Service · Digital Engineering · ARaaS® Toolbox

Know the failure
is coming.
Act before it does.

Most maintenance programmes respond to what has already happened. Predictive and prescriptive maintenance changes the operating model — detecting the early signature of failure, determining the optimal intervention and acting before the asset reaches functional failure. Optimal deploys this capability within the ARaaS® framework, integrated with the reliability strategy and directly connected to the maintenance plan.

Start a Conversation ARaaS® Framework →

Service Summary

Vibration analysis, thermography, oil analysis and ultrasound diagnostics

SCADA, historian and sensor data integration and analytics

Machine learning models for anomaly detection and remaining useful life

Prescriptive recommendations — what to do, when and why

Digital twin simulation for critical asset classes

Integration with CMMS work order workflow

Mining · Oil & Gas · Nuclear · Power · FMCG · Transport

SAP PM · IBM Maximo · OSIsoft PI · Infor EAM · Wonderware

RCM-integrated analytics · P-F curve methodology

The P-F Curve — The Basis for Everything

Every failure
announces itself.
Most operations miss it.

The P-F curve describes the relationship between a potential failure — the point at which the deterioration process begins to produce a detectable signature — and functional failure — the point at which the asset can no longer perform its required function. The interval between P and F is the window within which detection, diagnosis and action are possible.

Predictive maintenance is the discipline of detecting the signature at P. Prescriptive maintenance is the discipline of determining the optimal action to take between P and F. Optimal's programme deploys both — using the ARaaS® Toolbox to shift the intervention point as far left on the P-F curve as the data and technology permit.

Reactive ZoneActing after functional failure — defect correction, unplanned downtime, consequential damage. The most expensive maintenance strategy.

Predictive ZoneActing between P and F — planned intervention before failure, minimum disruption to production, controlled repair conditions.

Optimal ZoneActing as early as possible after P detection — maximum lead time, optimal parts and resource planning, lowest total intervention cost.

P-F Interval — Condition vs. Time

P — Detection window opens

F — Functional failure

P-F interval — intervention window

Why Reactive Maintenance Persists

Data is collected.
Intelligence is not extracted.

Most asset-intensive operations already collect enough data to support a predictive maintenance programme. SCADA systems capture process parameters. Vibration sensors are installed on critical rotating equipment. Oil analysis programmes are running. The CMMS contains years of work order history. The data exists. What is missing is the analytical infrastructure to convert it into actionable intelligence.

The result is an operation that is nominally data-rich but practically reactive — responding to alarms that trigger after the condition has already deteriorated beyond the optimal intervention point, replacing parts on a time basis without knowing whether the asset is actually deteriorating, and missing failure signatures that the data would have revealed if anyone had been looking.

Predictive and prescriptive maintenance is not primarily a technology problem. It is an integration and methodology problem. Connecting the right data sources, applying the right analytical methods for each failure mode and failure signature, and ensuring the output reaches the maintenance planning process in time to act — that is the Optimal programme.

Sensor data collected but never analysedVibration, temperature, pressure and flow data logged continuously by SCADA and historian systems — but reviewed only when an alarm triggers, by which point the failure signature may already be advanced. The P-F interval is consumed before detection.

Time-based maintenance applied regardless of conditionPMs scheduled at fixed intervals based on OEM recommendations or historical precedent — not on actual asset condition or failure mode P-F interval. Some assets are maintained too frequently (cost); others reach failure between planned interventions (risk).

Alarm thresholds set without P-F interval analysisSCADA alarm thresholds set at absolute limits — the point at which action is unavoidable — rather than at the earliest point of detectable deterioration. By the time the alarm fires, there is insufficient lead time for planned intervention.

No link between condition data and CMMS work ordersEven where condition monitoring data is reviewed and a deterioration trend identified, the process of raising a work order, sourcing parts and scheduling the intervention is manual and slow. The P-F interval is eroded by process friction rather than analytical failure.

