DE102019122727A1 - Method and device for monitoring a drive unit with a rotating drive component, in particular a motor of an internal combustion engine with a crankshaft, and the number of drive units with the device - Google Patents

Abstract

The invention relates to a method (1) for monitoring a drive unit with a rotating drive component, in particular a motor (110) of an internal combustion engine (100) with a crankshaft, the monitoring being a sound monitoring (10) of the rotating drive component by means of an acoustic sensor system (120 ) takes place on the drive unit, and within the scope of the sound monitoring (10): - a sound signal (130) of the rotating drive component is detected by means of the acoustic sensor system (120) and made available for evaluation for an evaluation device (200), - by means of a signal analysis of the sound signal In a signal analysis unit (140) from the sound signal (130) a number of signal identifications (141) is determined. According to the invention it is provided that - a number of events (131) is identified by means of a number of signal identifications (141) in the signal analysis unit (140) , in particular a signal marking an event for the turn the drive component is assigned, and a signal identifier (141) to be assigned to an event (131) is characterized in the signal analysis unit (140) by means of at least one metric (142), the metric (142) being defined by at least one characteristic parameter (143), which fixes the event on the sound signal (130), and the metric (142) is processed and provided in the signal analysis unit (140) for transfer to an evaluation unit (240) of an evaluation device (200), in particular the at least one signal analysis unit on the at least one Drive unit and the evaluation unit is available remotely from the drive unit, in particular are communicatively connected by a wireless connection.

Description

The invention relates to a method according to the preamble of claim 1, namely a method for monitoring a drive unit with a rotating drive component, in particular an engine of an internal combustion engine with a crankshaft, the monitoring being carried out as sound monitoring of the rotating drive component by means of acoustic sensors on the drive unit , and within the scope of sound monitoring: a sound signal of the rotating drive component is detected by means of the acoustic sensor system and made available for evaluation for an evaluation device, and a number of signal identifications are determined from the sound signal by means of a signal analysis of the sound signal in a signal analysis unit.

The invention also relates to a corresponding signal analysis unit and evaluation device for carrying out the method. The invention also leads to a drive unit and a system, in particular a fleet or farm, a number of drive units and an evaluation device with a number of signal analysis units.

An internal combustion engine of the type mentioned at the outset has, in particular, an engine and, in most cases, units that support the performance of the engine, such as turbochargers and exhaust gas and supply air ducts with internals. The power-transmitting units of a drive train connected to the internal combustion engine with clutches, transmission units, gears and other drive components also have rotating drive components mentioned at the beginning. An internal combustion engine can also be part of a complex drive unit, namely, for example, a gene set comprising a generator and the internal combustion engine. A rotating drive component and components driven by it, such as crankshafts and camshafts, connecting rods, valves, superchargers and gears or a rotor of a generator, are regularly essential for the operation of the drive unit and at the same time are subject to wear and tear or defects due to their intensive operating loads.

For example, a motor or the components of the drive train connected to it experience wear in the field and a wide variety of errors resulting from this, which can only be partially recognized and rectified by complex sensors or only by regular checks by service personnel. An example of a misalignment that gradually develops during operation, an undesired adjustment or a wear development is, for example, a valve clearance that changes during operation or an increasingly deflecting connecting rod bearing or connecting rod eye - this type of wear is within the scope of the normal design of the named components depending on the company profile. Serious errors or failures can also result as a result - to the extent that they are announced or even unannounced because they are undetected; For example, piston rings or inlet or outlet valves can break during operation and under load. Some wear developments or faults can directly result in the failure of the internal combustion engine, in particular the engine.

Some developments in wear and tear and defects are recognizable or announced, others are not; In any case, it is desirable to recognize a development of wear or a defect in advance of an obvious occurrence or at least immediately afterwards - this with as little effort as possible. This need arises not only for reasons of cost, but above all it should serve to make the service life of an internal combustion engine as effortless as possible, but in any case to avoid a total failure. Up to now, this has regularly meant that in internal combustion engines, such as large machines (e.g. large diesel or gas engines), in addition to a conventional monitoring sensor system, a minimum use of additional sensors is provided for extended sensor signaling and corresponding signal processing. Nevertheless, even with such large machines of a drive unit with a rotating drive component, costs play a role, not only for the use of the sensor system, but also for the associated signal processing of the sensor signaling and signal processing measured by the sensor system.

U.S. 5,230,316 describes, for example, a method known per se for detecting knocking conditions in an internal combustion engine by evaluating frequency spectra and depending on the angle of rotation of a crankshaft. A knock intensity is taken into account in relation to the level and frequency shift of a characteristic crank angle by means of a frequency spectrum in a measurement window of the rotation angle. Such a method works with the raw data of a vibration sensor both in terms of the angle of rotation and in terms of frequency resolution - the volume of data is correspondingly high.

In contrast, a method for detecting anomalies during operation of a rotating machine, which works with raw acoustic data from a sound sensor, is known from, for example U.S. 4,669,315 A . There it is described how an enveloping signal of an acoustic raw data signal of a sensor of a rotating machine can be obtained and used in the context of a diagnostic tool to detect certain anomalies. So there is one The monitoring mentioned at the beginning is described as a sound monitoring of a rotating drive component by means of an acoustic sensor system on the drive unit, with a sound signal of the rotating drive component being recorded by means of the acoustic sensor system and made available for evaluation for an evaluation device, with a signal analysis of the sound signal in a Signal analysis unit from the sound signal a number of signal identifications is determined. For this and other such methods known per se, however, different sensors and different diagnostic functions are often required for the detection of different errors, which is the case in the U.S. 4,669,315 A is quite complex and involves computational effort. In addition, installation and replacement - for example in the event of a sensor failure - cause unnecessary costs.

In addition, however, the detection of errors in the engine z. B. with a built-in on the engine diagnostic or control unit is generally expensive. On the one hand, the hardware is sometimes exposed to demanding operating conditions such as vibrations, temperatures and dirt, etc. and has to withstand these over extreme operating conditions and operating times. On the other hand, the diagnostic and control hardware installed on the engine is limited in terms of computing power for control and diagnostic devices and can only be increased to a limited extent. Updating fault or wear detection algorithms and updating the diagnostic and control devices on each engine is logistically complex, but in any case only cyclically or lengthy to a limited extent. In addition, it is only possible to react to acute changes in operating states or operating conditions to a limited extent.

Therefore, in the future, the aim is generally to relocate the signal processing effort externally to the drive unit, e.g. at a distance from the motor in such a way that computing effort becomes available and the operating conditions are appropriate. Processing of sensor data can take place, for example, in a centralized or regional manner, for example at a regionally advantageous location for affected vehicles in a mine, in a regional main computer of a ship for all engines of the ship or even centrally in a central data processing system of a provider or manufacturer for all drive units or Motors of a fleet or farm - in principle, powerful large standard computers or data processing systems with all the advantages of applications for extended data analysis, such as so-called back-up applications, big data applications or the like, can then be used. In this application, these are cheaper than small individual computers in relation to their performance and are logistically easier to update or adapt. In addition, data from different or similar engines can be compared - an option that is usually not available on individual engines. Trend analyzes and predictive analysis methods for the internal combustion engines in the field are particularly advantageous.

A method that is advantageous in this sense, with a corresponding computation load distribution, is shown in U.S. 5,157,610 A for controlling a motor - a monitoring effort is divided between a station installed in the vehicle and a stationary base station, the following method steps being carried out already in the station installed in the vehicle. Operating conditions of the vehicle are recorded, and of the data representative of the recorded operating conditions that do not require real-time processing, those that are intended to be processed in the base station are transmitted to the base station, the base station transferring these data from the im Vehicle installed station. These data are processed in accordance with the data stored in the base station, the base station transmitting the processed data to a receiver in the station installed in the vehicle and to a control device in the station installed in the vehicle which is connected to the receiver installed in the vehicle. The control device of the station installed in the vehicle is designed in such a way that it checks and / or displays the operating states of the vehicle in accordance with the processed data received.

