Switching device event detection system

Abstract

The invention relates to a switchgear event detection system comprising one or more vibratory / acoustic sensors (10), a processing unit (20) and an output unit (30). The one or more vibration / acoustic sensors are configured to acquire one or more vibration / acoustic signals related to the event and continued on a time window, and provide the one or more vibration / acoustic signals to the processing unit; the processing unit is configured to compare at least a portion of the one or more vibrational / acoustic signals with a plurality of reference vibrational / acoustic signals stored in a database (40), wherein the plurality of reference vibrational / acoustic signals are related to N different reference events; the processing unit is configured to classify the event as a particular reference event when a comparative matching degree between at least a portion of the one or more vibrational / acoustic signals associated with the event and one or more reference vibrational / acoustic signals associated with the particular reference event exceeds a first threshold; the output unit is configured to output information about an event.

Description

Technical Field

[0001] The present invention relates to a switchgear event detection system, a switchgear event detection method, and a computer program unit. Background Art

[0002] Multiple events can occur in switchgear, which may have a negative impact on the operation of a substation (high voltage, medium voltage, low voltage). This includes unauthorized manipulation of the switchgear, intrusion of humans or animals into the substation area. There are safety-related events, operation-related events, and environment-related events, such as:

[0003] (1) Safety-related events: Some events in electrical equipment can ultimately lead to catastrophic failures.

[0004] (2) Operation-related events: Certain events may damage the switchgear itself (e.g., incorrect operation of components, animals in the switchgear).

[0005] (3) Environment events: Environment events (e.g., earthquakes) can affect the switchgear.

[0006] It is very important to detect safety-related events, operation-related events, and environment events before they cause harmful consequences and take actions on them. Summary of the Invention

[0007] Therefore, it would be advantageous to have an improved switchgear event detection system.

[0008] The object of the present invention is solved by the subject matter of the independent claims, wherein further embodiments are incorporated into the dependent claims.

[0009] In a first aspect, there is provided a switchgear event detection system, comprising:

[0010] - one or more vibration / acoustic sensors;

[0011] - a processing unit; and

[0012] - an output unit.

[0013] The one or more vibration / acoustic sensors are configured to be mounted in and / or on the switching device. The one or more vibration / acoustic sensors are configured to acquire one or more vibration / acoustic signals related to events inside and / or outside the switching device and in the vicinity of the switching device. The one or more vibration / acoustic signals extend over a time window. The one or more vibration / acoustic sensors are configured to provide the one or more vibration / acoustic signals to a processing unit. The processing unit is configured to compare at least a portion of the one or more vibration / acoustic signals with a plurality of reference vibration / acoustic signals stored in a database. The plurality of reference vibration / acoustic signals are related to N different reference events, where information about the N different reference events is stored in the database. Each of the N different reference events is associated with a different one or more of the plurality of reference vibration / acoustic signals. The processing unit is configured to classify the event as the specific reference event when the degree of match between at least a portion of the one or more vibration / acoustic signals related to the event and one or more of the plurality of reference vibration / acoustic signals associated with a specific reference event among the N different reference events exceeds a first threshold. The output unit is configured to output information about the event, including using the information of the specific reference event stored in the database.

[0014] Thus, vibration and / or acoustic sensors are used to detect and classify events by calculating a feature-based event fingerprint and comparing it with a database. The events are then characterized and classified based on a classification algorithm.

[0015] It should be noted that the one or more vibration / acoustic sensors being configured to be mounted in and / or on the switching device includes the one or more vibration / acoustic sensors being configured to be mounted on components of the switching device (such as circuit breakers, earthing switches, etc.) or mounted to components of the switching device.

[0016] The first threshold can be regarded as an event detection criterion or an event detection metric or an event detection marker. The first threshold can be, for example, a simple threshold based on a single feature, a residual threshold, a threshold applied to a similarity metric / ML algorithm, etc. / a threshold operating on a feature set. The first threshold can depend, for example, on the way the similarity metric is defined.

[0017] Thus, the first threshold is generally a marker for detecting events.