Prescriptive capability absent — data without decision supportAnalytics platforms generate alerts, but no systematic process converts the alert into a recommended action — which tasks to perform, which parts to procure, what the urgency is. Engineers must interpret raw data signals themselves, without the decision support framework that converts detection into optimal action.

Condition Monitoring Techniques

The right technique
for every failure signature

Each failure mode produces a detectable signature — but the signature type determines which monitoring technique will detect it at the earliest point on the P-F curve. Optimal selects and deploys the technique set appropriate for each asset class and failure mode, integrating outputs into the unified analytics platform.

Rotating Equipment

Vibration Analysis

The primary predictive technique for rotating machinery — motors, pumps, fans, compressors, gearboxes. Detects imbalance, misalignment, bearing defects, looseness and resonance. Frequency spectrum analysis identifies the specific failure mode and its severity level. Deployed continuously via online sensors or periodically via route-based portable instruments.

Continuous onlinePortable route-basedFFT spectrumEnvelope analysis

Electrical & Thermal

Thermographic Imaging

Infrared thermography detects heat signatures from electrical connections, switchgear, transformers, motor windings and mechanical friction. A loose electrical connection generating heat is detectable months before it reaches the failure point. Deployed as a periodic survey programme on electrical assets and as continuous monitoring on critical high-voltage equipment.

Periodic surveyElectrical systemsMechanical frictionRefractory

Lubrication Systems

Oil & Lubricant Analysis

Oil analysis detects wear particle generation, contamination, degradation and additive depletion. Wear particle morphology and element concentration identifies the component generating wear — gears, bearings, cylinder liners — and distinguishes between normal wear and abnormal wear rate. Extended oil drain interval optimisation is a direct financial benefit alongside the predictive value.

Wear particle analysisViscosityContaminationOil drain optimisation

Structural & Pressure

Ultrasonic Testing

Ultrasound detects early-stage bearing defects at very high frequencies before they become detectable by standard vibration analysis — providing the earliest possible warning on the P-F curve. Also applied for compressed air and steam leak detection, valve seat leakage, cavitation in pumping systems and electrical arc/partial discharge detection in switchgear.

Airborne ultrasoundContact ultrasoundEarly bearing detectionLeak detection

Process & Performance

SCADA & Process Data Analytics

Process performance analytics applied to SCADA, historian and DCS data streams — detecting efficiency degradation, abnormal process parameter trends and developing thermal or hydraulic performance deviations before they reach alarm thresholds. Statistical process control and multivariate analysis applied to identify complex failure signatures that single-parameter monitoring misses.

Historian integrationSCADA analyticsSPCMultivariate

Fluid Systems

Motor Current Signature Analysis

MCSA detects mechanical and electrical faults in motors and driven equipment by analysing the current waveform for characteristic frequency signatures. Detects rotor bar defects, bearing eccentricity, mechanical load variations and broken rotor bars without requiring direct access to the motor. Non-intrusive — deployed using existing current measurement infrastructure.

Non-intrusiveRotor bar faultsMotor eccentricityLoad analysis

ARaaS® Toolbox · Analytics Capability

How the ARaaS® Toolbox
delivers predictive intelligence

Predictive and prescriptive maintenance is delivered through four integrated tools within the ARaaS® Toolbox — combining the reliability engineering foundation, cross-domain analytics, prescriptive decision support and field diagnostics confirmation. Together they form a complete programme from strategy to action.

Full ARaaS® Framework

Strategy · StrategiseReliability-Centred Maintenance (RCM)
RCM identifies the failure modes and P-F intervals that determine which condition monitoring technique to apply and what detection threshold to set. Without the RCM foundation, predictive monitoring is applied to the wrong assets using the wrong techniques at the wrong thresholds. RCM is the analytical basis for every monitoring decision in the programme.

Analytics · Strategise & MonitorPredictive Analytics
Cross-domain analysis applied to vibration, process, electrical and lubrication data simultaneously — detecting complex failure signatures that single-parameter monitoring misses. Unsupervised methods automatically identify anomalies across the asset population. Machine learning models trained on historical failure events provide remaining useful life estimates.