However, if signal processing is to be relocated externally to the drive unit, i.e. centrally or regionally, considerable amounts of sensor data must be transmitted. As described above, this takes place regularly by means of a transmission of the raw data from the sensor system. So-called data bloggers, with which it is possible to transmit raw data from the sensors as sensor data to a central or regional office, have been common so far. It can also be done in such a way that the data made available is limited in terms of its quantity or temporal resolution - data bloggers often work with low temporal resolutions in the range of, for example, 1 Hz. This allows a significant reduction in the amount of data to be achieved and an otherwise necessary reduction Broadband data transmission can be restricted for reasons of cost and effort. With the increasing complexity of the systems, however, it can be foreseen that the transmission, especially of large amounts of raw data, is either too expensive or limited due to a lack of capacity. It turns out that such approaches in spite of in themselves for example JP 2-130429 A well-known filter and analysis methods, also for acoustic data, which have so far been too specific and complex - they only offer limited possibilities for adequate evaluation and also only insufficiently reduce the amount of raw data.

It is desirable to identify abnormalities in the operation of a drive unit as generally as possible, such as different faults or signs of wear, with comparatively little effort on the drive unit. This means that only limited use of additional hardware and evaluation capacity, advantageously without additional sensors, should be required on the drive unit, and signal processing on the drive unit, in particular a motor for the detection and analysis of anomalies, should nevertheless be manageable and nevertheless sufficiently precise. A detailed detection and analysis of anomalies such as errors or wear and tear on the drive unit should be carried out externally to the drive unit - i.e. "remotely", if regional or central. It is desirable - be it on the drive unit or "remotely" from it - to keep a data volume and thus the transfer and / or analysis effort for data as limited as possible.

This is where the invention comes in, the object of which is to specify a method and a device of the type mentioned, which offers potential for improvement with regard to the disadvantages mentioned in the introduction in the sound monitoring of a drive unit with a rotating drive component, with a data volume to be transmitted for monitoring the Drive unit, in particular an amount of data to be transmitted externally, insofar as it is primarily “remote” to the drive unit for further analysis, is kept limited or reduced. In any case, some of the aforementioned aspects should be addressed. In particular, a device and a method of the type mentioned at the outset are to be specified, the task of which is also to offer the most general possible approach to the detection of anomalies in the context of sound monitoring with reduced transmission effort for sensor data.

This object is achieved by a method according to claim 1 and a corresponding signal analysis unit according to claim 15, in particular as part of an evaluation device according to claim 17.

• - ein Schallsignal der drehenden Antriebskomponente mittels der akustischen Sensorik erfasst und zur Auswertung für eine Auswerteeinrichtung bereitgestellt wird,
• - mittels einer Signalanalyse des Schallsignals in einer Signalanalyseeinheit aus dem Schallsignal eine Anzahl Signalkennzeichnungen festgestellt wird. Erfindungsgemäß ist vorgesehen, dass
• - eine Anzahl von Ereignissen mittels einer Anzahl von Signalkennzeichnungen in der Signalanalyseeinheit identifiziert wird. So kann im Rahmen einer Weiterbildung mit besonderem Vorteil vorgesehen sein, dass eine Signalkennzeichnung einem Ereignis für die drehende Antriebskomponente zugeordnet wird. Erfindungsgemäß ist weiter vorgesehen, dass
• - eine einem Ereignis zuzuordnende Signalkennzeichnung in der Signalanalyseeinheit mittels wenigstens einer Metrik charakterisiert wird, wobei die Metrik durch wenigstens einen Kenngrößenparameter definiert ist, der das Ereignis am Schallsignal festmacht.

With regard to the method, the invention is based on a method mentioned at the beginning for monitoring a drive unit with a rotating drive component, in particular an engine of an internal combustion engine with a crankshaft, the monitoring being carried out as sound monitoring of the rotating drive component by means of an acoustic sensor system on the drive unit, and Sound monitoring framework:

• - a sound signal of the rotating drive component is detected by means of the acoustic sensor system and made available for evaluation for an evaluation device,
• a number of signal identifications are determined from the sound signal by means of a signal analysis of the sound signal in a signal analysis unit. According to the invention it is provided that
• - a number of events is identified by means of a number of signal identifications in the signal analysis unit. Thus, within the scope of a further development, it can be provided with particular advantage that a signal identifier is assigned to an event for the rotating drive component. According to the invention it is further provided that
• a signal identifier to be assigned to an event is characterized in the signal analysis unit by means of at least one metric, the metric being defined by at least one characteristic variable parameter which attaches the event to the sound signal.

According to the invention, it is further provided that the metric is processed and provided in the signal analysis unit for transfer to an evaluation unit of an evaluation device.

• - ein Schallsignal der drehenden Antriebskomponente mittels der akustischen Sensorik erfasst und zur Auswertung für eine Auswerteeinrichtung bereitgestellt wird,
• - mittels einer Signalanalyse des Schallsignals in einer Signalanalyseeinheit aus dem Schallsignal eine Anzahl Signalkennzeichnungen festgestellt wird. Erfindungsgemäß ist weiter vorgesehen, dass
• - eine Anzahl von Ereignissen mittels einer Anzahl von Signalkennzeichnungen identifiziert in der Signalanalyseeinheit wird,
• - eine einem Ereignis zuzuordnende Signalkennzeichnung in der Signalanalyseeinheit mittels wenigstens einer Metrik charakterisiert wird, wobei die Metrik durch wenigstens einen Kenngrößenparameter definiert ist, der das Ereignis am Schallsignal festmacht, und
• - die Metrik in der Signalanalyseeinheit zur Übergabe an eine Auswerteeinheit einer Auswerteeinrichtung aufbereitet und bereitgestellt wird.

The invention is based on a signal analysis unit of an evaluation device for monitoring a drive unit with a rotating drive component, in particular a motor of an internal combustion engine with a crankshaft, the monitoring being carried out as sound monitoring of the rotating drive component by means of an acoustic sensor on the drive unit , and within the scope of sound monitoring:

• - a sound signal of the rotating drive component is detected by means of the acoustic sensor system and made available for evaluation for an evaluation device,
• a number of signal identifications are determined from the sound signal by means of a signal analysis of the sound signal in a signal analysis unit. According to the invention it is further provided that
• a number of events is identified in the signal analysis unit by means of a number of signal identifications,
• a signal identifier to be assigned to an event is characterized in the signal analysis unit by means of at least one metric, the metric by at least one Characteristic parameter is defined, which attaches the event to the sound signal, and
• - The metric is processed and provided in the signal analysis unit for transfer to an evaluation unit of an evaluation device.

Thus, within the scope of a further development, it can be provided with particular advantage that the at least one signal analysis unit is available on the at least one drive unit and the evaluation unit is available away from the drive unit, in particular communicatively connected by a wireless connection.

Put simply, the concept of the invention provides that the sound signal is subjected to a preliminary evaluation by means of a signal analysis in the signal analysis unit, in particular on the basis of the raw data of the sound signal. As a result, a number of events can be identified in the signal analysis unit by means of a number of signal identifications and a signal identification to be assigned to an event is characterized in the signal analysis unit by means of at least one metric. The metric is defined by at least one characteristic parameter.

After this preliminary evaluation, the metric is prepared and provided in the signal analysis unit for transfer to an evaluation unit of an evaluation device. For this purpose, the metric and / or the at least one characteristic variable parameter preferably receives assignment information about the event. An event is preferably characterized by means of a number of metrics that can be arranged in an event field such as an event vector. Several events can thus be combined in an event matrix based on the assigned signal identification in the signal analysis unit and transmitted to the evaluation unit. The sound signal itself and / or the signal identification, in particular the raw data of the sound signal and / or the signal identification, can advantageously remain in the signal analysis unit, in particular be overwritten in the signal analysis unit. Only the processed metric or the event vector is then actually transmitted to the evaluation unit. As a result, the concept provides - in abbreviated form - based on a generally available sound monitoring system with a broad field of application, an external detailed evaluation, whereby instead of the sound signal itself, namely the raw data of the sound monitoring system, metrics of assigned events obtained only from the local pre-evaluation are preferred must be transmitted to an evaluation unit for preferably external detailed evaluation. A corresponding signal analysis unit for transfer to an evaluation unit of an evaluation device for processing and provision is available locally on the drive unit as a device for pre-evaluation and transmission. A preferably external evaluation unit for detailed evaluation is part of an evaluation device and a system comprising an internal combustion engine and a computer unit for analyzing event vectors.