[0018] Thus, these signal characteristics can be described and distinguished by a set of features to characterize event duration, amplitude, repeatability. This results in a set of proposed features:

[0019] Analysis features based on the envelope of the time signal (signal energy, peak energy, signal duration),

[0020] Aligning events to detect temporal patterns by using time compression / stretching methods (such as dynamic time warping and / or circular shifting).

[0021] Frequency and / or time-frequency characteristics, such as specific coefficients or energies from Fourier transform, wavelet transform, or cepstral coefficients.

[0022] Estimation of the excitation frequency by autocorrelation.

[0023] In addition, the original measurement signal itself can also be used as a feature.

[0024] In one example, the N different reference events include at least one reference event inside and / or at least one reference event outside a reference switching device; and / or the N different reference events include reference events inside and / or outside a switching device.

[0025] In other words, one or more switching devices different from the switching device currently being monitored and appropriate vibration and / or acoustic sensors can already have been used, together with information related to what has happened when generating the signal, to generate and establish a fingerprint sensor signal database. Alternatively or additionally, the switching device itself being monitored using vibration and / or acoustic sensors may already have undergone a series of tests and operational runs, together with information related to what has happened when generating the signal, to generate and establish a fingerprint sensor signal database.

[0026] In one example, the database is contained within the processing unit.

[0027] In one example, the database is separate from the processing unit.

[0028] Therefore, the database can be located in the memory local to the processing unit or saved separately. Thus, if saved separately, an entire series of switching devices can access a database that contains common fingerprint sensor data and information related to what has happened when generating the signal fingerprints.

[0029] In one example, the processing unit is configured to classify an event as an unknown event when the degree of comparison match between at least a portion of one or more vibration / acoustic signals related to the event and each of a plurality of reference vibration / acoustic signals is less than a first threshold.

[0030] The database can be accessed, for example, regarding events classified as unknown events to update the database. Thus, for example, unknown events of circuit breakers in different switching devices (which can be circuit breakers of the same or different types) can be used to update the database.

[0031] In one example, the processing unit is configured to classify an event as an unknown event when the degree of match between at least a portion of one or more vibration / acoustic signals associated with the event and each of a plurality of reference vibration / acoustic signals is greater than a second threshold.

[0032] Accordingly, it can be determined that an event of some form has occurred to prevent the system from treating normal background noise as an event that cannot be classified.

[0033] The second threshold is similar to the first threshold in standard form, but its purpose is that not all comparisons below the first threshold are classified as unknown events, otherwise it would mean that background noise would be classified as an unknown event. Thus, the second threshold is only used to provide confidence that an event (rather than just noise) has been detected but cannot be classified as above and is therefore classified as an unknown event.

[0034] In one example, the processing unit is configured to control the output unit to output an alarm when the event is classified as an unknown event.

[0035] In this way, the operator can be aware that an unknown event has occurred and conduct further investigation.

[0036] In one example, the processing unit is configured to determine a severity level, including using information of a specific reference event stored in a database, and wherein the information about the event output by the output unit includes the severity level.

[0037] There can be a whole series of severity levels, for example ranging from low level, medium level, high level to severe level.

[0038] Accordingly, the operator can know that an event has occurred and obtain information about its severity level, so as to be able to take immediate action if necessary.

[0039] In one example, the processing unit is configured to control the output unit to output information related to the event when the event is classified as an unknown event. The system includes an input unit (50), and the input unit is configured to enable the operator to input information about the event, and the information about the event classifies the event. The processing unit is configured to update the database using one or more vibration / acoustic signals related to the event, and the one or more vibration / acoustic signals together with the information about the event that classifies the event are stored as new one or more reference vibration / acoustic signals.

[0040] In this way, a form of feedback loop is provided where a professional operator can update the database using the vibration and / or acoustic information of the fingerprint of an unknown event, but the operator can identify and / or already know what event it is, and then the event becomes a known reference event. When this happens again, the system will identify it and provide information about the event that has occurred. Additionally, the operator can train the classifier in this way by generating an event that initially outputs as unknown by themselves, and then the operator inputs what has happened. Then the operator can repeat the event multiple times, where slightly different vibration and / or acoustic signals are generated, and the system may initially not be able to identify the event and output an unknown event "alert". However, the operator can input that it is the same event and continue to repeat this process until the system can identify the vibration and / or acoustic signals of all variants of the event as the said event.