Decision Support · Deploy & ExecutePrescriptive Maintenance
Detection without action is wasted. Prescriptive maintenance converts the analytical output into a recommended course of action — which maintenance tasks to perform, which parts to procure, what the urgency level is and when the intervention window closes. The output integrates directly with the CMMS work order workflow, eliminating the manual interpretation step.

Field Verification · Deploy & ExecuteField Diagnostics
Remote analytics generates the alert and the recommended action. Field diagnostics — portable vibration instruments, thermographic cameras, ultrasound equipment and oil sampling — verify the condition on the asset, capture the failure signature in detail and confirm the intervention is warranted before the maintenance plan is executed. Closes the detection-to-action loop.

Analytics Maturity Model

Four tiers.
From reactive baseline
to prognostic intelligence.

Optimal assesses current analytics maturity at the start of every programme and defines the target maturity tier based on asset criticality, data availability and commercial objectives. Most organisations enter at Tier 1 or 2 and progress to Tier 3 within a 12–18 month programme. Tier 4 is applied to the highest-criticality asset classes where prognostic accuracy delivers measurable commercial return.

Tier 01

Automated Analytics

The Baseline

Standardised reporting on what has already happened.

Analyses streamed data from multiple sources to detect anomalies. Unsupervised — automatically draws inferences from data. Deployed as the foundation layer across all assets in scope. The entry point for every Optimal analytics programme.

Reactive

Tier 02

Condition-Based Rules

The Watchdog

Logic-driven alerts when specific boundaries are breached.

Captures operator and engineer experience in rules-based logic — set threshold boundaries derived from P-F interval analysis and failure mode consequence. Targets assets with well-understood failure modes and reliable sensor coverage. Automated alert routing to CMMS.

Prescriptive

Tier 03

Guided Analytics

The Investigation

Pattern recognition and trend modelling across multiple data streams.

Multi-variable analysis across vibration, process, electrical and lubrication data. Statistical models detect developing trends before threshold breach. Alert confidence scoring — distinguishing genuine developing failure from transient process variation. Target maturity for most critical asset classes.

Predictive

Tier 04

Advanced Analytics

The Oracle

Machine learning models forecasting future asset state.

Targeted at the highest-criticality assets. Trained on cleansed historical failure data — multi-variable machine learning with rigorous model validation and continuous refinement. Provides remaining useful life estimates and optimal intervention windows. Digital twin integration for what-if simulation.

Prognostics

How Optimal Delivers

From baseline to
operational programme

Optimal's predictive and prescriptive maintenance programme follows a four-phase implementation — from analytics readiness assessment and data architecture through technique deployment and model development to operational handover and continuous improvement governance.

Analytics Readiness Assessment

Evaluate existing data infrastructure — sensor coverage, data quality, historian architecture, SCADA configuration, CMMS work order structure and current analytics maturity. Identify gaps, prioritise asset classes by criticality and P-F interval opportunity, and define the target analytics maturity tier for each asset group. Output: programme scope, data architecture requirements and business case.

Data Architecture & Integration

Design and implement the data integration layer — connecting SCADA historians, sensor networks, CMMS and condition monitoring instruments into a unified analytics environment. Data cleansing, normalisation and validation applied before any model development begins. CMMS integration configured for automated work order generation from analytics alerts. Establishes the foundation that all subsequent analytical work depends on.

Model Development & Technique Deployment

Deploy condition monitoring techniques and develop analytics models per asset class — selecting technique, configuring thresholds based on P-F interval analysis, training machine learning models on historical failure data, building prescriptive logic from FMECA failure mode database. Alert routing configured to CMMS. Field diagnostics protocols established for on-site verification. Parallel run period with results validation against physical condition.

Operational Handover & Continuous Improvement

Transition programme to operational use — team training, alert interpretation procedures, CMMS workflow integration and governance documentation. Standing model improvement cycle: every confirmed failure event feeds back into model refinement. Monthly analytics review with engineering leadership. Annual maturity assessment against target tier progression. Programme value tracking — planned vs. reactive maintenance ratio, mean time between failures trend, total maintenance cost per unit.