The concept of the invention is initially based on the assumption that many faults in an internal combustion engine, in particular an engine, show acoustic signatures that do not occur during normal operation - this can be given, for example, in the form of a noise or a recurring special acoustic signal. In particular, large engines such as diesel or gas engines already use acoustic sensors such as knock or acceleration sensors as standard in order to detect harmful knocking combustion or to regulate it through the engine control system, as explained by way of example using the prior art mentioned at the beginning. Such sensors are therefore available on gas engines or can also be installed in other large engines such as diesel engines without any problems. It is also shown that such acoustic sensors can be installed and retrofitted comparatively inexpensively, since they are regularly designed for the detection of structure-borne noise and can be retrofitted on the outside of an internal combustion engine. Structure-borne sound detectors, in particular acoustic or sound detectors, can also be installed in areas of an internal combustion engine that are exposed to less extreme conditions in terms of temperature, pressure and vibration, and yet these structure-borne sound detectors or other acoustic sensors suitable for detecting raw acoustic data are suitable for delivering raw data in Form of acoustic signals that are suitable for evaluating and recognizing errors or wear and tear in an internal combustion engine or its components. These considerations also apply to components of a drive train on an internal combustion engine, such as a clutch, a transmission, and to the periphery of the engine such as a turbocharger, exhaust gas or supply air peripherals and rotating parts such as flaps or the like. In particular, however, the concept of the invention is used to use acoustic sensors in order to identify faults or wear on the engine, specifically to identify faults and wear beyond the usual knock detection.

Based on these considerations, the invention has recognized that acoustic data of a sound signal are temporally or spectrally highly resolved and the raw data quantities that arise can nevertheless be reduced - the invention is based on the consideration that several hundred or thousand samples per Working cycle leads to a raw data volume of raw sensor data that can be significantly reduced.

The concept of the invention proposes a method and a device with a comparatively reduced signal processing on the engine or the like drive unit for measured acoustic signals with the claimed solution, with an actual event on the engine corresponding to a sound event in the acoustic signal, ie a signal identification for the sound event the event - the reduced signal processing is based on the processing of the signal identification for the event that makes up the sound event, in particular only the signal identification for the event that makes up the sound event. A sound event is basically any noticeable noise component in the acoustic signal. As a result, the concept of the invention means that one or more suitable parameters can be assigned to each actual event on the engine, ie the corresponding sound event - in the present case these are described as one or more metrics that allow the noise to be expressed in the form of the sound event to characterize and distinguish from others as best as possible; To this end, the metric is defined by at least one characteristic variable parameter which fixes the sound event on the sound signal in the form of a characteristic signal identifier or a characteristic signal part.

The concept has also recognized that a number of metrics are combined to form an event and, for example, as an event vector - especially instead of the raw data - for an event or several metrics in an event vector for an event or several events to an external, preferably central or regional evaluation unit can be transmitted as an analysis point. In particular, only the parameters for one event - here at least one metric with at least one parameter parameter - are actually transmitted. In the following, with regard to the sound event, reference is primarily made to the signal identification or the signal part, the sound event corresponding to an actual event on the engine.

To this end, the concept of the invention proposes a preliminary evaluation or evaluation of the raw data up to at least one metric with one or more parameters for an event. This leads to a drastic reduction in the amount of raw data to be transmitted and therefore pre-evaluated in the form of metrics with one or more parameters for an event.

In contrast, the concept of the invention has recognized that, for example, within the framework of the claimed system, a further analysis can be carried out in the regional or central evaluation point, which allows a specific error to be optimized and developed over the running time of the engine new bug to apply bug detection. Nonetheless, it has been shown that the preliminary evaluation, apart from a metric for an event or an event vector for an event, can in principle also be used and implemented in the local control and computing unit on an internal combustion engine or a drive train, in particular an engine.

In short, the invention therefore provides for the raw data to be reduced to a metric for one or more events in the raw data of the acoustic sensors; Several metrics for an event can be combined in one event vector - in this respect several event vectors or several metrics can be provided for several events, which are then available in the context of a preliminary evaluation of an evaluation unit at the local control and processing unit. Only the metric or the event vector with a number of metrics is then transmitted to an external analysis unit, which is located centrally and regionally and is available for a more comprehensive analysis.

In order to achieve the object, the invention also leads to a drive unit according to claim 16 and a corresponding evaluation device according to claim 17 with the signal analysis unit and the evaluation unit.

The drive unit according to the invention with a rotating drive component has a signal analysis unit according to the concept of the invention, in particular the drive unit has a motor of an internal combustion engine with a rotating drive component in the form of a crankshaft, with the monitoring of a sound monitoring of the rotating drive component by means of acoustic sensors on the drive unit is provided.

The drive unit thus has a sensor system for sound monitoring and suitable hardware of the signal analysis unit for preprocessing the sound signal data within the scope of the signal analysis according to the invention and suitable hardware of a transmission device for processing, providing and transferring the metrics from the signal analysis unit to an evaluation unit of an evaluation device.

The evaluation device according to the invention has the signal analysis unit and the evaluation unit designed to carry out the method according to the invention. In particular, the evaluation device according to the invention has at least a signal analysis unit on the at least one drive unit, the evaluation unit being available remotely from the drive unit, in particular these are communicatively connected by a wireless connection. According to the concept of the invention, the evaluation device thus advantageously has a number of signal analysis units on and for a number of drive units.

The evaluation device has corresponding hardware of a receiving unit for receiving the metric from the signal analysis unit as well as a storage medium and a preferably external, in particular regional or central, evaluation unit for storing and evaluating the metric, in particular a number of event vectors. Nevertheless, the concept of the invention can also be implemented with a local evaluation unit at the location of the signal analysis unit.

To achieve the object, the invention also leads to a system according to claim 18, i.e. a system, in particular a fleet or farm, a number of drive units according to claim 16 and a number of signal analysis units according to claim 15 and the evaluation device according to claim 17.

One advantage of the concept of the invention results first of all from the fact that a sensor system for sound monitoring is comparatively inexpensive. For gas engines in particular, the sensor system is necessary for engine knock control anyway and is therefore available as standard, i.e. no additional sensor system is required. One and the same sensor system can be used to detect several different errors. Examples are a changed valve clearance on the inlet and outlet valve, breakage of piston rings - an impending fault can already be recognized here - broken valves, worn connecting rod bearings or misfiring. By determining fewer metrics for each detected event, the data can be transmitted with little effort to an external, preferably regional or central, evaluation unit as an evaluation point, where a more complex, detailed evaluation, e.g. B. can be carried out in the form of a trend analysis or forecast of further development. This keeps the signal processing effort on the motor low.

The method and the device according to the concept of the invention also has the advantage that it applies equally to all types of errors: standardized metrics are preferably determined for each detected noise and these metrics are transmitted. This means that there is no effort to adapt the method on the engine for new errors to be detected. It is only necessary to adapt the recognition algorithms in the external, preferably regional or central, evaluation unit as the evaluation point, which is logistically significantly simplified compared to previous procedures.

The method and the device according to the concept of the invention form the basis for a number of interesting diagnostic functions for current errors (which should or must be repaired immediately if necessary) and for predictive methods (also known under the subject keywords "Predictive Maintenance", "Condition Based Maintenance"), which make it possible to predict the future development of an impending fault by observing the metrics of an event. This saves service costs and prevents expensive consequential damage through early detection.

Advantageous further developments of the invention can be found in the subclaims and indicate in detail advantageous options for realizing the concept explained above within the scope of the task and with regard to further advantages.

As already explained, it is advantageously provided that a number of metrics are combined in an event vector or similar event field and assigned to the event. A plurality of events can therefore advantageously be combined in an event field with corresponding event vectors for the characteristic description of a number of signal identifications.

It is advantageously provided that the sound signal of the rotating drive component is recorded as structure-borne noise on the drive unit with an acoustic sensor. Structure-borne noise can be recorded without great effort and at comparatively harmless points on the drive unit with uncritical environmental conditions.

