The cost-effectiveness of predictive maintenance: from a reactive to a prescriptive approach

1. Introduction: Transformation of Asset Management in Ukrainian Industrial Production

Modern industrial production in Ukraine is on the threshold of significant changes in approaches to maintenance. Traditional strategies based on reactive actions or fixed schedules demonstrate limited effectiveness in the face of challenges of global competition and increasing performance requirements. The shift from reactive (performing repairs after failure) and preventive (planned work on schedule) to predictive and prescriptive maintenance is critical to optimizing operational efficiency and ensuring competitiveness. This transition involves the use of advanced monitoring, data analysis and artificial intelligence technologies to predict potential equipment failures and provide specific recommendations to prevent them. The goal is to maximize uptime, reduce operating costs, and extend the life cycle of production assets.

2. Scientific Foundations and Technologies of Prognostic Maintenance

The foundation of predictive maintenance (PRM) is based on continuous monitoring of the physical parameters of the equipment and analysis of the received data to detect early signs of degradation or potential failures. The main technologies include:

Vibration analysis: Uses vibration sensors to measure the amplitude, frequency, and phase of a vibration signal. Changes in these parameters (for example, exceeding 4.5 mm/s rms value for rotating equipment in accordance with DSTU ISO 10816-1:2004) may indicate imbalance, misalignment, defects of bearings or gearboxes.
Thermography: Infrared cameras detect abnormal temperature regimes (for example, local heating up to 80°C instead of the standard 40°C), which may indicate overload, poor electrical contacts, friction or coolant leakage. Applicable according to ISO 18434-1.
Acoustic analysis and ultrasound: Microphones and ultrasonic sensors detect changes in noise or ultrasonic waves caused by air/gas leaks, electrical discharges, or internal equipment defects.
Oil analysis: Chemical and spectral analysis of oil samples reveals the presence of metal particles of wear, water, oxidation or degradation of additives, which are indicators of the condition of the components (for example, a change in viscosity of more than 10% from the original).
Artificial intelligence and machine learning (AI/ML): Collected data is processed by AI/ML algorithms (neural networks, decision trees, support vectors) for:
- Anomaly detection: Identification of deviations from normal behavior.
- Classification of faults: Recognizing the type and cause of potential failure.
- Remaining Lifetime (RUL) Prediction: Time to Expected Failure Estimate with an accuracy of +/- 10-15%.

Integration of these technologies, according to ISO 17359:2018 "Condition monitoring and diagnostics of machines. General guidelines", allows you to create a comprehensive monitoring system that ensures early detection of defects and optimization of maintenance intervals.

3. Current State of Development

VET technologies have reached high levels of readiness (TRL - Technology Readiness Level) and are actively being implemented in advanced industries.

TRL Level 7-9: Basic sensor technologies (vibration, temperature, pressure) and data acquisition platforms are commercially available and widely used. Such systems can provide 99.5% reliability of data collection.
TRL Level 5-7: AI/MI-based systems for RUL prediction and prescriptive recommendations are under active implementation. Prototypes of such systems demonstrate 85-90% prediction accuracy.
Key players: Global leaders such as Siemens (MindSphere), General Electric (Predix), Rockwell Automation and ABB offer integrated platforms for VET. Specialized Ukrainian companies are developing that adapt these technologies to local conditions.
IoT ecosystem: The widespread distribution of inexpensive IoT sensors, the growth of computing power on the periphery (Edge Computing) and cloud platforms (Cloud Analytics) ensure the scalability and availability of VET. For example, edge devices can process up to 100,000 measurements/second, reducing decision-making latency to milliseconds.

This progress allows integration of VET not only as a separate system, but as an integral part of the overall asset management strategy, in accordance with the requirements of ISO 55001:2014 "Asset Management. Management systems. Requirements".

4. Potential Impact on MRO and Economic Efficiency

The implementation of predictive and prescriptive maintenance has a direct and significant economic impact on MRO operations:

Decreasing unplanned downtime: According to McKinsey & Company, VET can reduce unplanned downtime by 25-50%. For businesses where the cost of an hour of downtime can reach $100,000, this means savings of millions of euros annually.
Reduced maintenance costs: Deloitte reports show an 18-25% reduction in total maintenance costs by moving from planned to actual repair needs. The rate of return on investment (ROI) can reach 400-500% within 2-3 years.
Extending the life of assets: Optimized maintenance allows you to extend the life cycle of production equipment by 20-40%, deferring significant capital expenditures for the purchase of new equipment.
Optimization of spare parts inventory management: Prescriptive approach allows UNITEC-D and other suppliers to forecast demand for spare parts with high accuracy (up to 95%), minimizing inventory (down to 30%) and avoiding shortages of critical components. This reduces storage costs and risks associated with the obsolescence of spare parts.
Increasing labor safety: Prevention of unexpected equipment failures significantly reduces risks for personnel working on production lines, which meets the requirements of DSTU EN ISO 12100:2016 "Safety of machines. General design principles. Risk assessment and risk reduction".