Data Integration

Six sources.
One unified analytics picture.

Real-Time · Continuous

SCADA & DCS Historians

Process parameter streams — temperature, pressure, flow, speed, current, voltage — integrated from SCADA, DCS and OSIsoft PI historian systems. The highest-frequency data source in the programme. Configured for anomaly detection against P-F derived thresholds and multivariate statistical modelling for complex process failure signatures.

Continuous · Online

Online Vibration Networks

Permanently installed accelerometers on critical rotating equipment — continuous waveform capture and spectral analysis. Integrates with major online monitoring platforms. Data normalised for operating condition variation — speed, load, temperature — before trend analysis. Automated spectrum comparison against baseline and alarm threshold management.

Periodic · Route-Based

Portable Condition Monitoring

Route-based vibration, ultrasound and thermography data collected on a defined frequency by trained Optimal field technicians or client maintenance team. Data uploaded to the analytics platform after each route, triggering comparison against historical trends and machine learning model scoring. Extends predictive coverage to assets not warranting permanent sensor investment.

Periodic · Laboratory

Oil & Lubrication Analysis

Wear particle, viscosity, contamination and additive condition data from laboratory oil analysis. Integrated into the unified analytics platform with trend analysis and alert generation against programme-specific thresholds. Wear particle morphology data from ferrous debris monitors for high-criticality gearboxes and hydraulic systems with continuous sample ports.

Operational · CMMS

CMMS Work Order History

Work order history, failure codes, repair descriptions, parts usage and labour records from the CMMS — providing the historical failure signature database that machine learning models are trained on. CMMS integration also provides the output channel: analytics alerts automatically generate work orders with the prescriptive action, urgency and required parts appended.

Periodic · Infrared

Thermographic Survey Data

Infrared thermal imaging data from periodic electrical and mechanical surveys. Archived in the analytics platform with severity classification, trend comparison against previous surveys and automated work order generation for findings above intervention threshold. Integration with the electrical maintenance schedule to ensure survey frequency aligns with P-F interval for each circuit type.

The Prescriptive Difference

From alert
to action in one step

Predictive maintenance generates alerts. Prescriptive maintenance generates actions. The difference is significant — an alert requires an engineer to interpret the signal, diagnose the probable cause, determine the appropriate response, identify the required parts and raise a work order. That process takes time, and time is what the P-F interval has in finite supply.

Optimal's prescriptive capability encodes the diagnostic and decision logic that the engineer would otherwise apply manually — using the FMECA failure mode database to map the detected signature to the probable failure mode, and the maintenance task library to select the optimal intervention. The CMMS work order is generated automatically, with the task list, required parts and urgency level pre-populated. The engineer validates and approves. The lead time saved is used to plan and prepare the intervention rather than to decide what it is.

Alert generated from analytics model at threshold breach →

Failure mode probability scored from FMECA signature mapping →

Optimal intervention selected from RCM task library →

CMMS work order generated — tasks, parts, urgency pre-populated →

Engineer validates and approves — full lead time preserved for execution →

RCM Studies — the analytical foundation

Predictive maintenance analytics dashboard

Evidence of Delivery

Predictive maintenance
in practice

All case studies

Case studies below are anonymised. Client consent is required before specific project details are attributed publicly. Contact us to arrange reference calls.

Mining · Crushing & Processing · Southern Africa

Major Mining Group — Rotating Equipment Predictive Maintenance Programme

Six-site open-pit and underground mining operation. Critical rotating equipment — primary crushers, ball mills, conveyor drives and slurry pumps — maintained primarily on time-based schedule with significant unplanned failure rate. Online vibration monitoring installed on some assets but not integrated into a unified analytics programme. CMMS work order data available but not connected to condition data.