It is advantageously provided that the drive unit has an internal combustion engine and / or a generator, a motor of the internal combustion engine having the at least one rotating drive component as a crankshaft and / or camshaft and / or the generator having the at least one rotating drive component as a rotor. It can advantageously be provided that the acoustic sensor is arranged on a cylinder of the engine. The concept is therefore not limited to a drive unit in the form of an internal combustion engine - the acoustic sensor can be arranged, for example, on a pole of the generator.

It is advantageously provided that the sound signal is made available as transient or frequency-resolved raw data; both in the time domain and in the frequency domain or in Both areas can be used to generate raw data as a sound signal for sound monitoring. It is advantageously provided that the raw data have a transient sound signal curve and / or a frequency-resolved sound frequency spectrum for at least part of a working cycle of the rotating drive component. For example, it can also only provide a frequency spectrum as raw data for a pre-evaluation within the meaning of the concept of the invention for a specific time range of the work cycle. For example, it is advantageously provided that the transient sound signal curve of the working cycle has a sound signal amplitude of the acoustic sensor system over a rotation angle of the working cycle of the rotating drive component.

As part of a particularly advantageous development, it is provided that the metric characterizing the signal identification is processed as a representative of the sound signal, in particular the signal identification, for transfer to the evaluation device. This can, but does not have to be, the only information provided for transmission from the sound signal after its preliminary evaluation in the signal analysis unit, the advantages of the concept of the invention being particularly effective. It is advantageously provided, for example, that the sound signal itself and / or the signal identification, in particular the raw data of the sound signal and / or the signal identification, remain in the signal analysis unit, in particular are overwritten in the signal analysis unit.

A signal identifier to be assigned to an event, that is to say a sound event, with at least one metric is advantageously selected from the group consisting of the position, extent and energy of the event and can be characterized on the basis of the sound signal. In the sense of this development, this is a particularly advantageous characterization of each signal identification that is based on an event. To this end, it is advantageously provided that the signal identification and / or the associated metric is assigned to the event. It is advantageously provided that the metric is determined by means of one or more characteristic variable parameters of the sound signal.

In the context of a particularly advantageous development, it is provided to implement the concept that the signal analysis includes an amplitude analysis of transient or frequency-resolved raw data of the sound signal.

In a first particularly preferred variant, it is advantageously provided that the transient raw data of the sound signal are provided as absolute or relative sound amplitude. As explained on the basis of the exemplary embodiments of this first variant, the amplitude analysis in particular can include a threshold value analysis in relation to a number of threshold values.

In a second particularly preferred variant, it is advantageously provided that the transient raw data of the sound signal are provided as a cumulative sound amplitude. As explained with reference to the exemplary embodiments of this second variant, the amplitude analysis in particular can include a difference value analysis with respect to a number of references, in particular with respect to a cumulative sound amplitude.

It is generally advantageous, in particular also within the scope of the first and second variant, that one or more characteristic parameters are determined for the metric selected from the group consisting of position, extent and energy by means of an amplitude analysis of the sound amplitude; i.e. in terms of an absolute or relative or cumulative sound amplitude.

For a transient signal of an absolute or relative sound amplitude, a first and / or second or even further thresholds can be defined. By means of a mean position crank angle, a signal identification based on an event can be defined in the sound signal, i.e. a sound event. The extension can be defined as the start and end point of a crank angle. The energy of the event can be defined as the amplitude of the signal characterization. This has proven to be particularly advantageous with regard to an absolute or relative sound amplitude.

In general, a value on a horizontal axis that belongs to a greatest gradient or distance increase (point of maximum gradient or intensity maximum) can be used as a characteristic parameter for a metric, i.e. location of an event; this also has a certain relation to the duration of the event. With regard to a duration of the event, a value on the horizontal axis that belongs to the middle between the values at the beginning and end of the event on the vertical axis can be used as the position

With regard to the duration of the event, a middle point between the beginning and the end of the event on the horizontal axis can also be used as the location. These modifications have proven to be particularly advantageous with regard to a cumulative sound amplitude.

In the context of a particularly advantageous development, above all the evaluation unit - i.e. external to the signal analysis unit - it is provided that the characteristic parameter of the metric is formed for recognizing an event development based on a pattern development of the characteristic variable parameter of the metric assigned to the event in a parameter space. For example, a trend of a change can be determined over time in the noise component.

In the context of a particularly advantageous development, especially the evaluation unit - i.e. external to the signal analysis unit - it is provided that a pattern development of the characteristic parameter of the metric assigned to the event is assigned a list of causes or measures for the event and is stored. This is advantageous for and for displaying a cause recognition proceeding from the pattern development in the context of the evaluation device, which is designed to give feedback to the user or the control of the drive unit.

• 1 ein Schema einer bevorzugten grundsätzlichen Ausführungsform einer Brennkraftmaschine mit einer Signalanalyseeinheit zur Vorauswertung von Daten aus einer Schallüberwachung als Teil eines Systems und zur Übertragung von Ereignisvektoren im Rahmen einer Auswerteeinrichtung gemäß dem Konzept der Erfindung;
• 2A ein Beispiel gemessener akustischer Rohdaten in Form eines Schallsignals mit typischen Schallsignalamplituden aufgetragen über den Kurbelwellenwinkel als Grundlage für die Vorauswertung von Daten aus Schallüberwachung und Übertragung von Ereignisvektoren;
• 2B ein illustrierendes Beispiel für eine erste Ausführungsform einer Methodik zur Erkennung von Ereignissen, d.h. Schallereignissen im Schallsignal, im Schallsignal der 2A im Rahmen der Vorauswertung von Daten aus der Schallüberwachung mit einer Signalanalyseeinheit;
• 3A ein illustrierendes Beispiel für eine zweite Ausführungsform einer Methodik zur Erkennung von Ereignissen, d.h. Schallereignissen im Schallsignal, im Schallsignal der 2B im Rahmen der Vorauswertung von Daten aus der Schallüberwachung;
• 3B ein Detail zur zweiten Ausführungsform einer Methodik der 3A;
• 4 ein Schema eines Ereignisfeldes mit Metriken in Beziehung zu einem Ereignis und mit einer Liste von Ursachen oder Maßnahmen in Bezug auf das Ereignis;
• 5 ein Ablaufdiagramm eines bevorzugten grundsätzlichen Ablaufs eines Verfahrens zur Überwachung einer Brennkraftmaschine zur Vorauswertung von Daten aus einer Schallüberwachung und Übertragung von Ereignisvektoren gemäß dem Konzept der Erfindung.

Embodiments of the invention will now be described below with reference to the drawing in comparison to the prior art, which is also shown in part. This is not necessarily intended to represent the embodiments to scale; rather, the drawing, where useful for explanation, is shown in a schematic and / or slightly distorted form. With regard to additions to the teachings that can be seen directly from the drawing, reference is made to the relevant prior art. It must be taken into account here that various modifications and changes relating to the shape and detail of an embodiment can be made without deviating from the general idea of the invention. The features of the invention disclosed in the description, in the drawing and in the claims can be essential for the development of the invention both individually and in any combination. In addition, all combinations of at least two of the features disclosed in the description, the drawing and / or the claims fall within the scope of the invention. The general idea of the invention is not restricted to the exact form or the detail of the preferred embodiment shown and described below or restricted to an object that would be restricted in comparison to the object claimed in the claims. In the case of the specified measurement ranges, values lying within the stated limits should also be disclosed as limit values and be able to be used and claimed as required. Further advantages, features and details of the invention emerge from the following description of the preferred embodiments and with reference to the drawing; this shows in:

• 1 a diagram of a preferred basic embodiment of an internal combustion engine with a signal analysis unit for pre-evaluating data from sound monitoring as part of a system and for transmitting event vectors within the scope of an evaluation device according to the concept of the invention;
• 2A an example of measured acoustic raw data in the form of a sound signal with typical sound signal amplitudes plotted over the crankshaft angle as a basis for the preliminary evaluation of data from sound monitoring and the transmission of event vectors;
• 2 B an illustrative example of a first embodiment of a method for recognizing events, ie sound events in the sound signal, in the sound signal of FIG 2A as part of the preliminary evaluation of data from sound monitoring with a signal analysis unit;
• 3A an illustrative example of a second embodiment of a method for recognizing events, ie sound events in the sound signal in the sound signal of FIG 2 B as part of the preliminary evaluation of data from noise monitoring;
• 3B a detail of the second embodiment of a methodology of 3A ;
• 4th a schema of an event field with metrics related to an event and with a list of causes or actions related to the event;
• 5 a flowchart of a preferred basic sequence of a method for monitoring an internal combustion engine for preliminary evaluation of data from sound monitoring and transmission of event vectors according to the concept of the invention.