Moving to a prescriptive approach, where the system not only predicts the problem, but also suggests the optimal solution (for example, "Replace bearing #3 in pump A12 in 7 days, using spare bearing from warehouse 2, initiate order for new bearing"), provides maximum ROI.

5. Time Frames and Implementation Curve

The introduction of predictive maintenance in Ukrainian industry is a gradual process covering the period 2026-2035:

2026-2028: Pilot Projects and Criticality Assessment. At this stage, businesses focus on selecting the 10-20% most critical assets (generating 80% of downtime costs) for pilot projects. Initial investment in sensors and data analysis software. Expected ROI: 150-200% within the first 18 months.
2029-2032: Widespread Deployment and Integration. After successful pilots, the scaling of VET systems to other production lines and shops is taking place. The key is integration with existing ERP systems (as in the case of UNITEC's ERP-AI) and maintenance management systems (CMMS) to automate work processes. Estimated ROI: 300-400% by the end of the period.
2033-2035: Prescriptive Systems and Full Automation. Achieving maturity where systems not only predict failures, but also independently generate optimal repair strategies, spare parts ordering, and resource planning. Implementation of the principles of digital doubles for complex modeling of equipment behavior. It is possible to achieve ROI up to 500% and more.

The average annual growth rate (CAGR) of the VET market is projected at 17% until 2028, highlighting the global trend and economic attractiveness of the technology.

6. Challenges and Barriers

Despite significant advantages, VET implementation faces a number of challenges:

Initial investment: Sensor hardware, software and integration costs can be significant, especially for large productions. However, according to data from IoT Analytics, 27% of companies break even in less than 12 months.
Data quality and integration: Collection, storage and analysis of large amounts of heterogeneous data (from sensors, SCADA, ERP) requires reliable infrastructures and adherence to standards such as ISO 13374 (Data processing, exchange and presentation) and ISO 14224 (Collection and exchange of equipment reliability and maintenance data).
Cybersecurity: Connecting industrial equipment to networks creates new vectors for cyberattacks. The protection of VET data and systems is critical.
Skill shortage: The need for data engineers, AI/M&E specialists, and technicians trained to work with new technologies is high.
Organizational resistance: Changing established work processes and maintenance culture is often met with resistance from staff.

7. Recommendations for Chief Engineers

For the effective implementation of predictive maintenance, chief engineers of Ukrainian enterprises should take the following steps:

Perform an asset criticality analysis: Identify the equipment whose failure will have the greatest impact on production and safety. Focus pilot projects precisely on these assets, following the principles of ISO 17359.
Develop a data roadmap: Assess existing data infrastructure, identify data sources (sensors, SCADA, CMMS, ERP), integration and data storage requirements. Ensure compliance ISO 13374.
Initiate pilot projects: Start with small, manageable projects that will allow you to quickly demonstrate ROI and gain hands-on experience. Choose technologies that have CE certification and, if possible, UkrSEPRO.
Invest in personnel training and development: Provide training of engineers and technicians in data analysis, operation of new sensor systems and AI/MN algorithms.
Attract reliable suppliers: Cooperate with companies like UNITEC-D GmbH, who can provide not only quality spare parts, but also expertise in the selection of sensors, the implementation of components compatible with PTO systems, and the integration of these solutions into existing production processes. UNITEC-D ensures the supply of certified components for the entire range of industrial equipment.
Create a culture of continuous improvement: Encourage the sharing of knowledge and the constant search for new opportunities to optimize through data.

8. Conclusion

The transition from reactive to predictive and prescriptive maintenance is not just technological modernization, but a strategic necessity for Ukrainian industrial enterprises. This will allow not only to significantly increase operational efficiency and reduce costs, but also to strengthen competitive positions in the global market. Although the path is fraught with challenges, careful planning, use of proven scientific methods, and cooperation with reliable partners who provide quality and certified components make these benefits achievable. UNITEC-D GmbH is ready to be your partner in this transformational process, providing critical components and expert support.

Find out more about components and solutions for your production in the UNITEC-D E-Catalog.

9. List of Links

DSTU ISO 10816-1:2004 Vibration. Evaluation of machine vibration based on the results of measurements on non-rotating parts. Part 1. General guidelines.
DSTU EN ISO 12100:2016 Machine safety. General design principles. Risk assessment and risk reduction.
ISO 17359:2018 Condition monitoring and diagnostics of machines — General guidelines.
ISO 55001:2014 Asset management — Management systems — Requirements.
ISO 13374:2017 Condition monitoring and diagnostics of machines — Data processing, communication and presentation.
ISO 14224:2016 Petroleum, petrochemical and natural gas industries — Collection and exchange of reliability and maintenance data for equipment.
ISO 18434-1:2008 Condition monitoring and diagnostics of machines — Thermography — Part 1: General procedures.
McKinsey & Company Industry Reports on Predictive Maintenance ROI (various, e.g., "Industry 4.0: Reimagining production and assembly").
Deloitte Insights: "Predictive Maintenance: An IoT-enabled competitive edge".
IoT Analytics Reports on Predictive Maintenance Market (various).