Improving service quality to rural areas is challenging due to the high cost of traditional improvements with only marginal benefits to customers. National Grid in New York state is examining creative solutions such as applying distributed generation (DG), power conditioning, demand side management (DSM) and load desensitization to improve service quality in the most cost-effective manner.

One example is a project examining a power-quality/reliability upgrade on a rural portion of the system serving Old Forge, New York, U.S. This is a remote community located in the southwest corner of Adirondack Park, a state forest preserve with a low population density. The power source for the substations serving tourist towns in the region is a radial 46-kV line that snakes along 73 miles (117 km) of state highway 28 from Boonville to Blue Mountain Lake (Fig. 1).

THE TRIGGER

The winter operation of a downhill ski area originally triggered National Grid's attention to the power conditions in Old Forge. The local electric distribution is a 4800-V system with considerable impedance. The ski area motors — 300-hp air compressors and 125-hp water pumps for snow making, and a 100-hp ski lift motor (Fig. 2) — cause flicker while starting. This resulted in customer complaints. Upon examination, National Grid found that the ski area was using full-voltage starters resulting in starting current of six times the load current. To limit the starting currents, the ski area was directed to install electronic soft starters for the largest motors.

While this fix helped considerably, flicker complaints persisted to a lesser degree. In pursuit of further improvements, National Grid decided to perform a study to address both voltage flicker and broader issues of power quality and reliability in the region. National Grid selected EPRI PEAC Corp. (now EPRI) and Nova Energy Specialists, LLC to perform a two-year comprehensive study of the power quality and reliability in the region and to evaluate potential solutions. Key study objectives were to identify new techniques and, while looking at conventional solutions, to explore creative approaches that could include distributed generation, power conditioning, DSM and special controls.

MONITORING PROGRAM

The project team installed monitors at three sites: one on the 46-kV line and two on the distribution system. Interruptions (momentary and sustained), voltage sags, voltage swells, harmonic levels, voltage regulation, flicker and power demand were monitored. After two years of monitoring, the major findings were high service interruption rates on the 46-kV line, higher-than-expected seasonal loads and voltage flicker on the distribution line.

• Interruptions

  Over the period of the study, an average of 6.5 interruptions occurred on the 46-kV line each year with an average duration of 4.3 hours. The total accumulated interruption time each year had an average value of 27.9 hours. The values for system average interruption frequency index (SAIFI), customer average interruption duration index (CAIDI) and system average interruption duration index (SAIDI) were within the limits of typical values for long-line, rural and radial-feed systems. However, National Grid was looking to improve those reliability values.

• Demand

  The measurements show high demand periods during the summer holiday weekends when the tourist population increases in the Adirondacks (Fig. 3). Highest loading occurred whenever hot weather coincided with a holiday weekend. This condition is evident a few times during 2002 and 2003 (see the red circles in Fig. 3). The worst case measured was the holiday weekend on July 4, 2002, where the 46-kV system was unable to maintain voltage — the substation bus voltage dropped slightly below the ANSI C84.1 Range (A) lower limit. These measurements show that the peak loading for a few days each year slightly exceeded or nearly exceeded the voltage regulating capacity of the 46-kV line. Without any other interventions, a system upgrade would be required to alleviate this voltage condition.

• Voltage flicker

  Because of ski area motor operations, voltage flicker was slightly over the IEEE 519-1992 borderline of irritation and/or National Grid's own flicker curve limit (Table 1). A 7% to 8% voltage change on the primary feeder occurred when the 300-hp motors started. The ski lift, which starts and stops the most frequently, was causing roughly a 4% to 6% voltage change. The ski area had soft starters set about as low as they could be (250%) without stalling the 300-hp air compressors. The ski lift soft starters still had room for downward adjustment (they were set at 500%) and could probably be set as low as 300%.

SOLUTIONS

National Grid determined that 2 MVA of added capacity on the 46-kV line would solve the high load voltage problems (assuming no major subsequent load growth). The capacity problem on the 46-kV line was found to be voltage-drop related and not a thermal issue. Thus, an additional regulator bank located at a strategic point, plus some power factor correction, seemed to be the best option with a relatively low cost (US$300,000 to $500,000). This approach is lower in cost than installing DG or a new parallel line (Table 2).

It is notable that DSM and energy efficiency, in theory, might be an even better solution to the capacity shortfall. An aggressive DSM program could shave peak loads by 2 MVA while also saving energy. However, problems persist with DSM in implementation and maintenance. DSM is a “dispersed” solution that involves hundreds of participating customers. The tariff and regulatory mechanisms for feeder-focused DSM-style capacity support are not yet well established at this time.

