines. Bi-fuel is an attractive option for acute care hospital applications. Bi-fuel units are compression-ignited diesel engines that can operate on 100% diesel or on a mix of 25% diesel and 75% natural gas. This two-fuel configuration extends the generator’s run time by a factor of four, minimizing the risk of running out of fuel. It also means the site may opt for less on-site diesel, reducing fuel maintenance costs.
Keeping cost in check is everyone’s objective, whether the cost is capital or expense. Some of the cost savings have already been discussed in the comparison with a typical 2N approach. Looking at the generator’s dollar cost per kW at different power outputs illustrates why using multiple smaller generators introduces cost efficiencies. The engines typically used in this concept are mass-produced, which provides economies of scale. They also integrate the paralleling function into the generator so no external paralleling equipment is required. Figure 1 shows relative dollar cost per kW for diesel engine generators.
Installation costs are generally comparable to a large, single engine-generator of equivalent output. When considering installation costs, the following points help explain why the costs are comparable.
%%POINT%% Smaller units are easier to transport.
%%POINT%% Smaller units are easier to handle at the site.
%%POINT%% Both are terminating the same number of amps.
%%POINT%% Cabling is easier to terminate (lower ampere density).
%%POINT%% Control wiring is simplified into digital communications.
%%POINT%% Units can be placed and commissioned in days versus weeks.
The end result is paralleled generation at the installed cost point of a single engine-generator. Some end users with load levels of 1.8 MW have seen as much as a $500,000 savings using an N+1 approach (4 x 600 kW for 2.4 MW) versus a 2N approach using two 1.8 MW generators.
Due to global warming concerns, many healthcare companies and their customers are seeking emission improvement and control. By being able to more accurately match the existing loads with the N+1 approach, efficiency is improved and emissions are reduced. Another benefit is an improvement in the time required to get state regulatory agency approval for the generator installations.
What does this really mean?
The first step in creating an integrated approach to generator paralleling is to remove the control complexity of the traditional switchgear approach. This is accomplished by using one digital controller per generator to control the following generator functions:
%%POINT%% Speed governing
%%POINT%% Voltage regulation
%%POINT%% Genset alarm and monitoring
%%POINT%% Load sharing
The approach of using a single digital controller per generator significantly enhances system performance. Inherently unstable load sharing methods are replaced with stable control loops. Synchronizing processes are greatly enhanced by directly interacting with frequency control functions. Troubleshooting becomes a simple process of monitoring inputs and outputs using a laptop computer. Repairs that took hours or days are reduced to minutes using an on-site spare controller and a simple plug-and-play approach. Supervisory control and monitoring also is made easy by simply passing all information digitally. If desired, the controllers can be accessed remotely by a service technician with the same functions available at the remote site.
The second step in creating an integrated approach to generator paralleling is to integrate the paralleling switch function into the generator connection box, thus removing the cost and space of external switchgear. Once a generator becomes synchronized, the generator controller issues a close command to a paralleling switch that connects the unit to the generator bus. Historically, this switch is a motor operated breaker located in a large metal cabinet, and connected to a bus bar.
With an integrated paralleling system, the paralleling switch is a high-cycle-rated contactor specifically designed for switching power circuits. The paralleling switch is mounted on and wired directly to the generator, resulting in a higher degree of system integration.
From an interconnection standpoint, the paralleling switch is cabled to a common point, typically a generator distribution panel. This cabling replaces the functionality of the generator bus bar inside traditional switchgear. From the generator distribution panel, various automatic transfer switches are fed. Figure 2 shows a sketch of how a system might be configured.
Traditional switchgear uses programmable logic control (PLC) to coordinate the operation of every generator and automatic transfer switch in the emergency power system. An integrated paralleling system also needs a system controller for coordination. These functions include starting and stopping the generators, prioritizing load shedding, and collecting data for supervisory control by building management systems. A significant difference offered by an integrated approach is that most all communication to the system controller is digital versus the hardwiring required in many traditional systems.
This consolidation of functions significantly changes the issues surrounding parallel generation. What was a complex system becomes a simple plug-and-play module. No more hardwiring multiple controllers together. No more difficult calibration processes. No more inherently unstable control loops. No more pulling I/O points back to the master PLC just to secure basic supervisor monitoring capabilities.
|Kirchner is industrial training manager with Generac Power Systems.|
Due to the critical temperature and humidity requirements of much of today’s highly sensitive diagnostic and treatment equipment in acute care hospitals, many owners and designers require at least a limited number of HVAC chillers and air movement equipment to be on emergency power when the outside air temperatures exceed 95 F.
To assist in determining this need, the Joint Commission requires that all healthcare facilities cond uct a vulnerability analysis to determine their degree of risk and their inability to diagnose and care for patients in the event of long-term power outage. This generally applies to outages in excess of 1 hr.
This equipment is not currently classified as essential power and therefore is optional to be on emergency power. If the emergency power system is designed for the equipment and has the capacity to provide emergency power to the facility cooling system, it is allowed to be on a manual transfer.
With much of today’s HVAC cooling systems (chillers, pumps, air handlers, etc.) on variable frequency drives, the equipment can be brought online with very limited inrush current resulting in minimal system transients.
Instead of designing excess capacity or an additional generator for HVAC needs, the use of paralleling multiple gensets using excess power capacity in all generators allows for the entire investment of capital dollars to be spent on the emergency power system.
It is essential that in the event of a loss or failure of one or more generators in the paralleling emergency power system, the master controller for the system immediately notify the BAS to shut down the nonessential equipment so as to not overload the remaining generator.