SAIFI Explained: Measuring the Frequency of Electric Service Interruptions

SAIFI, or System Average Interruption Frequency Index, measures the average number of sustained electric service interruptions experienced by customers during a defined reporting period. Where SAIDI describes how long customers are without power, SAIFI describes how often it happens. Utilities use SAIFI because frequent interruptions disrupt customers and strain equipment and crews even when each outage is short.

SAIFI is an average across the customer population, built from outage records and customer counts. It depends on how interruptions are defined, which events are counted and how good the data is. This article covers the formula, a step-by-step calculation, the data and systems involved, interpretation, and limitations, and it explains how SAIFI differs from SAIDI and CAIDI. For broader context, see electric grid reliability and how utilities measure grid reliability. Exact reporting treatment varies by utility, jurisdiction and methodology.

What Is SAIFI?

SAIFI, or System Average Interruption Frequency Index, measures the average number of sustained electric service interruptions experienced by customers during a defined reporting period.

Key points:

• Full form: System Average Interruption Frequency Index.
• What it measures: how often the average customer experiences a sustained interruption.
• Why it is customer-oriented: it counts interruptions per customer, not per outage event or per device.
• “Average frequency”: a total count of customer interruptions spread over all customers served, so it is a population average.
• What counts as an interruption: a loss of supply to a customer that meets the reporting definition. Sustained interruptions last longer than a defined threshold, while shorter ones are momentary. The threshold differs between definitions and is often a few minutes.

SAIFI is commonly associated with distribution reliability, because most customer interruptions originate on distribution feeders and equipment.

SAIFI Formula

SAIFI = Total Number of Customer Interruptions ÷ Total Number of Customers Served

• Total number of customer interruptions is the sum, over all qualifying interruption events, of the number of customers affected by each event.
• Total number of customers served is the customer base for the area and period being reported, often an average over the period or a count at a stated date.
• Result: interruptions per customer, with no time unit, for the reporting period.

Exact calculation and reporting methodology can depend on utility and regulatory practice, including which interruptions qualify and how the customer base is defined.

How SAIFI Is Calculated

1. Identify qualifying interruptions. Select sustained interruptions that meet the reporting definition.
2. Determine affected customers. Establish how many customers lost supply in each event.
3. Count customer interruptions. Add up the affected customers across events.
4. Determine the customer base. Use the number of customers served for the same area and period.
5. Apply the SAIFI formula. Divide total customer interruptions by customers served.
6. Validate the interruption data. Check events, times and customer counts.
7. Report the resulting index with the period and methodology.

Three terms are easy to confuse, and the distinction is central:

• Number of outage events: how many separate outages occurred. One feeder fault is one event.
• Number of customers affected: how many customers lost supply in an event.
• Number of customer interruptions: the count of interruptions summed over customers. A customer interrupted three times in the period contributes three customer interruptions, and an event affecting 200 customers contributes 200.

SAIFI uses customer interruptions, not events. One large event can contribute more than many small ones.

SAIFI Calculation Example

Hypothetical example. These numbers are invented for illustration and are not real utility data.

A utility serves 1,000 customers. During the reporting period, Outage A affects 200 customers, Outage B affects 300 customers and Outage C affects 100 customers.

SAIFI = 600 ÷ 1,000 = 0.60 interruptions per customer

This means that, averaged over all customers, there were 0.6 sustained interruptions per customer in the period. It does not mean each customer had 0.6 of an interruption. Many customers may have had none and others one or more.

What Does a High SAIFI Mean?

A higher SAIFI generally indicates customers experienced interruptions more frequently during the period. Possible contributors include repeated outages, equipment failures, vegetation-related faults, weather events, protection operations, aging infrastructure, distribution equipment problems, maintenance-related interruptions and network configuration issues.

A high SAIFI does not identify the root cause by itself. It signals a need to look at which feeders, events and causes drive the count, and engineering investigation is required to find out why.

What Does a Low SAIFI Mean?

A lower SAIFI generally indicates fewer sustained interruptions per customer during the period, which can indicate better interruption-frequency performance. SAIFI alone does not tell the full story. A system could have few interruptions that are very long, which would show a low SAIFI but a high SAIDI. This is why SAIFI should be read with other reliability metrics.

