EP4153866A1

EP4153866A1 - Method for monitoring the operation of a fan, apparatus and fan

Info

Publication number: EP4153866A1
Application number: EP22718855.4A
Authority: EP
Inventors: Bjoern WENGER; Jacob KRAUTH; Fabian MOELLER
Original assignee: Ziehl Abegg SE
Current assignee: Ziehl Abegg SE
Priority date: 2021-04-16
Filing date: 2022-03-23
Publication date: 2023-03-29
Also published as: WO2022218482A1; CN117178122A; BR112023019002A2; DE102021203806A1; JP2024517383A

Abstract

The invention relates to a method for monitoring the operation of a fan, the method including the following steps: performing at least one measurement (1) by detecting at least one input signal (2.1, 8.1) over at least one time period, calculating parameters (2, 8) of an actual state on the basis of the measured input signals (2.1, 8.1), comparing (3, 9) the calculated parameters (2, 8) of at least one actual state with limit values (3.1, 9.1), classifying (4) the actual state on the basis of the comparison (3), either as a normal state (5) without anomalies or as a general anomalous state (6), and monitoring whether at least one specific anomalous state (12) is present.

Description

PROCEDURE FOR MONITORING THE OPERATION OF A FAN, DEVICE AND FAN

The invention relates to a method for monitoring the operation of a ventilator, a device for monitoring the operation of a fan and a fan with such a device.

Methods and devices for monitoring the operation of fans of the type in question have been known in practice for years. For example, it is known to carry out vibration analyzes on rotating machines for early detection of damage, for example using acceleration sensors or similar vibration sensors. It is also known to identify certain damage regulations. These are specific multiples of the rotational frequency of the rotating machine. Compared to the rest of the frequency spectrum of the vibration analysis, a conspicuous amplitude in the area of a specific damage order indicates a specific anomaly in the operation of the mechanical machine, for example bearing damage. Certain types of damage - and the associated bearing damage frequencies - can be assigned to individual bearing components, for example an outer ring, an inner ring, a rolling element or a cage of the respective rolling bearing. If a conspicuous amplitude is detected in the frequency spectrum as part of the vibration analysis, an automatic emergency shutdown is triggered regularly. In the best case, the defective component of the fan that caused the anomaly can then be repaired or replaced on the basis of the collected data, without a fatal event occurring, for example a system failure that causes high financial damage and/or a risk to people through consequential damage can represent people and the environment. Despite the downtimes when an anomaly is detected, the overall lifespan of the fan is extended.

The present invention is therefore based on the object of designing and developing a method and a device for monitoring the operation of fans of the type mentioned in such a way that the operation via the The service life of the fan can be further optimized and unnecessary downtimes can be reduced. Furthermore, an improved fan with a correspondingly equipped device is to be specified.

According to the invention, the above object is achieved by the features of claim 1. According to that, the method in question for monitoring the operation of a fan comprises the following steps:

performing at least one measurement by acquiring at least one input signal over at least one period of time,

Calculation of characteristic values of an actual state based on the measured input signals,

Comparison of the calculated characteristic values of at least one actual state with limit values,

classifying the current state based on the comparison either as an anomaly-free normal state or as a general anomaly state, and

Monitor for at least one specific anomaly condition.

In a manner according to the invention, it was initially recognized that immediate intervention is not absolutely necessary, even if an anomaly state is detected. Even if individual characteristic values change over the course of the fan's service life in such a way that at some point they lie outside a permissible value range defined by the limit values, this does not necessarily mean that a damaging event is imminent and must be averted in the near future. This reduces unnecessary emergency shutdowns and downtimes and further optimizes operation over the life of the fan.

In addition to avoiding unnecessary downtimes, monitoring can also be possible in particular when the fan is at a standstill and/or in a voltage-free state. The operation to be monitored therefore includes the entire use of the fan over the product service life and also periods of standstill. This not only extends the operation from a financial point of view, but also allows real-time monitoring of the system in the sense of passive observation. This can be linked to recommendations for action be in the sense of an active action, which can, for example, trigger automatic control actions or be directed at a user. This also reduces the application risk that can arise from a system failure.

The acquisition of the at least one input signal over a period of time can be done continuously or at defined points in time. A sensor can capture time data over the at least one time period. In the case of an acceleration sensor, the data can be a yaw rate and/or an acceleration in one, two or three spatial axes. The spatial axes can be defined as non-rotatable or circumferential. The acceleration sensor can be assigned to a rotating component or to a stationary component of the fan.