15%

Improvement in overall asset availability across the six processing sites within 18 months of full programme deployment — driven by reduction in unplanned failure events on primary and secondary crushing equipment

6 sites

Unified analytics platform deployed across all six sites with common RCM-based monitoring logic, alert routing to CMMS and field diagnostics programme — shared intelligence across the asset population

Power Generation · Energy Recovery · UK

ERF Portfolio Operator — Turbine & Rotating Plant Analytics

Eight energy recovery facility portfolio. Steam turbine generator sets, ID and FD fans, and combustion air compressors were the primary unplanned failure drivers. SCADA historian data existed but was used only for process alarm management. No predictive monitoring programme in place. Insurance target — maintaining greater than 92% availability across the portfolio — was being missed at three sites.

92–98%

Turbine and generator availability achieved across the portfolio following programme deployment — with three previously underperforming sites moving from below 90% to above 95% within 12 months

Tier 3

Analytics maturity achieved on the primary rotating equipment population — guided analytics with multi-variable failure signature detection and automated CMMS work order generation from alert events

Oil & Gas · FPSO · West Africa

Offshore Operator — Gas Compression Train Predictive Programme

FPSO with four gas compression trains as the primary production bottleneck. A single compressor train trip caused significant production deferral — quantified at over $500K per day at prevailing gas prices. Online vibration monitoring was installed but analysed manually, intermittently. No integration between the monitoring platform, SCADA historian and CMMS. Two unplanned compressor trips in the preceding 12 months had not been prevented.

Zero

Unplanned gas compression train trips in the 24 months following programme deployment — with two developing bearing defects and one rotor unbalance event detected and actioned through planned intervention before reaching functional failure

Prognostic

Tier 4 analytics maturity achieved on primary gas compression train bearings — remaining useful life modelling deployed with confidence intervals, enabling optimal intervention timing within the production schedule

FMCG · Manufacturing · Europe

Consumer Goods Manufacturer — OEE Improvement through Predictive Programme

High-speed packaging and filling lines. OEE below target across three sites. Mechanical failures on packaging line drives, filling heads and conveyor systems were the primary OEE constraint. Maintenance was predominantly reactive and time-based. No condition monitoring programme in place. SCADA data captured but not analysed beyond alarm management. Board-level requirement to demonstrate OEE improvement through structured maintenance investment.

OEE +8%

Overall equipment effectiveness improvement across the primary bottleneck lines — driven by reduction in unplanned mechanical failures and shorter planned maintenance interventions from condition-led timing and pre-staged parts

CMMS

Full CMMS integration achieved — analytics alerts automatically generating work orders with task list and required parts pre-populated, eliminating the manual interpretation and planning step between alert and action

Asset Performance Management

Predictive maintenance
is one part of a
wider APM programme.

The highest return from predictive and prescriptive maintenance is realised when it is deployed as part of the full ARaaS® Asset Performance Management programme — not as a standalone analytics project. The RCM study defines which failure modes to monitor and which techniques to apply. The maintenance strategy determines what actions the prescriptive system recommends. ISO 55001 governance ensures the programme is reviewed and improved. Without the connected programme, analytics generates alerts that go unacted on.

RCM FMECA defines monitoring techniques per failure mode

Maintenance task library provides the prescriptive action database

ISO 55001 governance embeds continuous improvement

Optimal360™ provides the performance visibility layer

GARPI™ benchmarks analytics maturity against global peers

Explore ARaaS® Framework Book a discovery call →

Programme Deliverables

What is in place
at programme completion

D01

Analytics Readiness Report & Programme Architecture

Documented assessment of data infrastructure readiness — sensor coverage gaps, data quality findings, historian architecture recommendations and CMMS integration requirements. Target analytics maturity tier per asset class with commercial justification. Programme architecture document defining the data integration layer, analytics model scope, technique selection per asset group and CMMS interface design.

D02

Condition Monitoring Programme & Route Schedules

Documented condition monitoring programme per asset class — technique assigned per failure mode and P-F interval, monitoring frequency, alert thresholds based on P-F interval analysis, escalation procedures and field diagnostics confirmation protocols. Route schedules for portable monitoring instruments. Integration specification for online sensor data. All traceable to the RCM FMECA failure mode database.

