The concept of the condition monitoring ( monitoring state ) is based on a periodic or continuous detection of the machine state by measurement and analysis of physical quantities, such as Vibrations, temperatures, Location / approximation.


Condition monitoring is twofold: security and machine efficiency. It is similar to structural health monitoring in static components.

Based on the given case analyzed in real-time sensor data, a reliable and very fast-acting safety system ( emergency shutdown, engl. "Trip" ) can be realized. In comparison, previous systems (eg simple vibration sensors ( "Earthquake " switches) ) generally imprecise, and then do not contribute to the elucidation of the causes of damage. Online condition monitoring (continuous condition monitoring ) allows an emergency shutdown based on the analyzed and stored data - and thus a subsequent analysis of the interference factor.

The machine health monitoring is the mandatory requirement for condition-based maintenance. This strategy replaces the previously conventional reactive or preventive maintenance. The latter were at fixed time intervals shut down the machine in question and checked or replaced components. This type of machine maintenance often led to defective parts replaced and residual terms were thus wasted.

Modern CM systems make high demands on sensors, data acquisition, forwarding, and automatic data processing (analysis, diagnosis) as well as plant-specific skills. However, it also offers the greatest potential for cost savings because the life of critical machine elements can be almost completely exploited, while necessary repair work can be scheduled in coordination with the production plan.

The condition-based maintenance as a cross-section tray from the fields of mechanics, acoustics, systems theory, electronics and computer science is located in a development. In particular, in the monitoring of individual components can, however, be very accurate today. In complex systems, it is becoming increasingly blurred, as overlaid with increasing system complexity, a growing number of signals of different origins. Here, pure expert systems prove to be the only professional solution for monitoring critical machinery. These specially designed for a machine type systems offer - depending on the configuration - maximum protection for people, the environment and the machine and the maximum utilization of component life cycles.

Another shortcoming was far too often the lack of suitable sensors to record signals directly into the wear or damage zones can. Here is capable of in the future perhaps, microsystems technology to remedy the situation, eg by sensors in thin-film technology, which can be mounted directly on the structure to be monitored.

The challenges of this strategy can be seen in:

  • Finding suitable measurement points and sensors,
  • Finding meaningful parameters ( state variables ) of damage to the components of interest,
  • The targeted application of signal analysis and pattern recognition,
  • As well as the enormous amount of data.

Or to sum it up in one sentence: " What needs when, where, how and what to monitor? "

What condition monitoring can NOT:

  • Recognising and avoiding spontaneous failures such as fatigue fracture of a wave

In this context it should be noted that rapid shutdown systems help to avoid costly damages of spontaneous failures. This means that the machine is scanned within a few milliseconds after the damage. Experience shows that the consequences of the Continue defective machines are usually more extensive than the actual initial defect. In some cases, however, it is not sensible to make scrams since the resulting disconnected systems or technical processes involve high risks by subsequent processes. Here it is necessary rather urgently, all the data from the monitoring systems prepared to report to the responsible operating personnel. Thereafter, targeted and coordinated as necessary measures for controlled shutdown must be initiated in order to prevent greater damage of the subsequent processes. Appropriate contingency plans or operating instructions for such cases must be available.

Some steps in the condition monitoring

Condition monitoring is composed of several sub-steps:

First state detection: The state detection is the measurement and documentation of the machine parameters, the current status of the production means ( or of the machining process ) reflected.

2 state comparison: The state comparison shows the comparison of the actual state with a predetermined reference value dar. This reference value can be both a target value to be complied with as well as a not -border limit. The nominal value is determined depending on the parameters surveyed, either at the machine acceptance or defined by predetermined sizes. Limits are usually determined by the manufacturer or user of the machine empirically.

State monitoring and state comparison substantially conform to the inspection according to DIN 31051st

3 Diagnosis: It is the task of diagnosis to locate the basis of the results of the comparison condition for potential errors as early as possible and their cause (s) to determine in order to plan the necessary maintenance measures can betimes.

Inspection sequence

Condition monitoring systems can be classified according to the inspection sequence. The inspections may be made either intermittently or continuously.

Intermittent monitoring may occur at regular or variable intervals. This can naturally capture only the inspection times state information. Long-term trends are therefore being determined, or occurring at short notice, however, transient events can not be detected. The inspection intervals are either preset by the manufacturer or must be determined by their own tests / experiences. One advantage of the intermittent monitoring is the possibility of using mobile measuring instruments which, compared with the full metrological instrumentation of all the monitored machines naturally brings savings.

Continuous (permanent ) monitoring systems detect the machine parameters continuously and in real time. Thus both long-term trends as well as abrupt or transient state changes are recorded and documented. The cost of such systems is - especially from the data management here - much higher than in intermittent systems. This additional expense is justified only when high demands are placed on the reliability of the monitored system, eg steam turbines and generators in power plants. In the monitoring of treatment processes, such as the tool breakage monitoring, continuous systems are also often unavoidable.

About the diagnostic skills of a surveillance system the subdivision, however, says in intermittently and continuously from nothing.

Process control and machine monitoring

In the condition monitoring is to distinguish between process control and machine monitoring. The process monitoring aims thus based on the quality of the machining process; important application is the monitoring tool during machining; whereas the machine monitoring is to protect the machine and its components to the target.

Basically, the strategies and tools are similar in both monitoring fields. Often the same sensors and signal processing mechanisms can be used. The significant difference, however, that in the machine monitoring the behavior of the machine structure is the source - as the transmission link between the process signal and the sensor, but is a problem in the process control. If property changes on the machine structure, eg due to wear of components, so this represents for machine monitoring one to be detected event dar. For process monitoring, however, this change of the transmission path leads to a signal change, which does not originate from the process.

Further, in the process control signals from the process-induced excitation of the machine tool system can be detected. For the machine monitoring the process signals are, however, regarded due to their dominance as a disturbance, so here must be a suggestion of the machine structure for system identification outside of the machining process.