1 shows schematically a combination in a block diagram 1000 a drive unit and an evaluation device, here an internal combustion engine as an example 100 in combination with a central or regional evaluation device 200 . The evaluation device 200 includes an evaluation unit 240 for further evaluation of data in the context of monitoring the internal combustion engine 100 , the evaluation unit 240 in the present case remote from the signal analysis unit 140 on the internal combustion engine 100 is ready - the signal analysis unit 140 on the internal combustion engine 100 and the evaluation unit provided externally for this purpose 240 the evaluation device 200 are over a wireless connection 300 connected to data transmitting communication.

In another exemplary embodiment, the drive unit does not necessarily have to include an internal combustion engine, but can be designed differently. So here is the Internal combustion engine 100 In principle, it is representative of any other component of a drive unit in other exemplary embodiments - this can, for example, be a component such as a generator or a genset that the internal combustion engine 100 and comprises a generator. In principle, a drive unit can also comprise an electric drive, a fuel cell or a similar drive unit with a rotating component.

The present is in 1 it can be seen schematically that the internal combustion engine 100 preferably within a system 1001 also part of a fleet or farm 101 with a number of drive units - presently a number of internal combustion engines 100.1 , 100.2 ... 100.n-, where the number n of each internal combustion engine 100.1 , 100.2 ... 100.n assigned analysis units 140.1 , 140.2 ... 140.n also part of the evaluation device 200 are. That is, the evaluation device 200 in this respect includes a number n of the drive units with the internal combustion engines 100.1 , 100.2 100.n arranged signal analysis units 140.1 , 140.2 ... 140.n , as well as an external, centrally or regionally, arranged evaluation unit 240 the evaluation device 200 .

In a further development of the concept of the invention, according to the embodiment 1 provided that the internal combustion engine 100 an engine 110 and an acoustic sensor system 120 for this engine 110 having. The motor 110 itself has a number of rotating drive components, of which only a few are shown symbolically here by way of example, such as a number of cylinders here 111 with corresponding piston drive, corresponding valves 112 for the cylinders 111 , a camshaft 113 and a crankshaft 114 . Such a basic design of an engine 100 is present in particular in diesel, gas or Otto engines - in the case of a diesel engine, the engines regularly have one or more injectors 115 for injecting fuel into the cylinders 111 on - similarly, gas metering devices or the like are provided in a gas engine.

The acoustic sensors 120 is present with a number of acoustic sensors 121 equipped, with at least one acoustic sensor 121 on a cylinder 111 of the motor 110 is arranged on the cylinder to detect the structure-borne noise. In the present case, the acoustic sensor detects 121 as one of a number of such sensors 121 the sensors 120 a sound signal 130 that provides, for example, transient or frequency-resolved raw data; in the present case after a filter 160 . In the following, a sound signal is provided as a transient raw data signal as an example to explain the preferred embodiment, but not in a limiting manner, the transient sound signal 130 a transient sound signal curve over a working cycle of the rotating drive component, namely a crankshaft here 114 represents and in 2A will be explained by way of example.

That by means of the acoustic sensors 120 detected sound signal 130 is for evaluation in the evaluation device 200 intended.

According to the concept of the invention, the evaluation device comprises 200 a signal analysis unit 140 for preliminary evaluation of the raw data of the transient sound signal 130 and a central evaluation unit 240 for detailed evaluation of data; but not the raw data, which will be explained in detail later.

First, the raw data of the sound signal 130 after the said filter 160 the signal analysis unit 140 , i.e. at the location of the engine 110 the internal combustion engine 100 made available. The signal analysis unit 140 is formed in the sound signal 130 based on the raw data of the same an event 131 when operating the engine 110 Signal identification to be assigned 141 ascertain.

Such a signal identification 141 is shown here schematically and is based on a metric 142 characterized, being a metric 142 at least one characteristic parameter 143 is assigned to which the event 141 attached to the sound signal. One an event 131 assignable signal identification 141 with at least one metric 142 is advantageously selected from the group consisting of the location, extent and energy of the event and can be characterized on the basis of the sound signal. In the sense of this development, this is a particularly advantageous characterization of each signal identification 141 on an event 131 going back. The signal identification 141 and / or the associated metric 142 so becomes the event 131 assigned, the metric being determined by means of one or more characteristic parameters of the sound signal.

An event basically means any process that occurs over time or has an extended frequency, which - caused by a normal or abnormal operating situation of the drive device - acoustic phenomena in the sound signal 130 causes above the noise floor; thus a signal identification 141 that by means of a metric 142 can be circumscribed. The number of events 131 is not limited to a single sound signal, but a large number of events can occur 131 identified and each with one or more signal identifiers 141 can be identified - each of the signal labels 141 can be with one or more metrics 142 - again definable with one or more parameters 143 - can be attached to the sound signal. The result is in the signal analysis unit for transfer to an evaluation unit 240 the evaluation device 200 the metric 142 processed, so z. B. with a label 144 provided which the event 131 is assigned and provided accordingly formatted for a transfer device. The transfer device 150 can have one or more interfaces for wireless communication with a corresponding receiving device 250 the central evaluation unit 240 be trained.

The central evaluation device 200 thus includes the receiving unit 250 , a memory 220 for the events 131 as part of one or more metrics 142 that act as event vectors 230 can be filed. The event vectors 230 can then the evaluation unit 240 can be made available for the application of further evaluation processes.

The evaluation unit 240 has a number of evaluation modules 260 for example, for pattern recognition, trend analysis or other further detailed evaluation based on the event vectors 230 are trained. The evaluation unit 240 also has a correlation stored there with a list 270 of causes or actions based on a particular pattern or trend or event vector 230 can be assigned.

As a result, with the combination 100 , 200 or the system 1000 a series of events with a considerable data reduction in relation to the raw data from a sound monitoring 131 and associated event vectors 230 identified and sent to an evaluation unit 240 be transmitted; the event vectors 230 are then used to identify trends, causes and measures related to an event 131 . This is explained below with the aid of a few specific examples.

In relation to 2A is within the scope of an exemplary embodiment of a first variant a first method for identifying events 131 described. In the 2A shown transient raw data of the sound signal 130 stand as absolute sound amplitude 130A over a crank angle 130W to disposal. In the proposed method, the sound signals should first 130 at several points on the engine - preferably on each cylinder 111 - in practice, each with a sensor for several cylinders. Knock or acceleration sensors are used for detection. The sound amplitude is preferably measured 130A Angular synchronous for each complete work cycle and with a resolution Δα of z. B. Δα = 1 ° or Δα = 0.1 ° crankshaft angle 130W . The sound signal 130 in the present case has some signal components that are to be assessed regularly or as normal, ie within the marking of the characteristic signal components E1 , E3 and E6 , which can also be understood as characteristic signal labels.

A typical sound signal 130 has an in 2A illustrated course in which several characteristic signal parts E1 , E2 , E3 , E4 , E5 , E6 are marked. Eg represents the sound signal 130 within the marking of the characteristic signal part E1 the impact noise of the closing exhaust valve as a first event, which is a sound event with the characteristic signal part E1 is recognizable. The sound signal within the marking of the characteristic signal part E3 represents the impact noise of the closing inlet valve as a third event, which is a sound event with the characteristic signal part E3 is recognizable. The sound signal within the marking of the characteristic signal part E4 represents the combustion noise as a fourth event, which is a sound event with the characteristic signal part E4 is recognizable. The sound signal within the marking of the characteristic signal part E6 In the normal state of the engine, it should not have any signal components with significant intensity; So a background can be a sound event with the signal part which is also characteristic in this respect E6 (here background / background noise) can be assigned to the normal state of the engine as the sixth event.

