Modern steel plants generate millions of PLC samples per shift. Presenting raw alarms or raw trends creates noise and operator fatigue. This framework converts second-by-second signals into structured, explainable summaries that help shift engineers answer only what matters: what needs attention now, how urgent it is, and since when the issue exists.
Traditional SCADA systems overwhelm operators with continuous data streams—temperature readings every second, vibration measurements 3,600 times per hour, pressure values flooding control rooms. While this granular data captures everything, it provides no intelligence. Operators face the impossible task of identifying critical patterns among millions of data points while managing production simultaneously.
Download Your OXmaint CMMS PDF — Start Improving Uptime Today
Ready to streamline maintenance operations, cut downtime, and take control of asset performance? Download the full OXmaint guide now — then book a live demo to see how our platform transforms maintenance for teams like yours.
This development specification defines the complete operational logic for transforming ABB PLC sensor data via OPC into intelligent shift logbooks. Designed for automation engineers, reliability managers, and CMMS/MES development teams in heavy industry, it provides implementation-ready algorithms for hourly aggregation, four-level equipment classification, and shift-level reporting that maintenance teams can trust and act upon.
Transform PLC Data Into Maintenance Intelligence
Convert millions of sensor samples into clear shift-level decisions—no operator fatigue, no missed critical issues.
01 Business & Operational Overview
Understanding the challenge facing modern steel plants and the intelligent solution framework.
The Challenge
Information Overload
Millions of PLC samples generated per shift create data overwhelm that operators cannot process effectively.
Operator Fatigue
Raw alarms and trends cause continuous noise, leading to missed critical issues and alarm desensitization.
No Prioritization
Shift engineers struggle to identify what truly needs attention versus normal operational variation.
Legacy Limitations
Traditional systems lack intelligent filtering, trend analysis, and automatic prioritization capabilities.
The Solution
Structured Summaries
Second-by-second signals converted into explainable hourly summaries with statistical validation.
Three Critical Answers
What needs attention now? How urgent is it? Since when does the issue exist? Clear, actionable answers.
4-Level Classification
NORMAL / OBSERVATION / ALERT / CRITICAL—precise categorization with clear operational meanings.
Episode Tracking
Records since-when and since-which-shift abnormality started, providing full historical context.
Need help implementing this framework? Schedule a technical consultation — Get expert guidance from automation engineers who have deployed this specification across 50+ steel plants worldwide.
02 End-to-End Logic Flow (From PLC to Shift Logbook)
Complete data transformation pipeline from raw sensor signals to actionable maintenance intelligence.
Data Flow Pipeline
PLC Data Collection
PLC publishes raw sensor values every second (temperature, vibration, current, pressure)
OPC Server Integration
OPC server exposes tags; collector timestamps and stores raw samples with quality flags
Hourly Window Grouping
Raw samples grouped into fixed hourly windows per tag (up to 3,600 samples/hour)
Statistical Aggregation
For each hour and tag: min, max, average, median computed from valid samples
Trend Metrics Calculation
Rate-of-change, trend direction, occurrences, excursions analyzed
4-Level Classification
Each tag classified into NORMAL / OBSERVATION / ALERT / CRITICAL bands
Episode Tracking
Records since-when and since-which-shift abnormality started
Equipment Correlation
All signals from same equipment correlated into single hourly health record
Shift Summary Rollup
Hourly records rolled up into comprehensive shift summaries
Priority-Sorted Output
Only abnormal equipment appears in Equipment Status table, sorted by priority
Want to see this pipeline in action? Schedule a live demo — Watch PLC data flow through each stage from raw samples to intelligent shift logbooks in under 30 minutes.
03 Hourly Aggregation Logic (Per Signal)
Hourly aggregation is the foundation of noise-free predictive maintenance. Each hour may contain up to 3,600 samples per signal (1 sample/second). Aggregates are computed only from valid samples.
| Metric | How It Is Calculated | Why It Matters |
|---|---|---|
| Minimum | Lowest value observed in the hour | Detects under-range conditions |
| Maximum | Highest value observed in the hour | Detects spikes and overloads |
| Average | Arithmetic mean of all samples | Represents general operating level |
| Median | Middle value after sorting samples | Robust against spikes |
| Rate of Change | Current hour median minus previous hour median | Detects acceleration |
| Trend Direction | UP / DOWN / NEUTRAL based on rate-of-change threshold | Shows worsening vs stabilizing |
| Occurrences | Count of samples outside defined limits | Measures persistence |
Need custom metrics for your equipment? Talk to our engineers — Get help defining custom aggregation logic, thresholds, and trend calculations specific to your equipment types.
04 Overall Equipment Condition Classification
Equipment condition is classified into four distinct levels based on signal analysis and trend behavior.
Overall Equipment Condition Classification Flow
Equipment progresses through severity levels as signals degrade. Episode tracking records since-when and since-which-shift abnormality started.
Questions about classification logic? Schedule an expert review — Our reliability engineers can review your specific use cases and recommend optimal threshold configurations for OBSERVATION, ALERT, and CRITICAL levels.
05 Shift Logbook – Equipment Status (Action List)
This table is the final operational output. It lists only equipment requiring attention. Issues may originate in the current shift or be carried forward from previous shifts.
| Priority | Equipment | PDM | Since (Shift) | Trend | Occur. | Key Metric | Overall |
|---|---|---|---|---|---|---|---|
| P1 | Motor XYZ (MTR-1123) |
6 hrs | Today Shift A | Fast Upward |
37 | Temp 82°C | CRITICAL |
| P2 | Gearbox Fan (FAN-044) |
2 days | Yesterday Shift C |
Upward | 15 | Temp 74°C | ALERT |
| P3 | Pump P-07 | 3 shifts | Today Shift B | Stable | 9 | Temp 68°C | OBSERVE |
Sorting Rule
Order by Priority (P1 → P4). Within the same priority, sort by lowest PDM first.
Key Logbook Components
Priority
P1-P4 ranking based on criticality and business impact
PDM
Predicted Days to Maintenance or time window for intervention
Since (Shift)
Episode start tracking—when abnormality first detected
Trend
Directional movement: Fast Upward, Upward, Stable, Downward
Occur.
Occurrence count—persistence measure of abnormal readings
Key Metric
Primary signal driving the classification decision
Want to customize the shift logbook? Contact our implementation team — Adapt the logbook format to match your shift handover procedures, add custom fields, and integrate with existing CMMS systems.
Ready to Eliminate Operator Fatigue?
Deploy OXmaint Factory AI's shift logbook framework and transform millions of PLC samples into clear, priority-sorted maintenance actions.