The measurements can be evaluated in order to calculate the characteristic values via a computing unit of the fan, in particular a motor driving the fan. The motor is preferably an electric motor, in particular a brushless DC motor designed as an external rotor motor—EC motor, electronically commutated motor.

According to a further embodiment, data relating to the measurement can be processed via defined interfaces (l ² C, Modbus, CAN, Wifi, ...) for further signal processing and/or signal utilization and/or signal storage--at certain times or linked to events other end devices are transmitted. Selected measurements associated with specific points in time can be backed up to internal memory on the device itself or external memories. This data can be used, for example, for reclamation purposes, in which case a service employee can preferably access the memory. Parameter data, comparison data, data relating to the classification of the actual state and/or other data relating to monitoring can also be transmitted to other terminals for further signal processing and/or signal utilization. In this way, monitoring of the operation of the fan can also be tracked via an app, for example on a smartphone or tablet. Furthermore, the data is optically processed and displayed to an operator on external screens/displays.

Characteristic values can be calculated from one or more measurements, which reflect the current actual status. If the input signals are acceleration or sound pressure data, amplitudes of the frequency spectrum, for example, can be determined as characteristic values. Especially when diagnosing an anomaly in a roller bearing, this includes bearing damage frequencies, which can depend on the one hand on the speed and load and on the other hand on the geometry of the roller bearing. Characteristic values can, for example, depend on the application, i.e. on specific use cases/areas of use/installations, but also on the respective environmental conditions/

environmental influences. Parameters can also be dependent, for example, on the applied rotor speed/rotational rate and/or on a load condition. Bearing geometry information can also be included in the calculation of the characteristic values.

The limit values required for the comparison can be determined in advance or in real time in experimental or numerical studies and limit the maximum permissible value range of the characteristic values. One or more upper threshold values and/or one or more lower threshold values can thus be defined for each input signal and/or each characteristic value in order to delimit the permissible range of values. A self-configuration of a classifier for classifying the states is also possible.

The comparison of calculated characteristic values with limit values allows classification into two or more states. The comparison can be made over several measurements in order to increase the robustness of the method with regard to environmental conditions/environmental influences. In the case of determined acceleration data, these interference signals can be external excitations, for example due to vibrating or moving bodies, different location factors or the absolute or relative speed of a moving application - for example when operating the fan in an airplane or a train , a motor vehicle or another means of transport. The detection of an anomaly condition can be falsely caused by temporary environmental factors or environmental conditions. For the comparison, the number of all limit deviations - ie the comparisons in which a characteristic value was recorded above an upper threshold value or below a lower threshold value for the respective limit value - can therefore also be considered per period. Furthermore, the intensity of a limit value deviation can be taken into account by considering the relative and/or absolute differences between the limit value and the characteristic value. A rate of change over time of a characteristic value profile can also be taken into account. If, for example, there are strong fluctuations or rapid changes - also known as high gradients - within one or more defined periods of time, the deviation from the limit value can be weighted accordingly and taken into account when classifying the current state.

The actual state can be classified as an anomaly-free normal state if regular operation of the fan and/or the EC motor and/or the application is diagnosed. This includes in particular natural wear and tear. The current status can be assessed iteratively by comparing characteristic values with limit values.

The actual state can be classified as a general anomaly state if a significant malfunction of the fan, its EC motor or the application is diagnosed. This condition can be attributed to increased wear, for example. The significant fault can be, for example, bearing damage, contamination of the lubricant, foreign objects in the roller bearing, or in the form of erosion of the roller bodies, e.g. due to bearing currents, an imbalance - e.g minor material failure, a defect or partial defect of electronic components and/or other irregular influences. However, the primary functions of the system are not necessarily impaired in this case. The metrics may include one or more of the following:

Results of one or more previous comparisons, a rotation rate of the fan, an acceleration in one, two or three spatial axes, non-rotatable and/or rotating, at least one temperature, a sound pressure, a torque, a pressure, in particular an operating or ambient pressure,

Moisture values, measured forces and/or virtual values, recorded for example using soft sensors.

The entirety of all data recorded by one or more sensors in a period of time is the result of a measurement. The temperature can be an ambient temperature. Additionally or alternatively, it can be an operating temperature of the fan, in particular a temperature on a specific component, namely on one or more transistors, capacitors, heat exchangers, coolants, computing units, resistors, coils, lubricants, mechanical components such as bearings , shafts, permanent magnets.