The proposed method now detects areas of the signal with significant signal intensity deviating therefrom with corresponding characteristic signal parts and identifies these as further "events" 131 . This detection is in 2 B shown as an example. There are - with the aforementioned events - six events 131 due to the characteristic signal parts E1 , E2 , E3 , E4 , E5 , E6 ( E5 , E6 here abnormal "real" sound events above the background / background noise) detected. This takes place in the context of an amplitude analysis, which in the present case provides a threshold value analysis explained below with reference to a number of threshold values.

The detection can e.g. B. by recognizing local maxima in the sound signal 130 which exceed a certain threshold So (above the "background / background noise" of the signal). The threshold So is indicated here as the upper dash-dotted line. The sound signal can be used before the local maxima are recognized 130 be rectified, for example, ie the sign of the negative signal components of the sound signal 130 is inverted and it can be in the said filter 160 a filtering of the sound signal 130 be provided about the strong fluctuations in the sound signal 130 to smooth.

• - eine erste Metrik 142.M1 in Form einer Lage der dem Ereignis 131 zugordneten Signalkennzeichnung 141 innerhalb des Arbeitsspiels, z.B. definiert durch einen Kenngrößenparameter in Form einer Position des lokalen Maximums auf der horizontalen Achse, z. B. für das erste Ereignis 131 beim Kurbelwellenwinkel α =43 °; der genannte Kurbelwellenwinkel α bildet somit vorliegend einen geeigneten Kenngrößenparameter 143 für eine Metrik 142; hier die erste Metrik 142.M1.
• - eine zweite Metrik 142.M2 in Form einer Ausdehnung der dem Ereignis 131 zugordneten Signalkennzeichnung 141, z. B. beschrieben durch einen Kenngrößenparameter in Form eines Winkelbereichs, den das Ereignis 131 abdeckt. Dieser kann z. B. aufgefasst werden als der Bereich, in dem die Signalwerte bzw. ihr Betrag eine zweite Schwelle Su mehrheitlich nicht unterschreiten, wobei diese zweite Schwelle Su hier als untere gestrichelte Linie dargestellt ist. Der Winkelbereich kann beschrieben werden durch Kenngrößenparameter in Form eines Startpunkts und eines Endpunkts des Ereignisses 131 oder durch seine Breite, z. B. für das vierte Ereignis 131, das durch den charakteristischen Signalteil E4 erkennbar ist, durch den Startpunkt beim Kurbelwellenwinkel α_S =355° und den Endpunkt beim Kurbelwellenwinkel α_E =485° bzw. durch die Breite, die gegeben ist durch einen Bereich von Kurbelwellenwinkeln 485° - 355° = 130°; die genannten Kurbelwellenwinkel α_S und α_E bilden somit vorliegend geeignete Kenngrößenparameter 143 für eine Metrik 142; hier die zweite Metrik 142.M2.
• - eine dritte Metrik 142.M3 in Form einer die Energie des Ereignisses 131 charakterisierende Größe ist z. B. beschrieben durch Kenngrößenparameter in Form einer Summe aller quadrierten Abtastwerte des Signals zwischen α_S und α_E. Aus Gründen der Ersparnis von Signalverarbeitungsaufwand kann die Quadratur der Abtastwerte durch eine Absolutwertbildung (weglassen des Vorzeichens) ersetzt werden; die genannte Summe aller quadrierten Abtastwerte des Signals zwischen Kurbelwellenwinkel α_S und α_E Kurbelwellenwinkel bilden somit vorliegend geeignete Kenngrößenparameter 143 für eine Metrik 142; hier die dritte Metrik 142.M3.Weitere Metriken 142.M1, 142.M2, 142.M3, ... 142.Mn können vorgesehen sein; die Aufzählung ist nicht abschließend.

For each identified event 131 can now use metrics 142 be determined who the event 131 based on its signal identification 141 characterize in more detail; this metric 142 each characterizes the corresponding signal identification 141 . In particular the following - -in 4th as part of an event vector I according to a first embodiment - metrics shown 142 are suggested:

• - a first metric 142.M1 in the form of a location of the event 131 assigned signal identification 141 within the work cycle, for example defined by a characteristic parameter in the form of a position of the local maximum on the horizontal axis, e.g. B. for the first event 131 at the crankshaft angle α = 43 °; the crankshaft angle α mentioned thus forms a suitable characteristic parameter in the present case 143 for a metric 142 ; here the first metric 142.M1 .
• - a second metric 142.M2 in the form of an extension of the event 131 assigned signal identification 141 , e.g. B. described by a characteristic parameter in the form of an angular range that the event 131 covers. This can e.g. B. can be understood as the range in which the signal values or their magnitude do not mostly fall below a second threshold Su, this second threshold Su being shown here as a lower dashed line. The angular range can be described by characteristic parameters in the form of a starting point and an end point of the event 131 or by its width, e.g. B. for the fourth event 131 , that by the characteristic signal part E4 can be recognized by the starting point at crankshaft angle α _S = 355 ° and the end point at crankshaft angle α _E = 485 ° or by the width, which is given by a range of crankshaft angles 485 ° - 355 ° = 130 °; the crankshaft angles α _S and α _{E mentioned} thus form suitable parameters in the present case 143 for a metric 142 ; here is the second metric 142.M2 .
• - a third metric 142.M3 in the form of the energy of the event 131 characterizing variable is z. B. described by parameter parameters in the form of a sum of all squared samples of the signal between α _S and α _E. In order to save signal processing effort, the quadrature of the sampled values can be replaced by an absolute value formation (omitting the sign); the stated sum of all squared sampled values of the signal between crankshaft angle α _S and α _E crankshaft angle thus form suitable parameter parameters in the present case 143 for a metric 142 ; here is the third metric 142.M3 .Other metrics 142.M1 , 142.M2 , 142.M3 , ... 142 Mn can be provided; the list is not exhaustive.

Based on 3A is within the scope of an exemplary embodiment of a second variant a second method for identifying events 131 explains what a second preferred alternative for identifying the events using the characteristic signal parts E1 , E2 , E3 , E4 , E5 , E6 offers - this can be implemented as follows. There is also a typical sound signal 130 of the 2A gained history of a cumulative sum S130 the absolute values of the sound signal 130 in 3A shown, in which in turn several characteristic signal parts E1 , E2 , E3 , E4 , E5 , E6 are marked; this time in terms of the cumulative total S130 the absolute values (all negative signs are inverted) of the sound signal 130 . This so-called sum signal S130 of the sound signal 130 has a course like that in 3A is shown by way of example as a solid line in comparison with reference straight lines shown in dashed lines - with respect to the reference straight line shown in dashed lines, a characteristic signal part can be E1 , E2 , E3 , E4 , E5 , E6 be recognized, so whether an event 131 present. While the reference line R1 , R2 , R3 , R4 (Not shown), R5 on the step-like plateaus P1 , P2 , P3 , P4 , P5 of the sum signal S130 of the sound signal 130 align parallel, runs in the case of a characteristic signal part E1 , E2 , E3 , E4 , E5 , E6 the cumulative signal S130 no longer parallel to the reference line R1 , R2 , R3 , R4 (Not shown), R5 . An area of a non-parallel part of the signal E1 , E2 , E3 , E4 , E5 , E6 shows an area of an event 131 - it runs in 3A sum signal of the sound signal shown 130 again parallel to the reference line, there is no event. In the example there are again six events E1 ... E6 shown their metric 142 is marked by means of characteristic parameters as the beginning and end each by a vertical dotted line. An algorithm that uses this principle to generate the metrics M142.M1, M142.M2, M142.M3 of the events in the second variant --in 4th Shown as part of an event vector II according to a second embodiment - determined can be implemented as follows. First of all, the reference straight line is taken offline (e.g. by means of signals evaluated a priori) R1 , R2 , R3 , R4 (Not shown), R5 on the step-like plateaus P1 , P2 , P3 , P4 , P5 of the sum signal S130 of the sound signal 130 determined by adopting their slope parallel and at the level of the plateaus - the step-like plateaus P1 , P2 , P3 , P4 , P5 of the sum signal S130 are using the reference line R1 , R2 , R3 , R4 (Not shown), R5 extrapolated. The distance A (shown as an example in 3A) preferably each point of the cumulative sum signal S130 to the reference line R1 , R2 , R3 , R4 (Not shown), R5 also over the step-like plateaus P1 , P2 , P3 , P4 , P5 of the sum signal S130 determined by simple subtraction. If necessary, the accumulated sum signal S130 before and / or after the summation still filtered in an above-mentioned filter 160 . If the distance to the reference line remains constant, there is no event. If the distance to a reference line increases R1 , R2 , R3 , R4 (Not shown), R5 beyond a parameterizable threshold value So, the start of an event 131 - a signal part E1 , E2 , E3 , E4 , E5 , E6 as one of the event 131 assigned signal identification 141-- marked. If the change in distance falls below a further parameterizable threshold value Su, the end of the event is 131 marked.

