IBM Maximo Health 9.2: The New Power Transmission Scoring Models Explained
A product manager’s deep dive into the new APM industry models — what they are, how they work, and why they matter
Note: This post focuses specifically on the Category E models shipped in 9.2. The full release also includes major performance improvements (80–90% faster score calculation), multi-tab configurable dashboards, AI-powered Condition-Based Insights, a real-time Alerts framework, and a new Ticket Application. Each of those deserves its own post.
The problem we’re solving
Utilities manage thousands of high-value electrical assets — power transformers, circuit breakers, switchgear, cables, support structures — yet most maintenance decisions are still based on asset age alone.
· The oldest assets get replaced first; the rest get inspected on a schedule, if at all
· When something fails unexpectedly, crews mobilise in emergency mode — and the business pays the price
· Most utilities already have the data to make better decisions: oil test results, inspection records, lab reports, electrical test results — all there, scattered across systems, never connected
· The question that's never answered: which asset do I need to act on, and when?
IBM Maximo Health is built to answer that question — and in version 9.2, we expanded coverage with Category E scoring models.
What is a scoring model?
A scoring model is a transparent, auditable engine that takes in asset data and produces three outputs:
• Health Score — a 0–100 score representing current condition
• End of Life — the statistical likelihood of an asset reaching end of life.
• Effective Age — how old the asset is actually behaving, based on condition (not calendar)
Each scoring model is a Jupyter notebook, with python code paired with a configuration file. Key principle: you cannot average away a bad component — a failing bushing pulls the entire transformer's score down, regardless of healthy oil readings elsewhere.
What are Category E scoring models?
The “Category E” designation refers to the new electrical power transmission scoring models that were built on a field tested scoring methodology. All new scoring models carry the Category E suffix in their notebook and configuration file names.
|
Feature
|
Earlier models
|
Category E scoring models
|
|
Scoring philosophy
|
Weighted averages
|
Each health score is built from real meter readings and inspection data — aggregated into a single result per asset
|
|
Child component integration
|
Limited
|
Full parent-child hierarchy — worst child pulls parent down
|
|
Effective age calculation
|
Age curve only
|
Condition-adjusted effective age
|
|
Asset families covered
|
Transformers, circuit breakers
|
30+ asset types across the full electrical network
|
How the scoring system works
The diagram below shows the end-to-end flow. Asset data enters through Maximo Manage, the scoring group triggers the notebook job, and scores are written back to the Maximo Health dashboard.
Transmission networks: a new frontier
One of the strategic milestones in 9.2 is explicitly extending IBM Maximo Health to Power Transmission organisations — not just distribution utilities.
For organizations that operate both transmission and distribution networks — and for independent transmission system operators (TSOs) — this means a single platform can now health-score the entire asset portfolio from generation interconnect to the low-voltage network.
Why this is different from what came before
Three things make the Category E approach genuinely different:
1. Built on real utility knowledge, not theory. The scoring methodology was developed with leading energy and utilities companies. It is not a generic algorithm applied to power assets.
2. Embedded in the system of record. The models run inside IBM Maximo, where asset records, inspection findings, and work orders already live. Health scores and risk rankings appear in the same dashboard the operations team uses every day.
3. Scales across the entire fleet. A manual condition assessment might cover the highest-risk assets. These scoring models score every transformer, every bushing, every circuit breaker in the network.
Getting started
The models are available in Maximo Application Suite (MAS) version 9.2. Each model ships with:
- A Jupyter notebook containing the scoring logic
- A configuration file for environment-specific setup
- Prerequisites documentation covering the database schema updates, domain values, and meter definitions needed in Maximo Manage
The first step is configuring the External Scoring Group in Maximo Health and connecting it to a Watson Studio project on Cloud Pak for Data. From there, each model can be configured and scheduled as a notebook job.
The models are available in MAS version 9.2. For setup instructions, prerequisites, and the full model list, see the IBM documentation for Maximo Health