D03

Integrated Analytics Platform & CMMS Interface

Operational analytics platform with data integration across all specified sources — SCADA historian, sensor networks, CMMS and condition monitoring instruments. Analytics models deployed and validated per asset class. CMMS work order generation interface configured and tested. Alert routing, escalation workflows and work order templates documented. Parallel run validation report confirming alert accuracy against known failure events.

D04

Prescriptive Logic Library & Alert Interpretation Guide

Prescriptive action library per failure mode — derived from the RCM task library, mapped to analytics alert types and configured in the CMMS interface. Alert interpretation guide for engineering and maintenance planning teams — describing each alert type, the failure mode it represents, the confidence scoring methodology and the required response. Used for team training and as the standing operating reference for the programme.

D05

Analytics Maturity Assessment & Tier Progression Plan

Documented current analytics maturity per asset class at programme completion — Tier 1 through Tier 4, with evidence. Target tier progression plan for the following 12 months — identifying which asset classes are candidates for maturity advancement, what additional data or model development is required, and the commercial case for each progression step. Used as the basis for the annual programme review.

D06

Programme Governance & Performance Framework

Standing governance framework for the predictive maintenance programme — monthly alert review, quarterly performance reporting, annual maturity assessment and model improvement cycle. KPI framework: planned vs. reactive maintenance ratio, P-F interval utilisation rate, alert-to-work-order conversion rate, mean time between failures trend and total maintenance cost per unit. Dashboard configuration in Optimal360™ or existing analytics platform.

Related Services

Services that form the
connected APM programme

All services

GARPI™ — Global Asset Reliability & Performance Index

How does your analytics
maturity compare globally?

GARPI™ Dimension 4 — Data & Digital Capability — measures whether your organisation has the data infrastructure, analytics maturity and digital integration to support a predictive maintenance programme. Dimension 3 — Maintenance Strategy & Execution — measures how well analytics drives action rather than generating unactioned alerts. Benchmark your current position against global peers in your sector — free, anonymous, 12 minutes.

Take the Free Survey Discuss your results

Dim 1

Asset Performance Outcomes

Dim 2

Reliability Governance

Dim 3 — Focus

Maintenance Strategy & Execution

Dim 4 — Focus

Data & Digital Capability

Dim 5

Lifecycle Value & Financial Alignment

Dim 6

Workforce Capability & Knowledge

Dim 7

Spares & Materials Management

Dim 8

Strategic Outlook

Next Steps

Ready to move
left of P?

Whether your starting point is a specific asset class causing production loss, a board mandate to move from reactive to predictive maintenance, or an existing analytics platform generating alerts that aren't being acted on — Optimal can design and deliver the programme. We start with an analytics readiness assessment so that the programme scope reflects your actual data infrastructure, not a generic deployment template.

The readiness assessment takes two to three days on site, produces a programme architecture document with commercial justification and is completed before any commitment to implementation scope or investment. Start there.

Book a Discovery Call GARPI™ Survey — Dim 3 & 4

Global Enquiries

enquiries@optimal.world
optimal.world/contact-us

Credentials

ISO 9001:2015 certified · IAM Member 1035342
ISO 55001 advisory · GFMAM aligned methodology

Practice Area

Digital Engineering — part of the Asset Reliability

Know the failureis coming.Act before it does.

Every failureannounces itself.Most operations miss it.

Data is collected.Intelligence is not extracted.

The right techniquefor every failure signature

How the ARaaS® Toolboxdelivers predictive intelligence

Four tiers.From reactive baselineto prognostic intelligence.

From baseline tooperational programme

Six sources.One unified analytics picture.

From alertto action in one step

Predictive maintenancein practice

Predictive maintenanceis one part of awider APM programme.

What is in placeat programme completion

Services that form theconnected APM programme

How does your analyticsmaturity compare globally?

Ready to moveleft of P?