Further characteristic values can preferably be calculated from one or more characteristic values. Other parameters can be:

Statistical variables, such as minimum and maximum values,

percentile values,

standard deviations,

mean values,

Classifying/evaluating variables, for example using a points system similar to a ranking as for energy efficiency classes, and/or combinations, for example a linear combination of several parameters, with uniform or different weighting factors. In this way, the values of the input signals determined in one or more previous measurements can be included in the calculation of individual characteristic values in a suitable manner, depending on the application.

On the one hand, combinations of the above variables can be defined and summarized in individual characteristic values, for example the circulating accelerations in one, two or three spatial axes, and the rotation rate of the fan, the temperature and humidity values or other suitable combinations.

On the other hand, new characteristic values can be generated from current and earlier characteristic values. This is particularly useful if conclusions about the current status or future status of the fan appear likely. For example, a characteristic value in the form of a counter can also be formed to classify the actual state. For this purpose, a number of previously defined limit value deviations is counted as a characteristic value in a counter over time. The counter determines the limit value deviations either continuously, i.e. over all previous measurements, or also for measurements within defined periods of time, in other words for the number of all limit value deviations over a defined period of time, for example for the measurements of the last 30 minutes.

The corresponding limit value or tolerance value for each characteristic value configured as a counter can be configured relative. For example, the characteristic value designed as a counter can lie outside the tolerance values if 10% deviations from the limit value have been counted over a defined period of time or over all measurements. A characteristic value functioning as a counter can also be absolute. For example, the counter can be outside the tolerance values if 20 limit deviations have been counted in the last 50 measurements or in all measurements.

In addition, the counter can also take into account current or earlier comparisons in which the respective characteristic values are within the limit values and no deviation from the limit value is determined. Then, for example, the counter can to be reduced or counted down. At the next limit deviation, the counter can be increased or vice versa. In other words, the counter can be incremented and/or decremented.

In addition to the frequency of limit value deviations or non-limit value deviations, the intensity of a limit value deviation for the increase or decrease of the counter can additionally or alternatively be taken into account. For example, a particularly severe limit deviation by a factor of 2 can lead to an increase in the counter by a value of 2, while a less intense limit deviation by a factor of 1.1 can only lead to an increase in the counter by a value of 1.1.

The characteristic value calculated in this way can then be compared with a limit value designed as a tolerance value. The classification of the actual state can not only be done by directly comparing current characteristic values from current input signals of the actual state with limit values, but also by comparing a counter with a tolerance value. In other words, with an indirect comparison of current and earlier characteristic values and limit values, a quota of limit value violations can be determined, which in turn can be a characteristic value.

A combination of direct and indirect comparison is also conceivable when considering several comparisons between characteristic values, which are assigned to the actual state, and limit values. The characteristic value designed as a counter can increase by the value 2/3, for example, if two of three limit values are exceeded by the other associated characteristic values.

According to one embodiment, monitoring whether a specific anomaly condition exists includes the following steps:

Calculation of specific characteristic values of at least one state based on the measured input signals, specific comparison of the calculated specific characteristic values with specific limit values, Assess whether there is a specific anomaly condition based on the specific comparison.

This further monitoring can also be carried out continuously or at defined points in time over defined periods of time with a defined sampling rate. The specific characteristic values that are decisive for assessing whether a specific anomaly state is present can be calculated from one or more measurements, preferably in the engine's computing unit. The specific characteristic values can differ from the general characteristic values for classifying the actual state into the anomaly-free state or the general anomaly state, or they can be partially or completely identical to these general characteristic values. The specific characteristic values can also depend on the application, i.e. on specific use cases/areas of use/installations, but also on the respective ambient conditions/environmental influences. The specific characteristic values can also be dependent on the applied rotor speed/rotational rate and/or on a load condition. Bearing geometry information can also be included in the calculation of the specific characteristic values, for example when monitoring specific anomaly states that can be assigned to a rolling bearing. In addition to the variants and design options for assessing whether a specific anomaly condition exists - or whether a specific anomaly condition does not exist - the described variants and design options for classifying the current condition either as an anomaly-free normal condition or as a general anomaly condition can also be referred to to get expelled.

