IIoT Platform Cutting Manufacturing Downtime 47% Across 6 Production Lines

The OEE Gap

OEE is not just a metric. It is a direct measure of lost revenue. At 54% OEE 46% of production capacity is wasted.

For Calvert:

• 6 CNC production lines
• 23 machines
• 85% capacity utilisation

The gap between:

• Current OEE: 54%
• Target OEE: 71%

= ~£2.8M in lost annual production capacity

The Root Problem

The issue wasn’t just downtime. It was lack of visibility.

• Downtime recorded on paper → ~40% data completeness
• No real-time machine state
• No failure pattern tracking
• Maintenance = reactive firefighting

You cannot optimise what you cannot measure.

System Architecture

This was designed as a full IIoT stack, not just sensor deployment.

Edge Computing Layer

Each of the 23 CNC machines was equipped with an industrial edge unit:

• Device: Advantech UNO-2473G
• Sampling rate: 100Hz

Sensor Inputs

Spindle Load

• Current draw monitoring
• Detects cutting resistance anomalies

Vibration (3-axis accelerometer)

• Identifies bearing wear and imbalance
• Key for spindle health prediction

Thermal Monitoring

• Motor housing hotspots
• Electrical panel heat signatures

Cycle Data (OPC-UA integration)

• Program ID tracking
• Cycle time accuracy
• Production state

Edge Processing Strategy

• Raw data processed locally at 100Hz
• Aggregated metrics sent to cloud at 1Hz
• Full-resolution data transmitted only on anomaly

Result: 94% reduction in data transmission cost while preserving diagnostic fidelity.

Real-Time OEE Visibility

Production managers now have live OEE tracking:

OEE Components

• AvailabilityPlanned vs. unplanned downtime
• PerformanceActual vs. ideal cycle time
• QualityFirst-pass yield and scrap rate

Operator Workflow Transformation

Paper → digital

• Tablet interface at each machine
• Downtime recorded in <15 seconds
• Structured reason tree

Data completeness improved: ~40% → >95%
This alone unlocked actionable insight.

Predictive Maintenance Layer

We built 6 machine-specific failure models, trained on:

• 18–36 months historical maintenance data
• Real-time sensor signals
• Operational load patterns

Failure Modes Detected

1. Spindle bearing wear
2. Coolant pump failure
3. Tool wear progression
4. Chuck jaw degradation
5. Ball screw backlash
6. Hydraulic pressure decay

Model Output

• Daily asset health score
• Failure probability window (5–14 days)
• Trigger-based alerting

Maintenance moves from: after failure → before failure

ERP Integration

Prediction must translate into action.

• Alerts automatically generate work orders
• Integrated with existing ERP system
• Maintenance teams see prioritised queue

Enables:

• planned maintenance windows
• reduced disruption to production

Performance Outcomes

OEE Improvement

• 54% → 71%
• +17 percentage points

Downtime Reduction

• 4.2 → 2.2 hours/day per line
• 47% reduction

Maintenance Efficiency

• Response time:4.2 hours → 47 minutes

Cost Savings

• Tool optimisation alone:£180K annual savings

Operational Impact

Before:

• Reactive maintenance
• Poor data quality
• Hidden inefficiencies
• High downtime cost

After:

• Real-time production visibility
• Predictive maintenance capability
• Structured downtime analytics
• Data-driven operations

Why This Worked

1. High-frequency data capture (100Hz)Enabled early detection of failure patterns
2. Edge-first architectureReduced cost while preserving insight
3. Operator integrationData quality improved dramatically
4. Focused failure modellingTargeted highest-impact downtime drivers
5. ERP integrationClosed the loop between insight and action

The Key Insight

Manufacturing inefficiency is rarely due to a single failure. It is:

• small delays
• repeated micro-failures
• untracked inefficiencies

IIoT makes these visible.
Visibility → Prediction → Action

Final Outcome

Calvert transitioned from:

• Reactive → Predictive operations
• Paper-based → Data-driven manufacturing
• Hidden losses → Measurable optimisation

Result:
A production system that is:

• more efficient
• more predictable
• significantly more profitable

Looking to Improve OEE?

Intagleo builds IIoT platforms that turn manufacturing data into measurable performance gains.

Start a conversation