[0041] In one example, the processing unit is configured to cut each of one or more vibration / acoustic signals associated with an event into multiple segments, each segment extending over a different part of a time window. The processing unit is configured to compare each segment of one or more vibration / acoustic signals with multiple reference vibration / acoustic signals stored in the database. The processing unit is configured to classify the event as a specific reference event among N different reference events when the degree of comparison match between two or more segments of one or more vibration / acoustic signals associated with the event and one or more reference vibration / acoustic signals associated with a specific reference event among the multiple reference vibration / acoustic signals exceeds a first threshold.

[0042] This can better identify and classify events because some events occur within a shorter time range, while some events occur within a longer time range. By segmenting, long events and short events can be identified and classified. Additionally, the same event occurring over a relatively long period of time may have very similar parts in some parts of the time window, while other parts of the signal may be different. By segmenting the signal and comparing each part, when multiple parts of the signal match the relevant parts of the stored fingerprint signal and the degree of match is higher than the threshold, a positive identification can be made, while in the case of the overall signal itself, the degree of match may not be higher than the positive classification threshold.

[0043] It should be noted that the segments extending over different parts of the time window include some segments that can overlap, but do not have to overlap.

[0044] In one example, the processing unit is configured to classify the event as an unknown event when the degree of comparison match between two or more segments of one or more vibration / acoustic signals associated with the event and each of the multiple reference vibration / acoustic signals is less than the first threshold.

[0045] In this way, it can be more surely shown that a certain event is truly unknown.

[0046] For example, the database can be accessed regarding events classified as unknown events to update the database. Thus, for example, unknown events of circuit breakers (which can be circuit breakers of the same or different types) in different switching devices can be used to update the database.

[0047] In one example, the processing unit is configured to classify an event as an unknown event when the degree of matching between two or more segments of one or more vibration / acoustic signals related to the event and each of a plurality of reference vibration / acoustic signals is greater than a second threshold.

[0048] The second threshold is similar to the first threshold in terms of standard form, but its use is that not all comparisons below the first threshold are classified as unknown events, otherwise it would mean that background noise would be classified as an unknown event. Therefore, the second threshold is only used to provide confidence that an event (rather than just noise) has been detected, but it cannot be classified as above, so it is classified as an unknown event.

[0049] In one example, at least one of the plurality of segments continues in a time period different from at least one other of the plurality of segments.

[0050] In a second aspect, there is provided a switching device including the system according to the first aspect.

[0051] In a third aspect, there is provided a method for detecting switching device events, the method including:

[0052] - Obtaining one or more vibration / acoustic signals related to events inside and / or outside the switching device and near the switching device through one or more vibration / acoustic sensors installed in and / or on the switching device, wherein the one or more vibration / acoustic signals continue over a time window;

[0053] - Providing the one or more vibration / acoustic signals to a processing unit through the one or more vibration / acoustic sensors;

[0054] - Comparing at least a part of the one or more vibration / acoustic signals with a plurality of reference vibration / acoustic signals stored in a database (40) by the processing unit, wherein the plurality of reference vibration / acoustic signals are related to N different reference events, and wherein information of the N different reference events is stored in the database, and wherein each of the N different reference events is associated with a different one or more of the plurality of reference vibration / acoustic signals;

[0055] -Classify the event through a processing unit configured to classify the event as the specific reference event when the degree of comparison match between at least a part of one or more vibration / acoustic signals related to the event and one or more reference vibration / acoustic signals among a plurality of reference vibration / acoustic signals associated with a specific reference event among N different reference events exceeds a first threshold; and

[0056] -Output information about the event through an output unit, including using the information of the specific reference event stored in the database.

[0057] In one example, the N different reference events include at least one reference event inside and / or outside at least one reference switching device; and / or, the N different reference events include reference events inside and / or outside the switching device.