Not every general anomaly condition is critical. Also, not every specific anomaly state is critical. A critical anomaly condition is the extreme case of a specific anomaly condition. An anomaly state that is not associated with urgent, immediate recommendations for action is therefore a non-critical anomaly state. However, timely maintenance can be recommended. On the other hand, there may also be an increased significant disturbance and require an emergency shutdown because a fatal event is imminent. A critical anomaly state can then be assumed. A critical anomaly can exist if an increased significant malfunction of the fan, its EC motor or the application is diagnosed. This condition can be due to a very high level of wear, for example. There can also be pronounced, irregular behavior, for example advanced bearing damage, a significant imbalance due to enormous contamination, severe damage to a rotor blade, advanced material failure, a significant defect or partial defect in electronic components or other increased irregular behavior A rivers act. Prompt intervention is regularly required here, because the system behavior is system-critical and can be dangerous for the application, living beings and/or the environment. A system - this can be a fan motor, its controller or the signal-processing peripherals - can intervene autonomously, for example by switching off the motor or by changing the operating point, for example in the case of high heat loads. However, intervention can also mean human action.

According to one embodiment, the described monitoring as to whether a specific anomaly state is present is carried out following the classification of an actual state as a general anomaly state. For example, classifying a current actual state as an anomaly state can be a condition for monitoring whether the current actual state is a specific anomaly state being carried out. The correct assessment of the presence of a general, non-specific abnormal condition or a specific abnormal condition can be over 95%. Preferably, a second monitoring of whether a second specific abnormal condition is present is performed subsequent to the judgment that a first specific abnormal condition is present. More preferably, a third monitoring as to whether a third specific anomaly state is present is only carried out if there is an assessment that initially a second specific anomaly state is present, etc. Several specific anomaly states can thus build on one another, in particular with regard to the severity of their severity. Each specific anomaly state - just like the distinction between anomaly free normal state and general anomaly state - represents a binary state classifier. Such a series connection of several diagnoses of a general and then an increasingly specific anomaly in several stages, which are preferably linked to one another by conditions, expands the number of classification states.

In the basic form with two stages - classifying as an anomaly-free normal state and general abnormal state and judging whether a specific abnormal state is present or not - the following three positive classifications are possible:

Anomaly-free normal state, general anomaly state and specific anomaly state

Two binary classifications connected in series thus allow a trinary classification. Furthermore, the negative classification can be determined, according to which a general anomaly state exists, but a certain specific anomaly state does not exist.

The following condition can preferably be formulated for the series connection: In order for the next classifier stage—ie the next more pronounced specific anomaly state—to be able to be reached, the previous stage must have identified an anomaly state, so that a cascade is created. This applies in particular to the highest level of the specific anomaly state - this is the special case of the critical anomaly.

More preferably, one or more of the other specific anomaly states can form intermediate stages between the general anomaly state and a critical anomaly state, in which there is a significant fault in the fan that requires immediate intervention. In this case, several intermediate states can be associated with different degrees of manifestation of a specific damage pattern or determinable damage patterns. At most, for example, a first intermediate stage can be an anomaly with a low level, a second intermediate level corresponding to an anomaly state of high magnitude and a third intermediate level corresponding to an anomaly state of dangerous magnitude before a critical anomaly state of catastrophic magnitude is reached.

Each of these anomaly states or each of these intermediate stages can be classified in binary. Here, too, the monitoring can be carried out independently for each anomaly state or each intermediate stage or depending on the fact that the presence of a more general or less pronounced anomaly state has been determined beforehand. Each anomaly state or each intermediate stage forms a binary state classifier. The characteristic values and limit values for each anomaly state or each intermediate stage can be defined differently or overlap in parts. The use of different counters and tolerance values is also possible within the framework of the various diagnoses.

According to another embodiment, the monitoring of whether a specific anomaly state is present can be independent of an actual state being classified as a general anomaly state on the first level. There is thus the possibility of diagnosing a specific anomaly independently of a classification of the current or a previous actual state as a general anomaly state—e.g. for monitoring natural wear and tear. Irrespective of continuous monitoring of the operation of the fan and independent of how the characteristic and limit values are adjusted automatically or manually after an anomaly has been determined for the first time, a prompt system failure can be detected over the entire service life of the fan. For example, if immediate action is required, a critical anomaly state is immediately identified, which further improves the diagnosis of possible anomalies in the operation of the fan and further increases operational reliability. It is also conceivable to carry out the monitoring for a specific anomaly state both when an actual state is classified as an anomaly state and independently of one such a classification - for example at regular times or when another internal or external condition occurs.

In this case, further specific anomaly states with different characteristics can preferably be defined. In this way, gradual damage or aging of the fan and/or its motor can be mapped and monitored. For example, a first specific anomaly state can be monitored based on a first specific comparison of a first combination of first specific characteristic values with first specific limit values. A second specific anomaly state can be monitored by a second specific comparison of a second combination of second specific characteristic values with second specific limit values. A third specific anomaly condition is monitored by a third specific comparison of a third combination of third specific characteristics with third specific limit values, etc.