Several options to improve reliability were also studied. These ranged from relaying adjustments to fault reduction to constructing redundant (parallel) circuits. Adding a redundant 46-kV line could potentially solve up to 95% of the interruptions, but at a cost of $4 million to $12 million, depending on whether it would be overhead or underground. The most promising reliability options involved the installation of DG and the creation of intentional DG islands subsequent to faults and during restoration.

Potential DG sources included a pumped storage hydro-UPS drawing from an existing lake or classic standby diesel DG. The best option overall was a simple diesel standby DG located at the substation that could carry the Old Forge Substation and the downstream 46-kV load as an intentional DG island. Its cost is between $2 million to $5 million and is lower than a redundant 46-kV line. The DG relaying and protection would need to deal with many issues including parallel versus islanded operation, effect of downstream reclosers on the island, cold load and inrush pickup, and so forth. But even with these complications, it would be workable with proper design. Table 3 highlights some of the key reliability options studied.

The most promising flicker mitigation option studied is a static VAR compensator (SVC). The SVC is a high-speed switching capacitor bank that can adjust its VAR output dynamically to compensate for motor-starting voltage drop. Its installed cost for the ski area load is $150,000 to $200,000, and it can mitigate at least 75% of the voltage flicker. Some of the other voltage flicker mitigation options are shown in Table 4.

The team considered upgrading the distribution feeder and substation to 13.2 kV. This was estimated to cost $1.2 million and would mitigate only about 55% of the voltage flicker. Another option was a dedicated 4800-V feeder directly to the ski area. This would mitigate 63% of the voltage flicker at a cost of $200,000 to $300,000, still not as good as the SVC. DG was considered for flicker mitigation because DG running in parallel with the system lowers the effective impedance of the source. However, simulations showed the DG would need to be at least 1000 kVA to provide suitable mitigation and the cost was estimated at $500,000 to $1 million — much more expensive than the SVC option. Some refinement of the ski-lift starter control settings could mitigate some of the flicker, but only up to about 30%.

An important option not mentioned in Table 4, but which shows out-of-the-box thinking, is the use of compact fluorescent lighting to solve flicker issues. Replacement of incandescent light sources in the afflicted area of the feeder with fluorescent lighting sources can solve the flicker issue about as well as the dedicated feeder option (63%) while saving energy. Studies show that fluorescent bulbs have one-third the flicker sensitivity of incandescent bulbs for a given change in voltage. Theoretically, this lighting technology solution was the lowest cost of all the solutions on a first-cost basis, but implementing and maintaining it is complex, so a utility-side SVC unit remains the solution of choice.

LESSONS LEARNED

The exercise of comparing new techniques with more conventional T&D upgrades as a means to improve rural power quality and reliability was instructive. This analysis showed that for the region of the Adirondack Park studied, a diesel standby DG configured to “intentionally island” is an effective option for reliability enhancement. For capacity enhancement to support the system voltage, a more classical 46-kV voltage regulator upgrade is the best choice. For voltage flicker mitigation, a distribution feeder-based SVC is a better choice than either a conventional dedicated feeder or DG.

It was clear in this case that DSM and energy efficiency could be the best options for solving certain capacity constraint and voltage flicker issues from a “total society benefits” perspective. Despite its superiority, this option is not easily implemented because the logistics of applying DSM/efficiency with large numbers of customers, and having a tariff/regulatory mechanism in place to proceed with feeder-focused DSM programs, is not available today. An important observation of this study is that there can be great value for electric utilities to have a mechanism for implementing DSM/efficiency projects for focused T&D support. Regulatory commissions and electric utilities may need to consider these techniques.

To correct the issues at Old Forge, National Grid has already moved forward with several system improvements. These include upgrading controls on the 46-kV reclosers, adding a new 4800-V voltage regulator bank on the ski area feeder, planning a voltage regulator for 2007 installation on the 46-kV system near White Lake, and making setting adjustments on some of the load tap changer controls at substations along the 46-kV line to improve voltage regulation. Tree trimming also has been performed along the 46-kV right-of-way. All these changes improve voltage conditions and system reliability. Other investment decisions such as whether or not to use a large substation DG island, add an SVC for flicker reduction or use transmission upgrades to improve reliability will be considered in the future.