[0058] It should be noted that one or more vibration / acoustic sensors installed in and / or on the switching device include installing one or more vibration / acoustic sensors on the components (such as circuit breakers, earthing switches, etc.) of the switching device or installing them onto the components.

[0059] The first threshold can be regarded as an event detection criterion or an event detection index or an event detection marker. The first threshold can be, for example, a simple threshold based on a single feature, a residual threshold, a threshold applied to a similarity metric / ML algorithm, etc. / running on a feature set. The first threshold can depend, for example, on the way the similarity metric is defined.

[0060] Therefore, the first threshold is generally a marker for detecting events.

[0061] Therefore, these signal characteristics can be described and distinguished by a set of features to characterize the event duration, amplitude, and repeatability. This results in a set of proposed features:

[0062] Analysis features based on the time signal envelope (signal energy, peak energy, signal duration),

[0063] Align the events by using time compression / stretching methods (such as dynamic time warping and / or cyclic shift) to detect temporal patterns.

[0064] Frequency and / or time-frequency characteristics, such as specific coefficients or energies from Fourier transform, wavelet transform, or cepstral coefficients.

[0065] Estimation of the excitation frequency through autocorrelation.

[0066] In addition, the original measurement signal itself can also be used as a feature.

[0067] In one example, the database is contained within the processing unit.

[0068] In one example, the database is separate from the processing unit.

[0069] In one example, the method includes classifying, by a processing unit, an event as an unknown event when a comparison match degree between at least a part of one or more vibration / acoustic signals associated with the event and each of a plurality of reference vibration / acoustic signals is less than a first threshold.

[0070] The database can be accessed, for example, regarding events classified as unknown events to update the database. Thus, for example, unknown events of circuit breakers (which can be of the same or different types) in different switchgear can be used to update the database.

[0071] In one example, the method includes classifying, by a processing unit, an event as an unknown event when a comparison match degree between at least a part of one or more vibration / acoustic signals associated with the event and each of a plurality of reference vibration / acoustic signals is greater than a second threshold.

[0072] Thus, it can be determined that an event of a certain form has occurred to prevent the system from regarding normal background noise as constituting an event that cannot be classified itself.

[0073] The second threshold is similar to the first threshold in terms of standard form, but its use is that not all comparisons below the first threshold are classified as unknown events, otherwise it would mean that background noise would be classified as unknown events. Therefore, the second threshold is only used to provide confidence that an event (rather than just noise) has been detected but cannot be classified as above, so it is classified as an unknown event.

[0074] In one example, the method includes controlling, by a processing unit, an output unit to output an alarm when the event has been classified as an unknown event.

[0075] In one example, the processing unit is configured to determine a severity level, including using information of a specific reference event stored in the database, and wherein the information about the event output by the output unit includes the severity level.

[0076] There can be a whole series of severity levels, for example ranging from low level, medium level, high level to severe level.

[0077] Thus, the operator can know that an event has occurred and get information about its severity level, enabling the operator to take immediate action if necessary.

[0078] In one example, the method includes controlling an output unit by a processing unit to output information related to the event when the event is classified as an unknown event, and wherein the method includes an operator inputting information about the event through an input unit, wherein the information about the event classifies the event, and wherein the method includes: updating a database by the processing unit using one or more vibration / acoustic signals related to the event, and the one or more vibration / acoustic signals are stored together with the information about the event that classifies the event as one or more new reference vibration / acoustic signals.

[0079] In one example, the method includes the processing unit segmenting each of one or more vibration / acoustic signals related to the event into a plurality of segments, each segment extending over a different part of a time window, and the method includes the processing unit comparing each segment of the one or more vibration / acoustic signals with a plurality of reference vibration / acoustic signals stored in the database, and the method includes when the comparison match degree between two or more segments of the one or more vibration / acoustic signals related to the event and one or more reference vibration / acoustic signals associated with a specific reference event among the plurality of reference vibration / acoustic signals exceeds a first threshold, the processing unit classifying the event as the specific reference event among N different reference events.

[0080] It should be noted that the segments each extending over a different part of the time window include that some segments can overlap, but do not have to overlap.