SAIFI vs SAIDI vs CAIDI

The relationship, when all three are calculated on the same basis and methodology, is:

CAIDI = SAIDI ÷ SAIFI

In short, SAIFI is frequency, SAIDI is duration, and CAIDI is duration per interruption.

SAIFI vs Momentary Interruption Metrics

Reliability reporting distinguishes sustained interruptions from momentary ones. A momentary interruption is a brief loss of supply, often caused by a recloser clearing a temporary fault. The Momentary Average Interruption Frequency Index, MAIFI, tracks them. SAIFI generally covers sustained interruptions, so it is not a substitute for momentary-interruption metrics. A system can have a low SAIFI and still produce many momentary interruptions that disturb sensitive customers. The threshold between momentary and sustained depends on the definition in use, so definitions should be checked before comparing figures.

Repeat Interruptions and the Customer Experience

Hypothetical example with invented numbers.

SAIFI counts every interruption a customer experiences, so repeat interruptions add up. Suppose a utility serves 1,000 customers, and 50 of them are on a feeder section that suffers four sustained interruptions in the period. Those 50 customers contribute 50 × 4 = 200 customer interruptions on their own. If no other customer is interrupted, SAIFI is 200 ÷ 1,000 = 0.20 interruptions per customer.

That 0.20 describes the average and looks low. Yet the 50 customers on that section each experienced four interruptions, which is far above the average, while the other 950 experienced none. This is why system-wide SAIFI can sit comfortably within a target while a small group has a poor experience.

Some utilities therefore supplement system averages with measures that show how many customers experienced a given number of interruptions. One example in industry guides is the Customers Experiencing Multiple Interruptions (CEMI) family of indices, although whether and how it is used depends on the utility and its regulator. Feeder-level SAIFI and lists of worst-performing feeders serve a similar purpose. The idea is simple: an average tells you about the system, and a distribution tells you about customers.

Seasonal and Time-Based Patterns in SAIFI

SAIFI is usually reported for a month or a year, but looking at it over shorter intervals can reveal patterns. Interruption frequency may rise in certain seasons because of storms, heat, vegetation growth or animal activity, and it may fall again when conditions change. Comparing like with like, for example the same month across years, avoids mistaking seasonal change for improvement or decline.

Trends are also affected by one-off events. A single severe storm can add a large number of customer interruptions in a day, which is one reason utilities analyze normal-day performance separately from major events. A rolling view over several periods tends to give a steadier picture than a single reading, and combining it with weather and outage-cause data helps explain why it moved.

What Data Is Needed to Calculate SAIFI?

The main data requirements are:

• Customer records, including the customer count.
• Customer interruption records.
• Outage events with start and end information.
• Affected customer counts.
• Network and device information.
• Event timestamps and restoration information.
• Planned or unplanned classification where applicable.
• Major event information where applicable.
• Data validation and reconciliation records.

Accurate customer-to-network mapping is important. Utilities must know which customers sit downstream of which device, or they cannot say how many customers an event affected.

How Utilities Collect SAIFI Data

Several systems contribute pieces of outage information:

• OMS (Outage Management System): tracks outages, affected customers and restoration.
• SCADA: reports equipment status, protection operations and alarms.
• AMI and smart meters: provide customer-level outage and restoration signals.
• Distribution automation: reclosers and switches report operations and switching.
• Field devices and protection systems: indicate faults and operations.
• GIS and network models: link devices to customers.
• Customer information systems, where relevant: hold customer counts and attributes.

Utilities do not all use the same architecture. Some rely mostly on the OMS, while others combine several sources.

SAIFI Data Flow

1. Grid event

   A fault or switching action occurs.

2. Field device or protection system

   Operates and records the event.

3. SCADA, AMI and OMS

   Report and collect event information.

4. Event validation

   Confirms the event and its times.

5. Customer impact identification

   Determines affected customers.

6. Customer interruption records

   Creates the count per event.

7. Reliability database

   Stores validated records.

8. SAIFI calculation

   Computes the index.

9. Reliability reporting

   Publishes results.

10. Planning and improvement

    Uses results to guide work.

Role of OMS in SAIFI

An outage management system can identify outages, track affected customers, track restoration, maintain outage events, support customer interruption calculations and provide operational records for reliability analysis. Because SAIFI is built from these records, OMS data quality directly affects the result. Wrong customer counts or unclosed events carry straight into the index.