Clayton Burns earned his BSEE and MSEE degrees from Lehigh University. His professional experience includes industrial electrical systems, electric generation, electric utility research, and his current assignment as a principal engineer in the Meter Engineering department at National Grid in Syracuse, New York. Burns is a registered professional engineer in the state of New York.
Clayton.Burns@us.ngrid.com

Phil Barker is the founder of Nova Energy Specialists, LLC, a Schenectady, New York-based consulting firm providing analytical services related to power systems, distributed generation and related energy technologies. He earned his BSEE and MSEE degrees from Clarkson University in Potsdam, New York. He is the author of more than 30 technical papers and articles and is a senior member of the IEEE.
pbarker@novaenergyspecialists.com

Tom Short is a part of an EPRI office located in Saratoga County, New York. Short holds a MSEE degree from Montana State University. He is the former chair of the IEEE working group on the Lightning Performance of Distribution Lines and led the development of IEEE Std. 1410-1997. Short authored the Electric Power Distribution Handbook by CRC Press (2004).
TShort@epri.com

Table 1. Voltage dip measurements.

Motor Name	Motor Rating and Starter (All motors shown are induction, Code-G types)	Voltage Dip Measured * (Primary voltage)
Air compressors	300 hp Soft starter set at about 250% current maximum	7% to 8%
Upper mountain water pumps	125 hp 65% voltage autotransformer start	5% to 6%
Lower mountain water pumps	75 hp Full-voltage start (600% current)	3% to 4%
Main ski lift	100 hp Soft starter but with high setting of about 500%	4% to 6%
*Measured at primary distribution feeder just outside the ski center.

Table 2. Solving subtransmission capacity issues (best short-term option is checked on the right).

Approach	Cost	Technical Performance	Issues/Comments	Overall Winner
Enhanced reactive support of transmission system and/or added 46-kV voltage regulator	$300,000 to $500,000	Marginal (perhaps a 10% to 15% improvement in capacity is possible)	More measurements and system modeling needed to confirm viability of this approach.	ϫ (pending more analysis)
DSM/energy efficiency	$200,000 to $500,000	Good but difficult to control	Might be lower cost than all other choices, but requires extensive customer participation, which is difficult to implement/maintain.
Distributed generation	$500,000 to $1 million	Capacity released depends on DG size	DG does not need to be as large as is required for reliability solution.
Redundant transmission (a second parallel line)	$4 to $8 million	Essentially doubles capacity	Provides solid massive capacity increase. Also an excellent reliability solution.

Table 3. Comparison of 46-kV reliability enhancement options (best option is checked on the right).

Approach	Cost (US$)	Technical Performance	Issues/Comments	Overall Winner
Enhance existing 46-kV line reliability (lightning protection, better insulators and mechanical strength)	$50,000 to $1 million	Poor to fair (mitigates only 5% to 30% of interruptions)	Only eliminates a fraction of outages (probably not a sufficient improvement to address the issue.
Add a redundant 46-kV overhead line	$4 million to $8 million	Fair to good (mitigates 40% to 80% of interruptions)	A second redundant line on opposite side of highway. Suffers from poor aesthetics.	A close second place
Add an underground 46-kV cable	$7 million to $12 million	Excellent (mitigates less than 95% of interruptions)	An underground cable is very expensive due to rock trenching in Adirondacks; however, it has best aesthetics and reliability of all options.
DG option: 200-hours per year standby diesel DG	$2 million to $5 million	Fair to very good (mitigates 50% to 90%)	Effectiveness depends on size and type of DG island. Cost is very competitive. Peak shaves in summer. Minor fuel issues.	ϫ
DG option: Hydro-UPS	$4 million to $8 million	Fair to very good (mitigates 50% to 90%)	Great use of indigenous hydro-renewable resources; needs more study. May peak shave in summer.
DG option: Full-time large molten carbonate fuel cell	$10 million to $50 million (not factoring subsidy)	Fair to very good (mitigates 50% to 90%)	Excellent demo of large-scale fuel cell. Extremely expensive without subsidy. Can't follow load or operate in standby mode. Cogeneration possible. Fuel issues. Provides capacity support.

Table 4. Comparison of voltage flicker mitigation options (best option is checked on the right).

Approach	Cost	Technical Performance	Issues/Comments	Overall Winner
Dedicated 4.8-kV feeder	$200,000 to $300,000	Very good (mitigates 63% of ΔV)	A simple, straightforward approach.
Upgrade to 13.2 kV	$1 million to $2 million	Good (mitigates 55% of ΔV)	Replacing substation transformer and upgrading feeder equipment to 13.2 kV makes it expensive.
Distributed generation	$500,000 to $1 million	Fair to very good (effectiveness depends on DG size)	DG must be at least 1000-kVA to be as effective as feeder upgrades.
Static VAR compensator	$150,000 to $200,000	Excellent (less than 75% ΔV mitigation)	As an added advantage, SVC can provide summertime reactive support.	ϫ
Additional starter adjustments	less than $10,000	Poor to fair (10% to 50% ΔV mitigation)	May only provide marginal improvement for compressors.