The parallel connection of several binary classifiers by independent monitoring of the various anomaly states allows an independent, unconditional binary classification of a state.

One or more binary classifiers can be activated and deactivated independently of one another. For example, this can be illustrated as follows: a first level - anomaly with a low degree - is inactive, a second level - anomaly with a strong degree - is inactive, a third level - anomaly with a dangerous degree - is active, a fourth level - more critical Catastrophic anomaly condition - is active.

An advantage of this can be that only anomalies that are assigned to different types of damage or damage patterns are monitored, for example in a roller bearing a specific anomaly assigned to the roller bearing grease, specific anomalies of rolling elements, their inner ring, outer ring, etc. Several classifiers can be provided that are not mandatory are related or dependent on each other. Depending on the application, the number of binary classifiers, i.e. the various defined and monitored specific anomaly states, can also be more than four, more than ten, more than 50, more than 100 or any number of more, as long as sufficiently fast data processing seems possible.

In the case of a classification of the actual state as an anomaly-free normal state and/or an assessment that a monitored specific anomaly state is not present, a signal feedback can take place. Preferably, the returned data can be used for at least one future measurement. The result of the classification is thus available and can be taken into account in the context of condition monitoring in the classification or assessment of future measurements.

The method can preferably also include an adjustment of the diagnosis, taking into account direct or indirect comparisons of characteristic values of at least one past actual state with limit values. The consideration of older characteristic values and comparative data collected over part or the entire service life of the fan and/or its motor can form a particularly good basis for decision-making. Furthermore, limit values—similar to the characteristic values described—can be dependent on applications/areas of use/installations, on a rotor speed/rotational rate and/or on a load condition. The classifiers can be adjusted during operation. Their configurations and data, such as calculated characteristic values and/or limit values, for example counters and/or tolerance values, are available to the computing unit at any time and can preferably be transmitted or retrieved via defined interfaces. Depending on the application, there may be different - parameterizable - limit values. Even if there are deviations from limit values, the system can react tolerantly in order not to immediately classify a current actual state into one of the two states and thus avoid or reduce incorrect classifications. For example, it can be determined statically or dynamically that one or more specific characteristic values for a certain number of measurements in a row are outside the limits defined by the limit values permissible value range before an actual state is classified as a general anomaly state. As described, a quota of limit value deviations over a certain number of past measurements is also conceivable. The sensor system can also be recalibrated here. The limit values can be reselected or defined so that a system tolerance against exceeding and/or falling below limit values can be changed. The calculation of characteristic values of future actual states, the comparison with limit values or the classification of the actual state based on the comparison can be adjusted. In other words, the diagnosis can be adapted by the operator or automatically as a self-learning system on the computing unit or via the defined interfaces as part of the external signal processing.

The characteristic values determined by measurements can be recorded in a log with a time stamp so that they can be taken into account in future adjustments to the diagnosis. Data can be passed on and/or secured/saved/deposited as part of signal processing, in particular with regard to the definition of anomaly states and classifiers. Classification results and classifier configurations can be saved together with a time stamp in an internal or external memory and can be read from there, and/or data can be transferred via communication interfaces, for example streaming to the cloud. In the case of constantly occurring limit value deviations, a self-diagnosis and a recalibration of the sensors can take place. If the sensor system is identified as faulty and cannot be recalibrated, information is given that further steps shown cannot be carried out. This information can be stored in logs on the one hand and communicated externally via the defined interfaces on the other. In the case of permanent limit value deviations, the current configuration of the classifier—ie the characteristic values and/or limit values associated with the respectively monitored general or specific anomaly state—can also be checked and, if necessary, adjusted.

With regard to the device for monitoring the operation of fans, the task mentioned above is achieved by a device with at least one nem sensor for detecting at least one input signal over at least one period of time to carry out at least one measurement, and a processing unit that is designed to run the method described above ren. The processing unit can be a processing unit of an electric motor of the fan, preferably an EC engines. The sensor can be one or more sensors internal to the fan or its motor and/or external sensors. The sensor system can include one or more of the following: a yaw rate sensor, an acceleration sensor, a temperature sensor, a microphone, a torque sensor, a pressure sensor, a moisture sensor, a force sensor and/or virtual sensors/soft sensors.

With regard to the improved fan, the object mentioned at the outset is achieved by a fan which comprises a device as described for monitoring its operation.

