The largest source of production losses in a manufacturing plant is due to unplanned downtime. During downtime, the machines are in an idle state, and they do not produce any value; but, the overhead costs continue to grow. Hence, an unplanned downtime incurs a large expense for a business and directly impacts its bottom line. According to Asset Performance Management: Blazing a Better Path to Operational Excellence, 2017, it is estimated that unplanned downtime can cost $ 10,000 to $ 250,000 per hour for industrial plants. Recent studies have also shown that industrial manufacturers spend around $ 50 billion each year due to unexpected asset repairs and failures (Forbes Technology Council).

Types of maintenance

Based on the asset type, maintenance goals, available budget and human resources, organizations use different maintenance strategies to increase the uptime of their manufacturing plants. In general, there are two main classes of maintenance strategies as reactive maintenance and proactive maintenance. In reactive maintenance, repair or replacement activities take place only when the asset has already broken down.

Reactive maintenance is usually expensive as machine failures often occur during peak working hours. Also, companies must rely on high-priced external services instead of in-house maintenance staff, to bring equipment back to working condition quickly. However, reactive maintenance sometimes can be cost-effective for low-value/low-priority assets which can be easily replaced rather than serviced.

In contrast, proactive maintenance focuses on taking steps to prevent and resolve component malfunctions before they occur. These maintenance practices are designed to monitor the behaviour of an asset, identify the underlying causes of failure and take proactive actions to minimize or eliminate those factors. Proactive maintenance strategies bring a significant ROIs, especially for high-value equipment as it minimizes equipment malfunctions, and unplanned downtimes and prevents consequential damages to assets. Preventive maintenance and condition-based maintenance are some popular proactive maintenance strategies.

Preventive maintenance

Preventive maintenance is about serving assets recurrently, whether the upkeep is necessary or not. In other words, it is a collection of best practices performed at a regular time interval, giving you the best opportunity to fix issues before they even start.

Following are the characteristics of preventive maintenance methods:

• Collection of planned activities
• Requires machine downtime to carry out
• Consists of a checklist of planned activities (Inspection, Cleaning, Calibration, Repair and Replacement)
• Takes place at a regular interval (Daily, Monthly, every 3-6 months, Yearly)
• Performs different tasks at different time intervals
• Takes place despite identifiable problems

In preventive maintenance, it does not consider the actual condition of the components/assets when performing maintenance activities. It can be inconvenient and can reduce the overall capacity of the manufacturing plant due to regularly planned downtimes. However, manufacturing companies prefer preventive maintenance over reactive maintenance, as it is highly effective in increasing the equipment lifespan and can prevent costly equipment failures.

Condition-based maintenance

Condition-based maintenance is a proactive maintenance strategy, and it happens only when the condition of the asset reaches unacceptable levels. Hence, it relies on real-time non-invasive sensor measurements, visual inspections and scheduled tests to get an understanding of the asset health. Compared to preventive maintenance, condition-based maintenance activities are performed on an as-needed basis.

Below are the characteristics of condition-based maintenance methods.

• Relies on real-time sensor readings
• Occurs only when a decline in asset status/condition is observed
• Is done only on components/assets that require maintenance
• Occurs while the asset is in working condition
• Low planned downtimes

The New Norm: Predictive maintenance

Predictive maintenance is a technique that detects anomalies in the equipment and forecasts future maintenance events by analysing the machine operation data. It relies on real-time sensor measurements to determine the condition of in-service equipment. It will accurately estimate and schedule future maintenance activities based on the collected data and predefined algorithms.

Characteristics of predictive maintenance methodologies are as follows.

• Predicts future breakdown events
• Is performed based on the analysis of collected machine data
• Occurs during machine operation. If downtime is required, it will be shorter and more targeted
• Addresses an actual problem

As the machine condition is detected and maintained in advance before a failure happens, predictive maintenance techniques reduce production hours loss due to unplanned downtime. The cost of predictive maintenance is comparatively low because it extends the life span of the assets, and the parts are only replaced when it is completely necessary. It also increases workplace safety by eliminating manual error and risks of injury.

According to the latest market research report from Grand View Research, the global predictive maintenance market is forecasted to be $28.24 billion by 2025. Evolving technologies, such as the Internet of Things (IoT) and Big Data analytics, drive the growth of the predictive maintenance market.

How does it work?

Predictive maintenance can be divided into two phases as data collection and data analysis.

1. Asset Monitoring and Data Collection

It requires quality data on the performance and health of the asset to predict when a breakdown will occur. The first step is to install appropriate sensors based on the parameters that you intend to measure and monitor in real-time. In general, parameters such as temperature, pressure, noise level and vibration are measured using IoT technologies and the collected data is sent to a central location for the analysis.

2. Machine Learning and Prediction

The next step is to build an analytical model to obtain actionable and meaningful insights about the state of the asset. Collected sensor data is fed to different models and algorithms to detect anomalies and make a prognosis for future machine failures. In other words, they extract information to identify signs of deterioration that can interrupt production and calculate when you need to do maintenance.

Use cases and applications

Vibration analysis, temperature/thermographic analysis, acoustic analysis, and motor circuit analysis are among the most popular predictive maintenance techniques used in the manufacturing industry. These analysis techniques are applied to detect short-circuits, engine degradations, rusting, disconnections, misalignments and imbalances, delamination, wear and tear of various heating and rotating equipment that might go unnoticed to the naked eye.

Predictive maintenance is not limited to the Manufacturing sector (chemicals, automotive, aerospace, food and beverage) and these techniques apply to other industry verticals such as energy and utilities, healthcare, aviation and railways, hospitality, IT, telecommunications, and many more.

Pros vs. cons

The below diagram illustrates the advantages and disadvantages of the above-discussed maintenance strategies.

Maintenance strategy and processes are core elements for any successful manufacturing company. According to Deloitte, poor maintenance strategies can reduce a plant’s overall productive capacity by 5% to 20%. Therefore, it is important to choose the appropriate maintenance strategy based on the value and urgency of the asset. Considering the above maintenance strategies, it is best if an organization can predict its future maintenance events in advance and obtain recommendations to prevent a breakdown.

Predictive maintenance breaks trade-offs of traditional maintenances techniques such as frequent disruptions to operations, catastrophic machine damage, risk of excess and obsolescence of spare parts. It also empowers companies to maximize the lifespan of their assets while avoiding unplanned downtime and unnecessary planned downtime. With the advancement of technology, it is time for companies to say goodbye to reactive or preventive maintenance and welcome predictive maintenance.

 – Written by Inuri Muthukumarana