• - eine Lage L1: Punkt oder Wertepaar, hier also Wert auf der horizontalen Achse, der zu der größten Abstandsvergrößerung (Punkt maximaler Steigung oder Intensitätsmaximum) gehört, die in der Dauer des Ereignisses, also dem im Schallsignal erkennbaren Schallereignis zu dem tatsächlichen Ereignis am Motor, auftritt,
• - eine Lage L2: Punkt oder Wertepaar, hier also Wert auf der horizontalen Achse, der zu der Mitte zwischen den Werten zu Beginn und Ende des Ereignisses, also dem im Schallsignal erkennbaren Schallereignis zu dem tatsächlichen Ereignis am Motor, auf der vertikalen Achse gehört,
• - eine Lage L3: Punkt oder Wertepaar, hier also Mitte zwischen Beginn und Ende des Ereignisses, also dem im Schallsignal erkennbaren Schallereignis zu dem tatsächlichen Ereignis am Motor, auf der horizontalen Achse.

As in 3B a first metric can be shown 142.M1 in the form of a location of the event 131 for this approach z. B. can be defined as

• - a location L1 : Point or pair of values, in this case the value on the horizontal axis that belongs to the greatest increase in distance (point of maximum slope or maximum intensity) that occurs in the duration of the event, i.e. the sound event recognizable in the sound signal to the actual event on the engine,
• - a location L2 : Point or pair of values, in this case the value on the horizontal axis that belongs to the middle between the values at the beginning and end of the event, i.e. the sound event recognizable in the sound signal for the actual event on the engine, on the vertical axis,
• - a location L3 : Point or pair of values, in this case the middle between the beginning and the end of the event, i.e. the sound event recognizable in the sound signal to the actual event on the engine, on the horizontal axis.

The second metric 142.M2 in the form of an extension, analogous to the first suggestion, is defined as the difference between the beginning and the end of the event.

The third metric 142.M3 in the form of an energy of the event 131 Descriptive quantity is defined as the difference between the values at the beginning and the end of the event on the vertical axis.

More metrics 142.M1 , 142.M2 , 142.M3 , ... 142 Mn can be provided; the list is not exhaustive.

In the context of both exemplary embodiments of the first and second variant, further metrics 142 Mn be provided to describe the shape of the signals for the individual events in more detail and to further delimit them, e.g. B. the rate of increase of the signal intensity or the like is conceivable for both approaches. In particular, the sum of all squared samples or, optionally, the sum of the absolute values of the entire sound signal 130 over a work cycle, i.e. again a sum signal S130 can be used as an interesting global metric 142 Mn serve to estimate whether a significant change has generally occurred in the signal or not.

In particular from the first metric 142.M1 a location of an event - i.e. a signal part E1 , E2 , E3 , E4 , E5 , E6 as one of the event 131 assigned signal identification 141-- can later be inferred to the cause of the event. For example, it is known that for crankshaft angles around 40 ° the impact noise of the closing exhaust valve is to be expected, whereas z. B. in the range of 400 ° the combustion noise of the corresponding cylinder is to be expected.

The metrics 142.M1 , 142.M2 , 142.M3 , ... 142 Mn for each event can now be in an event vector 230 combined and into one evaluation unit 240 as explained above. The evaluation unit 240 can e.g. B. be located centrally at the manufacturer. A regional main computer on a ship is also conceivable, which evaluates the data from all ship engines and makes the results available via the ship's automation system. A regional main computer can also be used in a mine, which evaluates the data from all mining vehicles for the region of the mine. Furthermore, a local evaluation directly on the engine is always conceivable as a special case.

The advantage of identifying events 131 and the description of the events 131 with few metrics 142.M1 , 142.M2 , 142.M3 , ... 142 Mn in relation to a signal part E1 , E2 , E3 , E4 , E5 , E6 as one of the event 131 assigned signal identification 141 makes so essential an event 131 characterizing information is summarized very compactly and can be transmitted to the evaluation unit with little effort (ie little transmission capacity).

For example, in relation to a data reduction, the consideration that in the full expansion stage of the method per work cycle for each cylinder depending on the resolution z. B. 720 to 7200 samples. At 1800 rpm, a work cycle like the one in 2A shown is about 66 ms. If each value is resolved with 32 bits, up to 70 Mbit / s of raw data are generated per second for a 20V motor. Are in the signal z. B. ten events 131 ( E1 ... E10 ) recognized each with four metrics 142.M1 , 142.M2 , 142.M3 , 142.M4 can only be written to 390 kbit / s on - a reduction in the amount of data generated by a factor of one hundred and eighty. In order to save signal processing effort or computing power and transmission bandwidth, it may be possible to dispense with recording every working cycle and only record every second or third etc. working cycle, or the resolution can be reduced.

A data structure made up of event vectors 230 and corresponding correlations points to an event 131 out that with a signal identification 141 in the sound signal 130 or the sum signal A130 or another such raw data signal is recognizable, as well as one of this event 131 in the sound signal 130 , S130 associated event vector 230 . The event vector 230 has a number of the signal identification 141 identifying metrics 142 , 142.2, 142.3 etc. - these are listed here as an example of the spatial expansion and energy of the event. More metrics 142 can also be provided - the list is not limited to the mentioned metric cards. Any of the metrics 142 can be defined by one or more characteristic parameters, some of which are explained in the above exemplary embodiments. For example, for the first metric 142.M1 In the present case, a characteristic variable parameter is specified in the form of the position as a rotational angle position in a working cycle, namely the crankshaft angle. Regarding the extent as a metric 142.M2 In the present case, a characteristic parameter is defined as a range on the crankshaft angle with a lower crankshaft angle and an upper crankshaft angle for the range as a characteristic.

For the energy of the event as a third metric 142.M3 In the present case, a characteristic parameter is defined as an amplitude of the event. The event vector formed in this way 230 with a number of metrics 142 and the associated characteristic parameters thus describe a number of signal identifications 141 each for one event 131 in the sound signal 130 .

Over the service life of the engine 110 a trend development, pattern recognition or similar development for an event identified in this way can now be performed on the basis of the development of the event vector 230 fixed, monitored and a cause, measure or other information for the operation of the engine 110 be assigned. A corresponding list 270 linked to causes and measures can be found in the list 270 be stored and to control the internal combustion engine on the engine 110 fed back what is in 1 is shown schematically.

The to the evaluation unit 240 as the central point of the evaluation device 200 transmitted event vectors 230 are then evaluated there. The event vectors 230 form patterns which, as far as known, form a list 270 can be stored and clearly different causes and thus different errors can be assigned via the list 270 . The change in individual metrics in certain event vectors enables the development of an event over time to be observed, i.e. - in physical terms - the change in a noise component over time. For example, from the observation of the temporal development or pattern of the metrics 142 in the event vectors 230 a trend analysis or (e.g. by simple extrapolation) a prediction of the future development of an error can be derived.

For example, from the pattern of an event vector 230 clearly recognized that this event vector 230 describes the impact sound of the exhaust valve. By observing the location metric as the first metric 142.M1 A change in the valve clearance can be determined over time. By extrapolation, z. For example, it can be estimated when the valve clearance will exceed a predetermined value and has to be readjusted by service personnel. As a result, for example, regular (and often superfluous) checking of the valve clearance through manual measurements on site by service personnel is superfluous.