Role of AMI and Smart Meters in SAIFI

AMI can improve visibility into customer outage events, restoration events, meter communication, customer-level status and outage validation. It can complement the OMS and other systems, for example by confirming that supply returned to specific customers.

AMI does not automatically provide perfect outage information. A communication gap can look like an outage when power is on, and a meter may fail to report an outage during a network problem. Data quality and cross-checking with other systems still matter. For more on smart metering, see key components of a smart grid.

Role of SCADA and Distribution Automation

SCADA and distribution automation can provide equipment status, breaker and recloser operations, fault indications, switching information and restoration information. This operational data shows what happened on the network. To calculate a customer-oriented index such as SAIFI, it must be combined with customer information, because the metric counts customers, not devices. See power system automation for how these systems work.

SAIFI and Data Quality

Data quality shapes SAIFI. Problems that can distort it include:

• Missing outage events, which understate frequency.
• Duplicate events, which overstate it.
• Incorrect customer counts for an event.
• Incorrect timestamps, which can affect classification.
• Incorrect feeder or customer mapping, which misassigns customers.
• Restoration data problems.
• Communication failures that hide or fake events.
• Incorrect event classification, such as planned versus unplanned.
• Inconsistent data between systems.

An automated calculation can give a precise-looking figure from flawed inputs, so validation and reconciliation matter.

SAIFI and Planned Interruptions

Utilities may have specific methodologies for planned interruptions in reliability reporting. Treatment can depend on the reporting methodology, jurisdiction and applicable reliability rules. Some utilities include planned interruptions, some report them separately, and some exclude them. No claim that all planned interruptions are always included or always excluded would be accurate, so any SAIFI value should state what it counts.

SAIFI and Major Events

Major event days or severe events may be handled differently depending on the applicable methodology. A severe storm can cause a large number of customer interruptions in one day. Utilities may therefore analyze normal operating performance, severe weather performance and major-event performance separately, because they reveal different things. No regulatory thresholds are given here, since they vary by jurisdiction.

SAIFI by Feeder, Region or Customer Group

Utilities can analyze interruption frequency at the system level and also by region, substation, feeder, customer class and geographic area. A system-wide figure may look acceptable while a particular feeder has far more interruptions per customer. Segmentation helps identify areas needing engineering attention. Small groups can swing widely from one event, so results should be read with care.

SAIFI and Distribution Reliability

SAIFI is especially useful for understanding distribution interruption frequency. Factors that influence it include feeders and distribution equipment, protection coordination, vegetation, weather, aging assets, automation, maintenance and network configuration. For example, protection that trips a whole feeder for a fault on a small branch increases the number of customers interrupted per event. SAIFI should not be treated as a complete measure of transmission-system reliability, which considers other aspects such as availability and security. For background, see transmission vs distribution networks.

SAIFI and Grid Reliability Improvement

Utilities can use SAIFI trends to find improvement opportunities, such as:

• Maintenance and asset replacement.
• Vegetation management.
• Distribution automation and FLISR.
• Protection improvements, such as better sectionalizing.
• Feeder reconfiguration.
• Undergrounding where appropriate.
• Weather hardening.
• Grid modernization and better outage detection.

No single solution automatically reduces SAIFI. The effect depends on the causes of interruptions on the particular network.

SAIFI and Grid Resilience

Reliability focuses on expected service performance and the interruption experience under expected conditions. Resilience focuses more on preparing for, withstanding, adapting to and recovering from severe disruptive events. SAIFI contributes to reliability analysis, but it does not by itself measure grid resilience. A utility could have a low SAIFI in normal years and still be exposed to rare severe events. For the distinction, see grid reliability vs resilience.

SAIFI and AI / Analytics

Analytics can help reliability teams with detecting unusual interruption patterns, identifying repeat outage locations, feeder-level trend analysis, correlating outages with weather, asset-risk analysis, reliability forecasting, maintenance prioritization and identifying recurring causes. SAIFI is a reliability metric. AI and analytics are tools for analyzing the data behind it. AI is not required to calculate SAIFI, and its findings need engineering review. See what causes power grid failures for common outage causes.

