Glossary

A

Anomaly Detection
A machine learning technique that identifies unusual patterns or outliers in data that deviate from normal behavior. In utilities, used to detect equipment malfunctions, cybersecurity threats, or operational irregularities.

ARIMA (AutoRegressive Integrated Moving Average)
A time series forecasting model that captures autoregressive patterns, trends (through differencing), and moving average components. Commonly used for load forecasting.

Asset Management
The systematic process of managing physical infrastructure assets (transformers, breakers, lines) throughout their lifecycle, from installation to replacement.

C

Classification
A machine learning task that predicts discrete categories or labels (e.g., “healthy” vs. “failure-prone” for equipment).

Clustering
An unsupervised machine learning technique that groups similar observations together without predefined labels. Used for customer segmentation and asset grouping.

Computer Vision
A field of artificial intelligence that enables machines to interpret and understand visual information from images or video. Used in utilities for automated infrastructure inspections.

D

Data Drift
Changes in the distribution of input data over time that can degrade model performance. Requires monitoring and periodic retraining.

Data Lake
A centralized repository that stores raw data in its native format. Enables analytics on diverse data types without predefined schemas.

Demand Response
Programs that encourage customers to reduce or shift electricity consumption during peak periods, helping utilities manage demand without building new generation capacity.

DER (Distributed Energy Resources)
Small-scale power generation or storage systems located near where energy is consumed, such as rooftop solar panels, battery storage, or small wind turbines.

Distribution Grid
The portion of the electric grid that delivers power from substations to end customers (homes, businesses).

E

EAM (Enterprise Asset Management)
Software systems that track and manage physical assets throughout their lifecycle, including maintenance records, inspections, and work orders.

Ensemble Methods
Machine learning techniques that combine predictions from multiple models to improve accuracy and robustness.

ETL (Extract, Transform, Load)
The process of extracting data from source systems, transforming it into a usable format, and loading it into a target system for analysis.

F

Feature Engineering
The process of creating new input variables (features) from raw data to improve model performance. Examples include creating time-based features (hour of day, day of week) or interaction terms.

Forecast Horizon
The time period into the future that a forecast predicts (e.g., 1 hour ahead, 24 hours ahead, 1 week ahead).

G

GIS (Geographic Information System)
Systems that capture, store, and analyze geographic and spatial data. Used in utilities to map infrastructure locations and analyze geospatial relationships.

Grid
The interconnected network of power generation, transmission, and distribution systems that deliver electricity from producers to consumers.

L

Load Forecasting
The process of predicting future electricity demand. Critical for generation scheduling, market participation, and grid operations.

LSTM (Long Short-Term Memory)
A type of recurrent neural network designed to capture long-term dependencies in time series data. Used for load and renewable generation forecasting.

M

MAPE (Mean Absolute Percentage Error)
A forecast accuracy metric that expresses error as a percentage of actual values. Commonly used in load forecasting.

MLOps (Machine Learning Operations)
Practices and tools for deploying, monitoring, and maintaining machine learning models in production environments.

Model Drift
Degradation in model performance over time due to changes in data distribution or underlying relationships. Requires monitoring and retraining.

N

Net Load
The difference between total electricity demand and behind-the-meter generation (e.g., rooftop solar). What the grid must actually supply.

NLP (Natural Language Processing)
A branch of artificial intelligence that enables computers to understand, interpret, and generate human language. Used in utilities to analyze maintenance logs and compliance documents.

O

Orchestration
The automated coordination of multiple tasks or workflows, managing dependencies, scheduling, and error handling.

Outage
An interruption in electric service to customers, caused by equipment failures, weather, or other factors.

Overfitting
When a machine learning model learns patterns specific to training data that don’t generalize to new data, resulting in poor performance on unseen examples.

P

Predictive Maintenance
A maintenance strategy that uses data and analytics to predict when equipment failures are likely to occur, enabling proactive intervention before failures happen.

Precision
A classification metric measuring the proportion of positive predictions that are actually correct. Important when false alarms are costly.

PMU (Phasor Measurement Unit)
High-speed sensors that measure voltage and current phasors across the grid, providing real-time visibility into grid conditions.

R

Recall
A classification metric measuring the proportion of actual positives that are correctly identified. Important when missing failures is costly.

Regression
A machine learning task that predicts continuous numerical values (e.g., load in megawatts, temperature in degrees).

Reinforcement Learning
A machine learning paradigm where an agent learns to make decisions by interacting with an environment and receiving rewards or penalties.

Reliability Metrics
Measures of grid performance, such as SAIDI (System Average Interruption Duration Index) and SAIFI (System Average Interruption Frequency Index).

S

SCADA (Supervisory Control and Data Acquisition)
Systems that monitor and control grid equipment in real time, collecting telemetry data and enabling remote control of switches, breakers, and other devices.

SARIMA (Seasonal ARIMA)
An extension of ARIMA that explicitly models seasonal patterns, essential for load forecasting which has strong daily and weekly cycles.

SHAP (SHapley Additive exPlanations)
A method for explaining individual model predictions by quantifying each feature’s contribution to the prediction.

Supervised Learning
Machine learning where models learn from labeled training data (inputs paired with known outputs).

T

Time Series
Data collected over time at regular intervals (e.g., hourly load, daily temperature). Requires specialized modeling techniques to capture temporal patterns.

Transformer
Electrical equipment that changes voltage levels in the power grid. Also refers to a type of neural network architecture used in natural language processing.

Transmission Grid
The high-voltage portion of the electric grid that moves power over long distances from generation plants to distribution substations.

U

Unsupervised Learning
Machine learning where models find patterns in data without labeled examples. Includes clustering and anomaly detection.

V

Voltage Control
The process of maintaining voltage levels within acceptable limits across the grid, typically through reactive power management.

W

Work Order
A formal request for maintenance, repair, or inspection of equipment, typically tracked in EAM systems.