There are now various possibilities for embodying and developing the teaching of the present invention in an advantageous manner. For this purpose, on the one hand, reference is made to the claims subordinate to claim 1 and, on the other hand, to the following de explanation of preferred exemplary embodiments of the invention with reference to the drawing. In connection with the explanation of the preferred exemplary embodiments of the invention from the drawing, preferred refinements and developments of the teaching are also explained in general. Show in the drawing

1 shows a schematic flowchart of a first embodiment of the method according to the invention.

2 shows a schematic flowchart of a second embodiment of the method according to the invention.

The same or similar features are marked in the figures with the same reference signs. 1 generally shows a flowchart. In the upper part there is a diagnosis of an anomaly A as part of the monitoring of a company of a fan shown schematically. Monitoring starts with a measurement 1 over a period of time. In the example shown here, an internal sensor can detect input signals 2.1 for measurement 1. In the case of an acceleration sensor, these are, for example, the yaw rate and/or acceleration in one, two or three spatial axes. The input signals 2.1 are detected at a defined sampling rate, either over a defined period of time per measurement 1 with subsequent defined pauses, or alternatively continuously. For example, input signals 2.1 can be recorded for measurement 1 once per minute over a period of 5 seconds with a sampling rate of 1 kHz—that is, almost continuously. A measurement 1 includes the entirety of all data recorded in the period from the input signals 2.1, which originate from one or more internal or external sensors. This measurement 1 will be examined below. Measurements 1 can be evaluated by the engine's arithmetic unit (not shown) and/or by transmission to other terminals via defined interfaces (l ² C, Modbus, CAN, Wifi, ...), so that status monitoring can also be tracked via app/displays is possible. Characteristic values 2 are calculated from a measurement 1, which represent a current actual state. In the case of acceleration or sound pressure data, this includes, among other things, amplitudes of the frequency spectrum.

An example for the calculation of further characteristic values from one or more characteristic values can be given in the following case for the rolling bearing: a) a first specific comparison leads to the assessment that a first specific anomaly state, which indicates a non-critical anomaly on the inner ring, is present, b) a second specific comparison results in an assessment that a second specific anomaly condition indicative of an inner ring anomaly is not present; and c) a third specific comparison results in an assessment that a third specific anomaly condition indicative of a indicates an uncritical anomaly in connection with the roller bearing grease.

A corresponding new further parameter can be defined depending on which and how many specific comparisons in which way and/or Indicate severity and/or severity of anomaly. In this example, it is defined that there is a limit deviation as soon as the majority opinion - namely the first and third comparison compared to the second comparison - detects an anomaly. From this it can even be concluded that a critical anomaly is present and a measure is necessary, even if the first and third specific anomaly states are only non-critical anomaly states.

Characteristic values 2 calculated from a measurement 1 are compared with limit values 3.1. Suitable limit values 3.1 for this comparison 3 can have been determined in experimental or numerical investigations and limit the maximum permissible value range of characteristic values 2. The comparison 3 of calculated characteristic values 2 with limit values 3.1 allows the actual state to be classified 4 or classified into two or more states, in particular as an anomaly-free normal state 5 or an anomaly state 6. The comparison 3 can be carried out over several measurements 1 in order to To increase the robustness of the method against disturbance variables from the environment or from the environment.

For comparison 3, the number and/or the intensity of limit value deviations per period and/or the rate of change of limit value deviations—gradients—are considered. If, for example, particularly strong fluctuations or rapid changes occur within a defined period of time, ie if there are high gradients, then the deviation of the characteristic values 2 from the limit values 3.1 is weighted accordingly. The system can then react tolerantly in order not to immediately classify the current state into one of the states. So with wrong classifications 4 are avoided. In the event of permanent deviations from limit values, the sensors perform a self-diagnosis and, if necessary, recalibrate. If the sensor system is identified as faulty and cannot be recalibrated, information is given that further steps shown cannot be carried out. This information is stored in logs and/or communicated externally via defined interfaces. If the actual state is classified as an anomaly-free normal state 5, a first signal feedback R1 for future measurements 1 can take place. The result of classifying the current state as anomaly-free normal state 5 or as a general anomaly state 6 is transmitted to the outside by means of an output signal 7 via defined interfaces for further signal processing 7.1. For example, a signaling device (display, LED, loudspeaker) that is optional, integrated or mountable in the EC motor can inform a user in real time analogously/digitally, acoustically and/or visually about the current actual state. The current status is recorded in a log together with a time stamp. This log can be read out in real time or for later maintenance purposes via defined interfaces or displayed with an app. In addition, the classification result can be saved in an internal or external memory and/or communicated externally via the defined interfaces. Furthermore, the adaptation of the binary classifiers based on the protocol is possible. This can include adjusting the characteristic values 2 and/or limit values 3 or the counters/tolerance values. However, if something has changed in the application of the system, an adjustment of the input signals 2.1 can also be useful, for example if a user uses a different inlet nozzle, positions the fan in a different way, etc. It is possible to adjust all the factors that contribute to the classification of the actual state 4. The signal processing

7.1 can take place alternatively or additionally in the computing unit and cause an emergency shutdown or operational changes. The signal processing 7.1 can also result in human intervention. This can be pointed out to the user, for example by signal lights, sirens or similar means.

