Inferensys

Glossary

System Average Interruption Duration Index (SAIDI)

The System Average Interruption Duration Index (SAIDI) is a regulatory reliability metric that quantifies the total minutes of sustained power interruption the average utility customer experiences during a defined reporting period, typically one year.
Developer building agentic RAG system, retrieval pipeline diagram on laptop, technical workspace with notes.
RELIABILITY METRIC

What is System Average Interruption Duration Index (SAIDI)?

SAIDI is the primary regulatory metric for quantifying the average total duration of sustained power interruptions a customer experiences over a defined period, typically one year.

The System Average Interruption Duration Index (SAIDI) is calculated by dividing the sum of all customer interruption durations by the total number of customers served. It measures the total minutes or hours of outage time for the average customer, providing a direct indicator of grid reliability. Utilities and regulators use SAIDI to benchmark performance against historical data and jurisdictional targets, often excluding major event days to isolate routine operational effectiveness.

SAIDI is a critical input for Distribution Automation (DA) and Self-Healing Grid initiatives, where automated Fault Isolation and Service Restoration (SR) algorithms aim to minimize the duration component of this index. A lower SAIDI value signifies higher reliability, directly correlating with effective Outage Management System (OMS) integration and rapid Cold Load Pickup (CLPU) management strategies.

SAIDI EXPLAINED

Frequently Asked Questions

Clear, technical answers to the most common questions about the System Average Interruption Duration Index, the definitive metric for power distribution reliability.

The System Average Interruption Duration Index (SAIDI) is a reliability metric that measures the total duration of sustained power interruptions the average customer experiences over a one-year period. It is calculated by summing the total customer-minutes of interruption and dividing by the total number of customers served. The formula is: SAIDI = (Sum of all Customer Interruption Durations) / (Total Number of Customers Served). The result is expressed in minutes or hours per customer per year. A sustained interruption is typically defined as any outage lasting longer than 5 minutes, excluding momentary interruptions. This metric provides a normalized view of system performance, allowing for direct comparison between utilities of different sizes and across regulatory jurisdictions. It is a cornerstone of performance-based ratemaking and is reported annually to bodies like the IEEE and the North American Electric Reliability Corporation (NERC).

RELIABILITY METRICS

Key Characteristics of SAIDI

The System Average Interruption Duration Index (SAIDI) is the primary regulatory metric for quantifying the total time customers spend without power. It directly influences utility capital planning, grid modernization investments, and performance-based rate making.

01

Definition and Calculation

SAIDI measures the total duration of sustained interruptions the average customer experiences in a calendar year. It is calculated as the sum of all customer interruption durations divided by the total number of customers served.

  • Formula: SAIDI = Σ (Customer Interruption Durations) / Total Customers Served
  • Unit: Typically expressed in minutes per year or hours per year
  • Threshold: Only counts interruptions lasting longer than 5 minutes; momentary outages are excluded and tracked via MAIFI
> 5 min
Sustained Outage Threshold
02

Major Event Day Exclusion

Regulatory reporting often distinguishes between normal operations and catastrophic events. A Major Event Day (MED) is a day where daily SAIDI exceeds a statistically derived threshold, typically calculated using 2.5 standard deviations above the historical daily average.

  • Purpose: Prevents hurricanes, ice storms, and wildfires from skewing baseline reliability trends
  • IEEE 1366: The governing standard for MED identification methodology
  • Impact: Utilities report both total SAIDI and SAIDI excluding MEDs to provide a normalized view of day-to-day grid performance
03

SAIDI vs. SAIFI

While often reported together, SAIDI and SAIFI measure fundamentally different aspects of reliability. SAIFI (System Average Interruption Frequency Index) counts how often the average customer loses power, while SAIDI measures how long they were out.

  • SAIFI: Σ (Customers Interrupted) / Total Customers Served — unit is interruptions per year
  • SAIDI: Σ (Customer Minutes Interrupted) / Total Customers Served — unit is minutes per year
  • CAIDI: Customer Average Interruption Duration Index — SAIDI divided by SAIFI, giving the average restoration time per outage
  • Relationship: A utility can have excellent SAIFI but poor SAIDI if restoration times are slow
SAIFI
Frequency Metric
SAIDI
Duration Metric
04

Regulatory Benchmarks and Penalties

Public utility commissions use SAIDI as a key performance indicator in performance-based regulation frameworks. Utilities that exceed target thresholds may earn financial incentives, while persistent underperformance triggers penalties.

  • UK RIIO Model: Ofgem sets explicit SAIDI targets for network operators with revenue adjustments tied to performance
  • US State PUCs: States like California and New York incorporate SAIDI into rate cases and grid modernization approvals
  • IEEE 1366 Benchmarks: Provides percentile rankings (e.g., 25th, 50th, 75th) for peer utility comparison by region and density class
05

Topology Optimization Impact on SAIDI

Dynamic feeder reconfiguration and self-healing grid automation directly reduce SAIDI by minimizing the number of customers affected by a fault and accelerating service restoration.

  • Fault Isolation: Automated sectionalizing switches isolate the faulted segment in seconds, preventing upstream customers from experiencing sustained interruptions
  • Service Restoration: Tie switches transfer de-energized customers to healthy adjacent feeders, reducing customer minutes interrupted
  • CVR Integration: Conservation Voltage Reduction, when coordinated with topology changes, maintains voltage compliance during reconfiguration without extending outage duration
06

Data Sources and Measurement

Accurate SAIDI calculation depends on granular outage data from multiple operational technology systems. Modern utilities integrate these sources to automate reliability reporting.

  • OMS: Outage Management System logs customer outage start and restoration timestamps
  • AMI: Advanced Metering Infrastructure provides last-gasp and restoration pings for precise duration tracking
  • SCADA: Supervisory Control and Data Acquisition confirms device operations and fault clearing times
  • GIS: Geographic Information System maps customer counts to feeder segments for accurate normalization
RELIABILITY METRIC COMPARISON

SAIDI vs. Related Reliability Indices

A comparison of SAIDI with other standard IEEE 1366 reliability indices used to benchmark distribution system performance.

FeatureSAIDISAIFICAIDIMAIFI

Full Name

System Average Interruption Duration Index

System Average Interruption Frequency Index

Customer Average Interruption Duration Index

Momentary Average Interruption Frequency Index

Measures

Total outage duration per customer per year

Number of sustained interruptions per customer per year

Average time to restore service per interruption

Number of momentary interruptions per customer per year

Unit

Minutes per customer per year

Interruptions per customer per year

Minutes per interruption

Momentary interruptions per customer per year

Interruption Type

Sustained (>5 minutes)

Sustained (>5 minutes)

Sustained (>5 minutes)

Momentary (<5 minutes)

Major Event Days Included

Primary Use Case

Regulatory penalty assessment and long-term trend analysis

Infrastructure investment justification and reliability planning

Crew response efficiency and restoration process optimization

Power quality assessment and sensitive load impact analysis

Typical US Benchmark

120-150 minutes

1.0-1.5 interruptions

90-120 minutes

4-8 momentary interruptions

Calculation Denominator

Total number of customers served

Total number of customers served

Total number of interrupted customers

Total number of customers served

Prasad Kumkar

About the author

Prasad Kumkar

CEO & MD, Inference Systems

Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.

His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.