Predictive operations

Move From Reactive Maintenance to Predictive Operations

IntelliGrid combines historical maintenance records, live equipment telemetry, alarms and operating behaviour to identify emerging risks before they interrupt facility operations.

Request a Demo See Live Operations

The business problem

Fixed schedules service the calendar, not the equipment

Time-based maintenance treats every asset the same regardless of how hard it has been working. Equipment in light duty is serviced unnecessarily while equipment under stress deteriorates between visits, and the first sign of a real problem is often an occupant complaint or an emergency call-out at a premium rate.

Unnecessary servicing on lightly used equipmentDeterioration missed between scheduled visitsUnexpected failures during occupied hoursHigh emergency repair cost and avoidable downtime

The IntelliGrid approach

Condition, history and behaviour scored together

IntelliGrid establishes a baseline per asset from its own operating history, then watches for statistically meaningful deviation across telemetry, alarm frequency and runtime. Risks are scored against asset criticality so the queue reflects operational impact, and a recommended action can be raised as a work order with one step.

Per-asset baselines rather than generic thresholdsRisk scored against criticality and service impactRecommended action attached to each flagged assetOne-step conversion from risk to scheduled work

AI risk queue

What the operations team sees each morning

Illustrative operational data.

CH-01 Chiller

Supply temperature 12% above established baseline over 9 days

High Risk

FP-02 Fire Pump

Abnormal vibration pattern detected during weekly test run

Medium Risk

AHU-03

Filter differential pressure increasing steadily across 3 weeks

Medium Risk

LIFT-07

Door cycle count approaching service interval threshold

Low Risk

The predictive intelligence screen

Predictive intelligence dashboard with asset risk scores and failure forecasts

Interface shown with illustrative demo data.

Capabilities

Predictive capabilities

Capability depends on the telemetry available for each asset class. Deployments typically start with the equipment that carries the most operational risk.

Degradation detection

Gradual decline in performance identified against the asset’s own established operating baseline.

Temperature trend analysis

Rising supply, return or bearing temperatures tracked over time rather than judged by a single reading.

Vibration abnormalities

Changes in vibration signature on rotating equipment such as pumps, fans and motors.

Pressure anomalies

Differential pressure drift across filters, strainers and pumping systems.

Power anomalies

Current and power-factor changes indicating mechanical load or electrical fault conditions.

Runtime behaviour

Cycling frequency, run hours and duty patterns compared against expected operation.

Repeated alarm analysis

Recurring alarms on the same asset treated as a deterioration signal, not isolated events.

Asset history analysis

Fault and maintenance history weighted into the current risk position.

Predictive risk scoring

A single comparable score per asset so a portfolio can be ranked consistently.

Remaining-life indicators

Where sufficient history and telemetry exist, an indicative remaining-life position to support replacement planning.

Workflow

From data to scheduled work

STEP 1

Data Collection

Telemetry, alarms, runtime and maintenance history gathered per asset.

STEP 2

Baseline

Normal operating behaviour established from the asset’s own record.

STEP 3

Anomaly Detection

Deviation identified across one or more signals.

STEP 4

Risk Scoring

Severity weighted by asset criticality and service impact.

STEP 5

Recommended Action

A specific intervention proposed for review.

STEP 6

Work Order

Approved action scheduled, assigned and tracked to closure.

Predictive position across a demo portfolio

Illustrative demo data — not verified customer outcomes.

94%

Model confidence on flagged assets

Assets at high risk this week

1,248

Assets under monitoring

21 hrs

Indicative downtime avoided

Business benefits

Intervention scheduled into planned hours instead of emergency response Maintenance effort concentrated on equipment that needs it Evidence for capital replacement decisions built from operating data Fewer occupant-visible failures on critical equipment Asset risk comparable across an entire portfolio

Connects with

IoT sensors BMS CCMS Energy meters Asset register Work order engine BI export

See integration architecture →

Frequently asked

How does predictive maintenance actually work here?

IntelliGrid builds a baseline of normal behaviour for each asset from its own telemetry, alarm and maintenance history. Deviation from that baseline is scored against the asset’s criticality, and a recommended action is proposed. It is a decision-support queue reviewed by your team, not unattended automation.

How much data is needed before predictions are useful?

Anomaly detection against a baseline becomes meaningful within weeks of continuous telemetry. Longer-horizon indicators such as remaining-life positions require more operating history and are introduced per asset class as data accumulates.