A further diagnosis of a specific anomaly B is shown schematically in the lower part of FIG. 1 . If the current state is classified as a general anomaly state 6, then specific characteristic values 8 are calculated from the current and/or previous measurements 1, for example in the engine's computing unit. These specific characteristic values 8 can differ from the general characteristic values 2 and can take into account the same input signals 2.1 or other input signals 8.1. In particular, the specific characteristic values 8 can be a subset of the general characteristic values 2 . The specific characteristic values 8 are in turn used for a comparison 9 with specific limit values

9.1 used. On the basis of this comparison 9 an assessment 10 is made that whether it is the general anomaly state 6 that is determined is a general, non-specific anomaly state 11 that is not a specific anomaly state, or whether it is a specific anomaly state 12 . If the actual state is classified and assessed as a general, non-specific anomaly state 11, a second signal feedback R2 can take place for future measurements 1. The result of this assessment 10 can also be transmitted to the outside as an output signal 13 via defined interfaces for further signal processing 13.1 or in a internal and/or external storage can be backed up. A general but not a specific anomaly state 11 exists if a general anomaly state 6 in the form of a fault in the EC motor or the application is diagnosed, which is not to be assessed as a specific anomaly state 12 . The other diagnoses of specific anomalies C and D are intermediate stages between diagnoses A and B. They each describe the monitoring carried out depending on the presence of a corresponding anomaly state on the previous stage A, C, D, i.e. shown above in FIG the presence of other specific abnormal conditions that are more severe than general abnormal condition 6 in diagnosis A and less severe than specific abnormal condition 12 in diagnosis B. Also the results of assessments in diagnosis C and D can be transmitted to the outside as an output signal for further signal processing 14, 15. In particular, the diagnoses A, B, C, D form a cascade. In particular, the specific anomaly state 12 within the scope of the diagnosis B can be a critical anomaly state.

2 also generally shows a flowchart. In the left part, a diagnosis of a general anomaly A is shown schematically, which corresponds to the diagnosis of a general anomaly A from FIG. However, the diagnosis of a specific anomaly B is not shown below the diagnosis of a general anomaly A in FIG. 2, but in the right part of FIG. 2. The difference between the two embodiments of the method from FIG. 1 and FIG in that the monitoring of whether a specific anomaly state 12 is present is set up in the method according to FIG. 2 independently of the diagnosis of an anomaly A. That means the diagnosis of a specific abnormality B, ie a specific anomaly state 12, can also be carried out if the current actual state and/or previous actual states have not been classified as a general anomaly state 6. This enables the realization of independent or unconditional classifiers. Since monitoring for the presence of a specific anomaly state 12 does not require a classification 4 of the actual state as a general anomaly state 6, the information that a general, non-specific anomaly state 11 is present is not transported as part of the signal feedback R2, but only the Information that a specific anomaly state 12 is not present 16. A general, non-specific anomaly state 11 can be diagnosed with the method according to FIG however, is not assessed as a specific anomaly state 12 on the basis of the characteristic values 8 .

If the specific anomaly state 12 is monitored independently of a previous classification 4 of an actual state as a general anomaly state 6 - for example to monitor natural wear and tear - then the diagnosis of a specific anomaly B, which examines a timely system failure, can be carried out at specific times or continuously . The general diagnosis of a general anomaly A and/or the diagnosis of a specific anomaly B can also only be carried out individually. The diagnosis of a general anomaly A can be made without further checking for the presence of a specific anomaly condition 12 . Thus, a possibly general anomaly state 6, which can possibly be classified as a general, non-specific anomaly state 11, can also be recognized if a specific anomaly state 12—possibly recognized on the basis of other characteristic values or other limit value deviations—has already been recognized. The further diagnoses C and D in FIG. 2 can relate to other binary classifications and, in particular, independently of the diagnoses A and B, monitor further deviating specific anomaly states which relate to deviating and specific damage patterns. On the other hand, one or more other diagnoses C, D at intermediate levels between the monitored with the diagnosis A general anomaly state 6 and the with the specific anomaly condition 12 being monitored with the diagnosis B, in particular if the specific anomaly condition 12 of the diagnosis B is a critical anomaly condition. These intermediate stages or the respective presence of their specific anomaly states are then independently monitored.

