SARFI: System Average RMS Variation Frequency Index
What is SARFI?
SARFI is an acronym for System Average RMS Variation Frequency Index. It is a power quality index that provides a count or rate of voltage sags, swells, and/or interruptions for a system. The size of the system is scalable: it can be defined as a single monitoring location, a single customer service, a feeder, a substation, groups of substations, or for an entire power delivery system. There are two types of SARFI indices. SARFIX and SARFICURVE.
SARFIX
SARFIX corresponds to a count or rate of voltage sags, swell and/or interruptions below a voltage threshold. For example, SARFI90 considers voltage sags and interruptions that are below 0.90 per unit, or 90% of a system base voltage. SARFI70 considers voltage sags and interruptions that are below 0.70 per unit, or 70% of a system base voltage. And SARFI110 considers voltage swells that are above 1.1 per unit, or 110% of a system base voltage. The SARFIX indices are meant to assess short-duration rms variation events only, meaning that only those events with durations less than 60 seconds are included in its computation.
SARFICURVE
SARFICURVE corresponds to a rate of voltage sags below an equipment compatibility curve. For example SARFICBEMA considers voltage sags and interruptions that are below the lower CBEMA curve. SARFIITIC considers voltage sags and interruptions that are below the lower ITIC curve. Lastly, SARFISEMI considers voltage sags and interruptions that are below the lower SEMI curve. These curves do not limit the duration of an rms variation event to 60 seconds; therefore, the SARFICBEMA, SARFIITIC, and SARFISEMI are valid for events with durations greater than ½ cycle.
Example Calculations
Consider the following rms variation event summary table, which was hypothetically measured at a single site:
Table 1: List of RMS Variation Events Measured at a Single Monitoring Site
| Time Stamp | Minimum Voltage | Event Duration |
|---|
| Jul-01-1997 09:48:52 | 73% | 9 cycles |
| Jul-01-1997 09:50:16 | 73% | 9 cycles |
| Jul-07-1997 14:20:12 | 0% | 82 cycles |
| Jul-10-1997 15:55:23 | 13% | 100 cycles |
| Jul-21-1997 09:48:52 | 0% | 2.6 seconds |
| Aug-08-1997 07:35:02 | 49% | 34 cycles |
| Sep-02-1997 08:30:28 | 0% | 41 seconds |
| Sep-08-1997 10:30:40 | 59% | 40 cyc |
The count of voltage sags and interruptions that would be included in the SARFI90 is 8, as there were 8 voltage sags and interruptions measured at this location that were had a minimum voltage below 0.9 per unit (90 percent) and between ½ cycle and 60 seconds in duration. This can be expressed as a rate of 3.93 events per 30 days. This is computed by dividing the 8 events by the 92 days between July-01-1997 and Oct-01-1997, and then multiplying by 30 to normalized to events per 30 days.
Table 2: SARFIX Rates Computed from Table 1
| Index | Count | Rate per 30 Days |
|---|
| SARFI90 | 8 | 2.61 |
| SARFI70 | 6 | 1.96 |
| SARFI50 | 5 | 1.63 |
| SARFI10 | 3 | 0.98 |
Technical Paper
The following paper introduced the SARFIX index and provides an indepth technical explanation of its use.
Reference
D. L. Brooks, R. C. Dugan, D. D. Sabin, S. Williams, Reliability Benchmarking Methodology. EPRI Report TR-107938, Palo Alto, California, May 1997.
D. L. Brooks, R. C. Dugan, M. Waclawiak, and A. Sundaram, "Indices for Assessing Utility Distribution System RMS Variation Performance," IEEE Transactions on Power Delivery, vol. 13, no. 1, January 1998, pp. 254-259.
Abstract
For many years, electricity distribution companies have used sustained interruption indices as indicators of the reliability of service provided on their systems. Today, however, many electricity consumers are adversely affected by more subtle voltage disturbances such as sags and swells. Many utilities are well aware of such service quality concerns and are implementing extensive monitoring systems to detect such disturbances and assess service quality in this regard. This paper presents a subset of work completed which provides utilities with tools to make more complete service quality assessments. Indices developed to reflect system service quality with respect to all rms variations are presented. Example values for the indices are calculated using data from a national distribution power quality data collection project. Finally, an example application of the indices currently being made by a distribution utility is discussed.
Download Paper
Download paper as Adobe Acrobat file (90KB)
This page was last updated on April 19, 2000