Ongoing Research · Grid Intelligence

Predicting Grid Failure
Before It Happens.

A software-only decision-support platform that watches grid conditions and internal plant parameters, forecasts disturbances, and recommends pre-positioning actions. Built for Nigeria, where 222 grid collapses have occurred since 2010.

Scroll
Back to Research
Project Overview

The Problem We Saw in Nigeria's Power Sector

Nigeria's national grid has collapsed 222 times between 2010 and 2022. On 65.3 percent of days analyzed, frequency exceeded alert thresholds. The grid operates outside safe parameters most days, and power plants have no advance warning of incoming disturbances.

Plant Availability Factor sits at approximately 39 percent. When the grid collapses, generation drops to zero in seconds. Restoration takes hours. Every collapse event costs millions in damaged equipment, lost generation, and emergency repairs. Existing monitoring systems serve the transmission system operator, not the individual power plant.

We are building a system that gives power plants advance warning of grid disturbances so operators can pre-position before collapse.

What We Are Building

Four monitoring layers, each tracking a different signal, all feeding a single alert cascade for operators.

Grid Frequency Predictor

Rolling statistical predictor with adaptive lookback windows. Provides 10-minute lead time on frequency deviations. Validated against historical NERC data with strong predictive accuracy.

Market Stress Index

Composite score combining frequency deviation, generation-demand gap, capacity shortfall, and voltage deviation. Day-ahead horizon with 0.803 correlation to critical days.

Thermal Parameter Monitor

Rolling regression on plant cooling system parameters. Predicts thermal breaches 72 hours ahead with 70.5 percent detection rate and 1.9 percent false alarm rate.

Gas Supply Monitor

Tracks gas pressure trends from SCADA data. Provides 30 to 120 minute lead time on gas supply disruptions, which are a primary cause of forced outages in gas-dominated grids.

Why Nigeria First

Africa's largest power market, structurally underserved by digital solutions, with a grid in crisis.

$30B
Power Market Value
2026, growing to $70B by 2035
222
Grid Collapses
2010 to 2022, NERC data
65%
Days Outside Safe Band
Of 731 days analyzed
39%
Plant Availability
Q1 2026, 61% unavailable

28 grid-connected plants, no plant-centric monitoring

Existing enterprise solutions from GE, Siemens, and ABB target transmission system operators and OEM-controlled plants. No competitor serves the individual generator facing an unstable grid. This is a structural gap.

Regulatory reform creating opportunity

The Electricity Act 2023, NISO inauguration, and Grid Code v3 draft are modernizing the sector. Operators are actively seeking digital grid solutions. The window for plant-centric intelligence is open now.

Thermal dominance aligns with the system

Natural gas supplies over 90 percent of thermal generation, which is 77.5 percent of installed capacity. The thermal parameter monitoring layer directly addresses the dominant generation type in Nigeria.

Software-only, no hardware required

The system connects via existing OPC UA or Modbus interfaces. No plant shutdown for installation. No sensor procurement. Operational on day one with no training data required from the plant.

How It Works

The system watches external grid conditions and internal plant parameters simultaneously. When any layer predicts a disturbance, it triggers an alert cascade with specific operator actions.

1
Normal. All layers within stable thresholds. Standard operations continue.
2
Watch. Any layer shows trending deviation. Increase monitoring frequency, review auxiliary systems.
3
Alert. Frequency predictor or thermal monitor predicts an event. Reduce load on vulnerable units, increase spinning reserve, pre-start backup fuel.
4
Critical. Multiple layers predict critical events within lead time. Prepare for controlled separation, execute restart procedures, notify maintenance.

Design Principles

Four principles that shape every decision in the system.

Plant-centric, not TSO-centric. Existing systems serve the system operator. This system tells the plant what is coming and what to do about it.
Software-only deployment. No hardware installation, no plant shutdown, no sensor procurement. Connects via existing industrial protocols.
Validated, not assumed. Every layer is backtested against historical data. Linear threshold models achieve near-random accuracy on Nigerian grid data, proving that adaptive statistical methods are necessary.
Operator action mapping. Every alert comes with a specific recommended action. The system does not just warn, it tells the operator what to do.
Plan Moving Forward

Where This Goes Next

The research is validated. The next phase is pilot deployment at a major Nigerian power plant.

01
Pilot at a major power plant. Deploy the four monitoring layers against live SCADA data. Validate predictions against actual grid events over a defined period.
02
Refine alert thresholds. Calibrate the alert cascade to the specific plant's operating characteristics and historical event patterns.
03
Scale across the GENCO fleet. Extend to additional plants within the same operating group. Each plant adds validation data, improving model accuracy.
04
Industry platform. Offer the system to other Nigerian GENCOs facing the same grid instability. The problem is universal across all 28 grid-connected plants.

This Research Needs the Right Partners

The technology is validated. The grid crisis is ongoing. What comes next needs Nigerian power producers, investors, and grid experts who want to build something that creates real value.

Explore the Laboratory

See the systems

Explore the operational and financial environments we are developing.

Explore Systems →

Examine the projects

See how individual problems become working intelligence systems.

View Projects →

Explore autonomous agents

See how domain-specific agents operate inside defined workflows.

Explore Agents →

Discuss a partnership

If your organization supports research, technology development, or infrastructure, there may be an opportunity to contribute.

Start a Conversation →