Condition monitoring: Effective early warning system for machines

Detect machine failures before they happen? Condition monitoring makes it possible. With intelligent sensors and real-time data, you can keep an eye on the status of your systems at all times and intervene early if necessary. In this article, we show you how condition monitoring works, what advantages it offers and how you can use it successfully in your company.

Condition monitoring (CM) refers to the continuous monitoring of machine conditions using sensors. Condition monitoring enables you to detect, monitor and track incidents that lead to a deterioration in the condition of machines or their components as early as possible. This preventive, condition-based maintenance strategy gives you important insights into the health of your machines.

The central goal of condition monitoring is to optimise machine condition through monitoring. Effective condition monitoring enables you to intervene at an early stage and rectify faults before they lead to machine downtime or damage. It plays a crucial role in Industry 4.0. A further development of this approach is predictive maintenance, which uses data science to predict future maintenance requirements.

In short, the condition monitoring process involves recording important status parameters of machines and detecting changes that indicate impending faults. We will take a closer look at exactly how this works and which sensors and systems are relevant for this purpose below.

Sub-steps in the condition monitoring process Condition monitoring is based on a structured process that is divided into four central sub-steps:

In condition monitoring, existing sensors are used first, if possible, to monitor the machine status. If necessary, additional sensors are installed on the systems to collect relevant measurement data. Depending on the type of machine, different condition monitoring sensors are used:

• Vibration sensors: Monitor vibrations on rotating machines and thus detect wear on axles and bearings at an early stage.
• Pressure sensors: Measure the pressure in hydraulic and pneumatic systems and help to identify anomalies such as leaks.
• Ultrasonic sensors: Measure high-frequency signals and locate friction or malfunctions on moving parts of the machine.
• Temperature and infrared sensors: Analyse heat development on machine parts, for example to detect overheating in electrical components using thermal imaging cameras.
• Acoustic sensors: Detect changes in operating noise. In a complex machine, for example, the sound emitted by the machine changes depending on the component defect.
• Oil sensors: Examine oil quality by measuring particles, moisture or viscosity.
• Current and voltage sensors: Monitor electrical currents and voltages to detect faults or overloads in electrical systems in good time.
• Position and displacement sensors: Measure movements and positions of machine parts to detect deviations or misalignments.
• Emission sensors: Monitor exhaust gases and pollutant emissions from machines and systems to ensure compliance with environmental regulations.

The exact parameters that are monitored depend heavily on the machine in question. Frequently monitored variables include vibrations, noise, temperatures and pressure. These environmental and system variables help to analyse the condition of the machines accurately and identify problems at an early stage.

A complex condition monitoring system usually consists of the following components:

• Sensors: Sensors are either wired or wireless and connected to the systems. They collect critical operating data that is necessary for evaluating the condition of the system or machine.
• Data transmission: The sensor data collected can be transmitted either to a system directly on the machine or to a central system. For individual systems, a machine-based system is often sufficient, while a central system is more appropriate for monitoring several machines. The frequency with which the sensor data is transmitted varies depending on the operating mode. Some systems enable alarm-controlled transmission.
• Data analysis: A specialised condition monitoring algorithm evaluates the collected data and identifies trends and deviations. Many modern condition monitoring systems offer the option of storing and analysing the data in the cloud. The analysis enables potential anomalies to be detected at an early stage.
• Warning systems: If the system detects irregularities, warning or error messages are automatically generated. These notifications enable maintenance teams to respond immediately.

In addition to all the advantages and opportunities, the disadvantages or challenges of condition monitoring should not be overlooked. Condition monitoring is generally suitable for all companies that operate plants and want to monitor their condition. However, implementation is often complex and requires careful planning and adaptation to the specific conditions of the plant. Depending on the industry, company size and available resources, you will face different hurdles in the condition monitoring process that need to be overcome.

A key challenge is determining which machines and plant components should be monitored. Not every machine is equally important for production. Therefore, start by asking yourself the following questions:

• What are our critical assets?
• Which measurement parameters are meaningful and necessary to achieve my goals?
• What expectations of machine reliability are reasonable and practical?