[0081] In one example, the method includes the processing unit classifying the event as an unknown event when the comparison match degree between two or more segments of one or more vibration / acoustic signals related to the event and each of the plurality of reference vibration / acoustic signals is less than the first threshold.

[0082] The database can be accessed, for example, regarding events classified as unknown events to update the database, so that unknown events of circuit breakers in different switchgears (these circuit breakers can be of the same or different types) can be used to update the database.

[0083] In one example, the method includes the processing unit classifying the event as an unknown event when the comparison match degree between two or more segments of one or more vibration / acoustic signals related to the event and each of the plurality of reference vibration / acoustic signals is greater than a second threshold.

[0084] The second threshold is similar to the first threshold in terms of its standard form, but its purpose is that not all comparisons below the first threshold are classified as unknown events, otherwise it would mean that background noise would be classified as an unknown event. Thus, the second threshold is only used to provide confidence that an event has been detected (and not just noise), but it cannot be classified as above, and thus it is classified as an unknown event.

[0085] In one example, at least one of the plurality of segments extends over a different time period than at least one other of the plurality of segments.

[0086] In a fourth aspect, there is provided a computer program component for controlling a system, the computer program component being configured to perform the method according to the third aspect when executed by a processor.

[0087] In one example, there is provided a computer-readable medium storing the computer component of the fourth aspect.

[0088] The computer program component can, for example, be a software program, but can also be an FPGA, a PLD or any other suitable digital device.

[0089] The above aspects and examples will become apparent and be elucidated with reference to the embodiments described below. Description of the Drawings

[0090] Exemplary embodiments will be described below with reference to the following drawings:

[0091] Figure 1 An exemplary switchgear event detection system is shown;

[0092] Figure 2 An exemplary switchgear event detection system is shown;

[0093] Figure 3 A schematic diagram of creating or generating a reference vibration / acoustic signal or fingerprint is shown;

[0094] Figure 4 A detailed example of the processing of the time vibration / acoustic signal of an event is shown. Feature calculation / classification; and

[0095] Figure 5 Examples of different sensor placements inside or outside the switchgear are shown, and many other locations can also be used for sensor placement. Detailed Description of the Invention

[0096] Figure 1 An exemplary switchgear event detection system is shown, which includes:

[0097] - one or more vibration / acoustic sensors 10;

[0098] - Processing unit 20; and

[0099] - Output unit 30.

[0100] The one or more vibration / acoustic sensors are configured to be mounted in and / or on the switching device. The one or more vibration / acoustic sensors are configured to acquire one or more vibration / acoustic signals related to events inside and / or outside the switching device and in the vicinity of the switching device. The one or more vibration / acoustic signals extend over a time window. The one or more vibration / acoustic sensors are configured to provide the one or more vibration / acoustic signals to the processing unit. The processing unit is configured to compare at least a portion of the one or more vibration / acoustic signals with a plurality of reference vibration / acoustic signals stored in a database 40. The plurality of reference vibration / acoustic signals are related to N different reference events, where information about the N different reference events is stored in the database. Each of the N different reference events is associated with a different one or more of the plurality of reference vibration / acoustic signals. The processing unit is configured to classify an event as the specific reference event when the degree of match between at least a portion of the one or more vibration / acoustic signals related to the event and one or more of the reference vibration / acoustic signals associated with the specific reference event among the N different reference events and the plurality of reference vibration / acoustic signals exceeds a first threshold. The output unit is configured to output information about the event, including using the information of the specific reference event stored in the database.

[0101] Thus, the vibration and / or acoustic sensors are used to detect and classify events by calculating a feature-based event fingerprint and comparing it with a database. The events are then characterized and classified based on a classification algorithm.

[0102] It should be noted that the one or more vibration / acoustic sensors being configured to be mounted in and / or on the switching device includes the one or more vibration / acoustic sensors being configured to be mounted on components (such as circuit breakers, earthing switches, etc.) of the switching device or being mounted to components of the switching device.

[0103] The first threshold can be regarded as an event detection criterion or an event detection metric or an event detection marker. The first threshold can be, for example, a simple threshold based on a single feature, a residual threshold, a threshold applied to a similarity metric / ML algorithm, etc. / a threshold operating on a feature set. The first threshold can depend, for example, on the way the similarity metric is defined.