Overall, one or more diagnoses A, B, C, D connected in parallel or in series can be carried out continuously or at defined times based on one or more current or previous characteristic values with regard to one or more components of the fan and/or its motor , whereby one or more types of damage to one or more components can be monitored.

With regard to further advantageous configurations of the device according to the invention, reference is made to the general part of the description and to the appended claims in order to avoid repetition.

Finally, it should be expressly pointed out that the above-described exemplary embodiments of the device according to the invention only serve to explain the claimed teaching, but do not restrict it to the exemplary embodiments.

Reference List

A Diagnosis of a general anomaly

1 measurement .1 input signal

2 characteristics

3 Comparison .1 Limits

4 Classify

5 Anomaly-free normal state

6 General anomaly state R1 First signal return

7 Output Signal .1 Signal Processing B Diagnosis of a specific anomaly

8 Specific characteristics .1 Input signal 9 Comparison .1 Specific limit values

10 assessment

11 General non-specific abnormal condition

12 Specific anomaly condition

13 output signal

signal processing

R2 Second signal return

C Diagnosis of a specific anomaly as an intermediate stage

D Diagnosis of a specific anomaly as an intermediate stage

14 Signal Processing

15 signal processing

16 No specific anomaly condition

Claims

Expectations

1. A method of monitoring the operation of a fan, the method comprising the following steps:

Carrying out at least one measurement (1) by recording at least one input signal (2.1, 8.1) over at least one period of time,

Calculating characteristic values (2, 8) of an actual state based on the measured input signals (2.1, 8.1),

Comparison (3, 9) of the calculated characteristic values (2, 8) of at least one actual state with limit values (3.1, 9.1),

Classifying (4) the actual state based on the comparison (3) either as an anomaly-free normal state (5) or as a general anomaly state (6), and

Monitoring whether at least one specific anomaly condition (12) is present.

2. The method according to claim 1, wherein the characteristic values (2, 8) include one or more of the following:

Results of one or more previous comparisons (3, 9), rotation rate of the fan, acceleration in one, two or three spatial axes, at least one temperature, sound pressure, torque, pressure, in particular operating or ambient pressure,

Moisture values, measured forces, and/or virtual values, recorded for example by means of soft sensors.

3. The method as claimed in claim 1 or 2, wherein at least one further characteristic value (2, 8) is calculated from a plurality of characteristic values (2, 8).

The method of any one of claims 1 to 3, wherein monitoring whether a specific anomaly condition (12) is present comprises:

Calculation of specific characteristic values (8) of at least one state based on the measured input signals (2.1, 8.1), specific comparison (9) of the calculated specific characteristic values (8) with specific limit values (9.1),

Assessing (10) whether a specific anomaly state (12) is present using the specific comparison (9), and/or wherein the monitoring whether a specific anomaly state (12) is present is independent of an existing classification of an actual state as a general one anomaly state (6).

5. The method according to any one of claims 1 to 4, wherein further specific anomaly states (12) with different characteristics are defined and preferably form intermediate stages between the general anomaly state and a critical specific anomaly state, in which there is a significant malfunction of the fan, the immediate immediate action required.

6. The method according to any one of claims 1 to 5, wherein in the case of a classification (4) of the actual state as an anomaly-free normal state (5) and/or assessment (10) that a general anomaly state is not at the same time a specific anomaly state (11) and/or assessment (10) that a specific anomaly state is not present (16), a signal feedback (R1, R2) takes place for at least one future measurement (1).

7. The method according to any one of claims 1 to 6, further comprising the step:

Adaptation of the diagnosis (A, B, C, D) taking into account comparisons (3, 9) of characteristic values (2, 8) of at least one past actual state with limit values (3.1, 9.1). 8. The method according to claim 7, wherein the characteristic values (2,

8) be recorded in a log, preferably with a time stamp, to be taken into account in future adjustments to the diagnosis (A, B, C, D).

9. Device for monitoring the operation of a fan, comprising: at least one sensor for detecting at least one input signal (2.1, 8.1) over at least one period of time for carrying out at least one measurement (1), and a computing unit that is designed to to carry out the method according to any one of claims 1 to 8.

10. Fan with a device according to claim 9.