5 shows a flow chart for a method 1 according to the invention for monitoring a drive unit 100 as previously described, the monitoring being a sound monitoring 10 the rotating drive component is carried out by means of acoustic sensors on the drive unit. As part of sound monitoring 10 will be a first step 20th a sound signal of the rotating drive component is detected by means of the acoustic sensor system and then made available for evaluation for an evaluation device 30th . By means of a signal analysis of the sound signal in a sound signal analysis unit, a number of signal identifications is determined from the sound signal 40. According to the concept of the invention, provision is made for a signal identification to be assigned to an event to be characterized in the signal analysis unit by means of at least one metric 41, the metric by at least a characteristic parameter is defined 42 that attaches the event to the sound signal. The metric is then transferred to the signal analysis unit 50 processed to an evaluation unit of an evaluation device 51 and provided 52 .

Another evaluation 60 with an evaluation algorithm takes place in the evaluation unit 240 and the result of a pattern recognition or trend analysis is based on a list 270 juxtaposed with a catalog of measures or causes 70. In this way, as part of a feedback 80 The operation of the internal combustion engine can be influenced 100 for controlling 81 the same or for the transmission of tax information or other additional information 82 to the operator or the operator of the drive unit.

List of reference symbols

1001

system

1000

combination

142.M1, 142.M2, ... 142.Mn

Number of metrics

140.1, 140.2 ... 140.n

Number of signal analysis units

100.1, 100.2 ... 100.n

Number of internal combustion engines

300

Wireless connection

270

list

260

Number of evaluation modules

250

Receiving device

240

Evaluation unit

230

Event vector

220

Storage

200

Evaluation device

160

filter

150

Transfer device

144

Labelling

143

Parameter

142

Metric

141

Signal identification

140

Signal analysis unit

131

event

130

Sound signal

130A

Sound amplitude

130W

Crank angle α

121

acoustic sensor

120

acoustic sensors

114

crankshaft

113

camshaft

112

Valve

111

cylinder

110

engine

101

farm

100

Internal combustion engine

82

Transmission of tax information

81

Control of the internal combustion engine

80

Feedback

70

Comparison of the metrics

60

evaluation

52

Provision of the metric

51

Preparation of the metric

50

Transfer of the metric

42

Definition of the metric

30th

evaluation

20th

first step

10

Sound monitoring

A130

Sum signal

E1, E2, ... En

characteristic signal parts or signal identification

L1, L2, ... Ln

location

P1, P2, ... Pn

step-like plateau

R1, R2, ... Rn

Reference line

S130

cumulative total

QUOTES INCLUDED IN THE DESCRIPTION

This list of the documents listed by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent literature cited

• US 5230316 [0006]
• US 4669315 A [0007]
• US 5157610 A [0010]
• JP 2130429 A [0011]

Claims (18)

Method (1) for monitoring a drive unit with a rotating drive component, in particular a motor (110) of an internal combustion engine (100) with a crankshaft, the monitoring being a sound monitoring (10) of the rotating drive component by means of an acoustic sensor system (120) on the drive unit takes place, and within the scope of the sound monitoring (10): - a sound signal (130) of the rotating drive component is detected by means of the acoustic sensor system (120) and provided for evaluation for an evaluation device (200), - by means of a signal analysis of the sound signal in a signal analysis unit ( 140) a number of signal identifications (141) is determined from the sound signal (130), characterized in that - a number of events (131) is identified by means of a number of signal identifications (141) in the signal analysis unit (140), in particular a signal identifier Assigned event for the rotating drive component et is, and - a signal identifier (141) to be assigned to an event (131) is characterized in the signal analysis unit (140) by means of at least one metric (142), the metric (142) being defined by at least one characteristic parameter (143) which defines the Event on the sound signal (130), and - the metric (142) is processed and provided in the signal analysis unit (140) for transfer to an evaluation unit (240) of an evaluation device (200), in particular the at least one signal analysis unit on the at least one drive unit and the evaluation unit is available remotely from the drive unit, in particular are communicatively connected by a wireless connection. Procedure according to Claim 1 , characterized in that a number of metrics (142.M1, ... 142.Mn) are combined in an event vector (230) or the like event field and assigned to the event. Procedure according to Claim 1 or 2 , characterized in that the sound signal (130) of the rotating drive component is recorded as structure-borne noise on the drive unit with an acoustic sensor. Method according to one of the preceding claims, characterized in that the drive unit has an internal combustion engine and / or a generator, wherein an engine of the internal combustion engine has the at least one rotating drive component as a crankshaft and / or camshaft and / or the generator has the at least one rotating Has drive component as a rotor. Method according to one of the preceding claims, characterized in that the sound signal is made available as transient or frequency-resolved raw data. Method according to one of the preceding claims, characterized in that the raw data have a transient sound signal curve and / or a frequency-resolved sound frequency spectrum for at least part of a working cycle of the rotating drive component. Method according to one of the preceding claims, characterized in that the metric (142) characterizing the signal identification is processed as a representative of the sound signal (130), in particular the signal identification (141), for transfer to the evaluation device (200). Method according to one of the preceding claims, characterized in that the sound signal itself and / or the signal identification, in particular the raw data of the sound signal and / or the signal identification, remain in the signal analysis unit (140), in particular are overwritten in the signal analysis unit. Method according to one of the preceding claims, characterized in that a signal identification to be assigned to an event with at least one metric (142) is selected from a group consisting of position (142.M1), extent (142.M1) and energy (142.M3) , and is characterized on the basis of the sound signal (130). Method according to one of the preceding claims, characterized in that the signal analysis comprises an amplitude analysis of transient or frequency-resolved raw data of the sound signal. Method according to one of the preceding claims, characterized in that the transient raw data of the sound signal (130) is provided as an absolute or relative sound amplitude (130A), in particular the amplitude analysis is a threshold value analysis with reference to a number of threshold values (So, Su ) includes. Method according to one of the preceding claims, characterized in that the transient raw data of the sound signal (130) are provided as cumulative sound amplitude (S130), in particular the amplitude analysis comprises a difference value analysis with respect to a number of references, in particular with respect to a cumulative sound amplitude. The method according to any one of the preceding claims, characterized in that the Characteristic parameter of the metric (142) is designed to recognize an event development on the basis of a pattern development (260) of the characteristic parameter of the metric assigned to the event in a parameter space. Method according to one of the preceding claims, characterized in that a list (270) for causes or measures for the event is assigned and stored to the pattern development of the characteristic variable parameter of the metric assigned to the event. Signal analysis unit (140) of an evaluation device (200) for monitoring a drive unit with a rotating drive component, in particular a motor of an internal combustion engine with a crankshaft, the monitoring being carried out as sound monitoring of the rotating drive component by means of acoustic sensors on the drive unit, and as part of the sound monitoring - a sound signal of the rotating drive component is detected by means of the acoustic sensor system and provided for evaluation for an evaluation device, - a number of signal identifications is determined from the sound signal by means of a signal analysis of the sound signal in a signal analysis unit, characterized in that - a number of events by means of a Number of signal identifications is identified in the signal analysis unit, in particular a signal identification is assigned to an event for the rotating drive component, and one is assigned to an event assigning signal identifier is characterized in the signal analysis unit by means of at least one metric, the metric being defined by at least one characteristic parameter which fixes the event on the sound signal, and - the metric is processed and provided in the signal analysis unit for transfer to an evaluation unit of an evaluation device, in particular the at least one signal analysis unit is available on the at least one drive unit and the evaluation unit is remote from the drive unit, in particular are communicatively connected by a wireless connection. Drive unit with a rotating drive component and a signal analysis unit according to Claim 15 , in particular a motor (110) of an internal combustion engine (100) with a crankshaft, a sound monitoring of the rotating drive component by means of an acoustic sensor system on the drive unit being provided for monitoring. Evaluation device (200) with the signal analysis unit (140) and the evaluation unit (240) designed to carry out the method according to one of the Claims 1 to 14th . System (1001), in particular a fleet or farm, a number of drive units (100.1 ... 100.n) Claim 16 and a number of signal analysis units (140.1 ... 140.n) Claim 15 and the evaluation device (200) Claim 17 .

Images