[0104] Thus, the first threshold is generally a marker for detecting events.

[0105] Therefore, these signal characteristics can be described and distinguished by a set of features to characterize event duration, amplitude, and repeatability. This results in a set of proposed features:

[0106] Analysis features based on the time signal envelope (signal energy, peak energy, signal duration),

[0107] Aligning events by using time compression / stretching methods (such as dynamic time warping and / or cyclic shift) to detect temporal patterns.

[0108] Frequency and / or time-frequency characteristics, such as specific coefficients or energies from Fourier transform, wavelet transform, or cepstral coefficients.

[0109] Estimation of the excitation frequency by autocorrelation.

[0110] In addition, the original measurement signal itself can also be used as a feature.

[0111] In one example, the N different reference events include reference events inside at least one reference switching device and / or outside at least one reference switching device; and / or, the N different reference events include reference events inside and / or outside the switching device.

[0112] In other words, one or more switching devices different from the switching device currently being monitored and appropriate vibration and / or acoustic sensors may have been used, together with information related to what has occurred when generating the signal, to generate and establish a fingerprint sensor signal database. Alternatively or additionally, the current switching device itself monitored by vibration and / or acoustic sensors may have undergone a series of tests and operational runs, together with information related to what has occurred when generating the signal, to generate and establish a fingerprint sensor signal database.

[0113] In one example, the database is contained within the processing unit.

[0114] In one example, the database is separate from the processing unit.

[0115] Therefore, the database can be located in the local memory of the processing unit or saved separately. Thus, if saved separately, an entire series of switching devices can access a database that contains common fingerprint sensor data and information related to what has occurred when generating the signal fingerprint.

[0116] In one example, the processing unit is configured to classify an event as an unknown event when the degree of comparison match between at least a portion of the one or more vibration / acoustic signals related to the event and each of a plurality of reference vibration / acoustic signals is less than a first threshold.

[0117] The database can be accessed, for example, with respect to events classified as unknown events to update the database. Thus, for example, unknown events of circuit breakers in different switching devices (these circuit breakers can be of the same or different types) can update the database.

[0118] In one example, the processing unit is configured to classify an event as an unknown event when the degree of match between at least a portion of one or more vibration / acoustic signals associated with the event and each of a plurality of reference vibration / acoustic signals is greater than a second threshold.

[0119] Thus, it can be determined that an event of a certain form has occurred to prevent the system from treating normal background noise as an event that cannot be classified.

[0120] The second threshold is similar to the first threshold in terms of standard form, but its use is that not all comparisons below the first threshold are classified as unknown events, otherwise it would mean that background noise would be classified as an unknown event. Thus, the second threshold is only used to provide confidence that an event (rather than just noise) has been detected but cannot be classified as above, and thus is classified as an unknown event.

[0121] In one example, the processing unit is configured to control the output unit to output an alarm when the event is classified as an unknown event.

[0122] In this way, the operator can be aware that an unknown event has occurred and further investigate.

[0123] In one example, the processing unit is configured to determine a severity level, including using information of a specific reference event stored in the database, and wherein the information about the event output by the output unit includes the severity level.

[0124] There can be a whole series of severity levels, for example ranging from low level, medium level, high level to severe level.

[0125] Thus, the operator can know that an event has occurred and obtain information about its severity level, so as to be able to take immediate action if necessary.

[0126] In one example, the processing unit is configured to control the output unit to output information related to the event when the event is classified as an unknown event. The system includes an input unit 50, and the input unit is configured to enable the operator to input information about the event, and the information about the event classifies the event. The processing unit is configured to update the database using one or more vibration / acoustic signals related to the event, and the one or more vibration / acoustic signals together with the information about the event that classifies the event are stored as new one or more reference vibration / acoustic signals.