SAIFI and Grid Modernization

Grid modernization technologies include smart meters, AMI, distribution automation, FLISR, sensors, digital substations, advanced distribution management, real-time visibility, data platforms and analytics. They can improve operational visibility and the ability to analyze reliability performance. Better outage detection can also change recorded figures, for example by revealing outages that previously went unreported, so trends should be interpreted with the data-collection change in mind. See grid modernization and real-time visibility.

Practical SAIFI Interpretation Examples

All three examples are hypothetical and use invented values.

Example 1: Low SAIFI With Long Interruptions

A system shows a SAIFI of 0.2 interruptions per customer, but a CAIDI of 240 minutes. SAIDI would then be 0.2 × 240 = 48 minutes per customer. Interruptions are rare but long. An engineer might investigate restoration times, fault location methods, crew response, access and switching options.

Example 2: High SAIFI With Short Interruptions

A system shows a SAIFI of 3.0 with a CAIDI of 10 minutes, giving a SAIDI of 30 minutes per customer. Interruptions are frequent but quickly cleared. An engineer might investigate causes of repeated faults, protection coordination and whether temporary faults are being cleared quickly.

Example 3: High SAIFI Concentrated on One Feeder

A system-wide SAIFI of 0.8 looks acceptable, but one feeder has a SAIFI of 3.5. An engineer might review that feeder’s event history, vegetation, equipment condition, protection settings and exposure to weather, and consider targeted maintenance or automation.

SAIFI Improvement Framework

1. Measure

   Calculate SAIFI consistently.

2. Validate

   Check events, counts and mapping.

3. Segment

   Break results down by feeder, area and group.

4. Identify repeat problems

   Find locations with recurring outages.

5. Investigate root causes

   Analyze causes with engineering review.

6. Prioritize assets and feeders

   Rank where work should go first.

7. Implement improvements

   Carry out maintenance, automation or other changes.

8. Monitor the SAIFI trend

   Check whether results change.

Common Mistakes When Interpreting SAIFI

• Treating SAIFI as outage duration. It counts interruptions, not time.
• Confusing SAIFI with SAIDI. One is frequency, the other duration.
• Ignoring data quality. Errors flow into the index.
• Comparing utilities without understanding methodology. Definitions differ.
• Ignoring customer counts. The denominator matters.
• Treating every outage event as one customer interruption. The count depends on customers affected.
• Ignoring feeder-level patterns. Averages hide them.
• Assuming a low SAIFI means perfect reliability. Duration and momentary issues still matter.
• Using SAIFI alone for resilience analysis. It does not measure resilience.

Limitations of SAIFI

• It does not measure interruption duration.
• It does not fully describe momentary interruptions.
• It does not explain outage causes by itself.
• It can hide differences between customer groups.
• It depends on data quality.
• Methodology can vary.
• A system-wide average can hide feeder-level problems.

Because of these limits, multiple reliability indicators should be used together.

SAIFI vs Other Reliability Indicators

SAIFI counts sustained interruptions per customer. SAIDI totals their duration per customer. CAIDI gives the average length of each. MAIFI counts brief interruptions. ASAI expresses the fraction of time service is available.

How Utilities Use SAIFI in Reliability Planning

Utilities use SAIFI for trend analysis, feeder prioritization, asset replacement, maintenance planning, automation planning, reliability investment, performance monitoring, regulatory reporting where applicable and reliability improvement programs. It is one input among several, usually combined with SAIDI, CAIDI, outage causes and asset condition. No financial figures or regulatory requirements are given here, since they vary.

SAIFI in Modern Utility Operations

SAIFI connects to the OMS, AMI, SCADA, DMS or ADMS, GIS, reliability databases, analytics platforms and asset management systems. Operational systems detect and manage outages, GIS links devices to customers, the reliability database holds validated records, and analytics and asset management systems turn the results into decisions. The information flow is therefore end to end, from field event to improvement program.

Current, Emerging and Future Approaches

Adoption varies by utility. Future items are possibilities, not guaranteed outcomes.

Frequently Asked Questions About SAIFI

Conclusion

SAIFI measures interruption frequency and is a customer-oriented reliability indicator. It should be interpreted alongside SAIDI, CAIDI and other reliability metrics, and it depends on accurate outage and customer data. Modern AMI, OMS, SCADA and analytics can improve reliability visibility, and SAIFI supports engineering and utility reliability decisions. It is not a complete measure of grid performance or resilience.