To identify the critical components and measurement data, it is important to understand the failure modes, failure rate and detectability of the failure. Proven methods such as the following can help you with this:

• Reliability-Centred Maintenance (RCM): RCM is a structured process that examines the functions of a system and determines which failures are most critical. It analyses how these failures occur, what consequences they have and which preventive maintenance strategies are most appropriate. Maintenance is prioritised based on actual risks and needs.
• Failure Mode and Effects Analysis (FMEA): FMEA identifies possible failure modes of a system and evaluates how likely they are to occur, what effects they have and how serious they are for the overall system. Based on this analysis, a risk priority number (RPN) is calculated. This helps to define targeted measures to prevent these errors or minimise their effects.

In addition to collecting data, another often underestimated problem is how to use it effectively. It is not enough to identify anomalies or potential errors – staff must be able to interpret the data and know what measures to take. Condition monitoring itself does not solve problems; it only identifies them – the cause of the error must still be rectified. When an error message is issued, it is the responsibility of the staff to respond appropriately and as quickly as possible. A successful system therefore requires not only good technology, but also training for staff in how to handle the data. Ask yourself the question: ‘How can I ensure that my employees understand the data and know what to do?’

Choosing the right technology and condition monitoring tools plays a key role in the success and efficiency of condition monitoring. Ensure that the sensors and networking solutions you use meet the specific requirements of your machines. In addition, integrating the system into existing IT structures often presents a technical hurdle, especially in companies with older machines.

Another aspect that is particularly important for smaller companies is cost. Depending on the scope and complexity of the monitoring, the purchase and installation of a condition monitoring system can often be very expensive. In addition to the costs for sensors and software, the effort required to integrate the system into existing machines must also be taken into account. It becomes particularly expensive if machines do not meet the requirements for condition monitoring and you have to retrofit them first. Carry out a careful cost-benefit analysis to determine whether and to what extent condition monitoring makes economic sense for your company.

To implement condition monitoring effectively in your company, you should follow a clearly structured process:

The use of AI in industry and AI in production is no longer a rarity. There are also promising AI use cases in condition monitoring. Here, AI offers one thing above all else: significant time savings. AI is able to evaluate large amounts of data in a very short time and identify patterns and potential problems that are often not immediately apparent to humans. AI's ability to identify subtle signs of irregularities at an early stage is also particularly valuable. This helps to detect and rectify errors before they lead to major problems. However, AI as a condition monitoring tool not only reveals problems, but also provides specific recommendations for action.

Condition monitoring is primarily used in industry to continuously monitor machines and systems. Typical areas of application include rotating machines, bearings, gearboxes, pumps and compressors. Condition monitoring helps to detect wear at an early stage and avoid unplanned downtime, especially in areas such as manufacturing, power generation and process engineering.

Condition monitoring in bearings and gearboxes

In a manufacturing plant for automotive components, condition monitoring is used on the bearings and gearboxes of conveyor belts. Sensors continuously monitor the vibrations and temperature of the bearings. After a few weeks, the system indicates an increase in vibrations in a specific gearbox, which is due to contamination of the lubricant. In addition, a slight eccentricity is detected, indicating that the bearing is not properly centred due to incorrect installation.

This data enables the maintenance team to act quickly: the lubricant is replaced and the bearing reinstalled before it is permanently damaged by the increased loads. This early intervention not only prevents expensive machine downtime, but also significantly extends the service life of the affected components.

Condition monitoring in vibration measurement

In a metalworking factory, condition monitoring is used during the milling of precision parts. Special sensors monitor the vibrations of the milling machine in real time. During a manufacturing process, the system detects an unusually strong vibration, which indicates a blunt milling head or incorrect machining settings.

Detailed vibration measurement enables maintenance personnel to intervene immediately, replace the milling head and adjust the settings. This not only prevents poor surface quality of the machined parts, but also protects the machine from further damage caused by uncontrolled vibrations. Condition monitoring helps to maintain product quality and avoid expensive rework.

Condition monitoring offers companies a powerful method for monitoring machines more efficiently and optimising their maintenance. Using modern sensors and analysis tools, problems can be identified at an early stage and targeted maintenance measures can be initiated before costly breakdowns occur.

Despite some challenges, such as selecting the right measuring devices and integrating them into existing systems, condition monitoring is an essential tool for maximising the service life of machines, reducing costs and increasing safety. With the right implementation, companies benefit from improved efficiency and reliability in the long term thanks to condition monitoring.