[0127] In this way, a form of feedback loop is provided, in which a professional operator can update the database using the vibration and / or acoustic information of the fingerprint of an unknown event, but the operator can identify and / or already know what event it is, and then this event becomes a known reference event. When this happens again, the system will identify it and provide information about the event that has occurred. In addition, the operator can train the classifier in this way by generating an event whose initial output is unknown by himself, and then the operator inputs what has happened. Then the operator can repeat this event multiple times, in which slightly different vibration and / or acoustic signals will be generated, and the system may not be able to identify this event initially and output an unknown event "alarm". However, the operator can input that it is the same event and continue to repeat this process until the system can identify the vibration and / or acoustic signals of all variants of this event as the said event.

[0128] In one example, the processing unit is configured to cut each of one or more vibration / acoustic signals related to an event into multiple segments, each segment continuing on a different part of the time window. The processing unit is configured to compare each segment of one or more vibration / acoustic signals with a plurality of reference vibration / acoustic signals stored in the database 40. The processing unit is configured to classify the event as a specific reference event among N different reference events when the comparison matching degree between two or more segments of one or more vibration / acoustic signals related to the event and one or more reference vibration / acoustic signals associated with a specific reference event among the plurality of reference vibration / acoustic signals exceeds a first threshold.

[0129] This can better identify and classify events because some events occur in a shorter time range, while some events occur in a longer time range. By segmenting, long events and short events can be identified and classified. In addition, the same event occurring over a relatively long period of time may have very similar parts in some parts of the time window, while other parts of the signal may be different. By segmenting the signal and comparing each part, a positive identification can be made when multiple parts of the signal match the relevant parts of the stored fingerprint signal and the matching degree is higher than the threshold. While in the case of the overall signal itself, the matching degree may not be higher than this positive classification threshold.

[0130] It should be noted that each of the segments continuing on different parts of the time window includes some segments that can overlap, but do not have to overlap.

[0131] In one example, the processing unit is configured to classify the event as an unknown event when the comparison matching degree between two or more segments of one or more vibration / acoustic signals related to the event and each of the plurality of reference vibration / acoustic signals is less than the first threshold.

[0132] In this way, it can be more surely indicated that an event is indeed unknown.

[0133] The database can be accessed, for example, regarding events classified as unknown events to update the database. Thus, for example, unknown events of circuit breakers (which can be circuit breakers of the same or different types) in different switching devices can update the database.

[0134] In one example, the processing unit is configured to classify an event as an unknown event when the degree of matching between two or more segments of one or more vibration / acoustic signals related to the event and each of a plurality of reference vibration / acoustic signals is greater than a second threshold.

[0135] The second threshold is similar to the first threshold in terms of the standard form, but its use is that not all comparisons below the first threshold are classified as unknown events, otherwise it would mean that background noise would be classified as an unknown event. Therefore, the second threshold is only used to provide confidence that an event (rather than just noise) is detected, but it cannot be classified as above, so it is classified as an unknown event.

[0136] In one example, at least one of the plurality of segments extends over a time period different from at least one other of the plurality of segments.

[0137] It is obvious from the above that the switching device can adopt the system as described in the reference Figure 1 as described.

[0138] Similarly, it is obvious from the above description of the system described in the reference Figure 1 that a related method for detecting switching device events may include:

[0139] - Obtaining one or more vibration / acoustic signals related to events inside and / or outside and near the switching device through one or more vibration / acoustic sensors installed in and / or on the switching device, wherein the one or more vibration / acoustic signals extend over a time window;

[0140] - Providing the one or more vibration / acoustic signals to the processing unit through the one or more vibration / acoustic sensors;

[0141] - Comparing at least a part of the one or more vibration / acoustic signals with a plurality of reference vibration / acoustic signals stored in the database by the processing unit, wherein the plurality of reference vibration / acoustic signals are related to N different reference events, and wherein information of the N different reference events is stored in the database, and wherein each of the N different reference events is associated with a different one or more of the plurality of reference vibration / acoustic signals;

[0142] - Classify by a processing unit configured to classify an event as the specific reference event when a comparison match degree between at least a part of one or more vibration / acoustic signals related to the event and one or more reference vibration / acoustic signals among a plurality of reference vibration / acoustic signals associated with a specific reference event among N different reference events exceeds a first threshold; and

[0143] - Output information about the event through an output unit, including using information of the specific reference event stored in a database.

[0144] In one example, the N different reference events include at least one reference event inside and / or at least one reference event outside a reference switching device; and / or, the N different reference events include reference events inside and / or outside a switching device.

[0145] It should be noted that one or more vibration / acoustic sensors installed in and / or on a switching device include installing one or more vibration / acoustic sensors on components (such as a circuit breaker, a grounding switch, etc.) of the switching device or installing them onto the components.

[0146] The first threshold can be regarded as an event detection criterion or an event detection index or an event detection marker. The first threshold can be, for example, a simple threshold based on a single feature, a residual threshold, a threshold applied to a similarity metric / ML algorithm, etc. / running on a feature set. The first threshold can depend, for example, on the way the similarity metric is defined.

[0147] Therefore, the first threshold is generally a marker for detecting an event.

[0148] Therefore, these signal characteristics can be described and distinguished by a set of features to characterize event duration, amplitude, and repeatability. This results in a set of proposed features:

[0149] Analysis features based on the time signal envelope (signal energy, peak energy, signal duration),

[0150] Align events by using time compression / stretching methods (such as dynamic time warping and / or cyclic shift) to detect temporal patterns.

[0151] Frequency and / or time-frequency characteristics, such as specific coefficients or energies from Fourier transform, wavelet transform, or cepstral coefficients.

[0152] Estimation of excitation frequency by autocorrelation.

[0153] In addition, the original measurement signal itself can also be used as a feature.

[0154] In one example, the database is included within the processing unit.

[0155] In one example, the database is separate from the processing unit.

[0156] In one example, the method includes classifying, by the processing unit, an event as an unknown event when a degree of match between at least a portion of one or more vibration / acoustic signals associated with the event and each of a plurality of reference vibration / acoustic signals is less than a first threshold.

[0157] The database can be accessed, for example, regarding events classified as unknown events to update the database. Thus, for example, unknown events of circuit breakers (which can be of the same or different types) in different switchgear can be used to update the database.

[0158] In one example, the method includes classifying, by the processing unit, an event as an unknown event when a degree of match between at least a portion of one or more vibration / acoustic signals associated with the event and each of a plurality of reference vibration / acoustic signals is greater than a second threshold.

[0159] Thus, it can be determined that an event of a certain form has occurred to prevent the system from treating normal background noise as an event that cannot be classified.

[0160] The second threshold is similar to the first threshold in terms of standard form, but its use is such that not all comparisons below the first threshold are classified as unknown events, otherwise it would mean that background noise would be classified as an unknown event. Thus, the second threshold is only used to provide confidence that an event (rather than just noise) has been detected but cannot be classified as above and is therefore classified as an unknown event.

[0161] In one example, the method includes controlling, by the processing unit, an output unit to output an alarm when the event has been classified as an unknown event.

[0162] In one example, the processing unit is configured to determine a severity level, including using information of a specific reference event stored in the database, and wherein the information about the event output by the output unit includes the severity level.

[0163] There can be a whole series of severity levels, for example ranging from low level, medium level, high level to severe level.

[0164] Thus, the operator can know that an event has occurred and obtain information about its severity level, enabling the operator to take immediate action if necessary.

[0165] In one example, the method includes controlling an output unit by a processing unit to output information related to an event when the event is classified as an unknown event, and wherein the method includes an operator inputting information about the event through an input unit, wherein the information about the event classifies the event, and wherein the method includes: updating a database by the processing unit using one or more vibration / acoustic signals related to the event, and the one or more vibration / acoustic signals together with the information about the event that classifies the event are stored as new one or more reference vibration / acoustic signals.

[0166] In one example, the method includes the processing unit segmenting each of one or more vibration / acoustic signals related to the event into a plurality of segments, each segment extending over a different part of a time window, and the method includes the processing unit comparing each segment of the one or more vibration / acoustic signals with a plurality of reference vibration / acoustic signals stored in the database, and the method includes when the comparison match degree between two or more segments of the one or more vibration / acoustic signals related to the event and one or more reference vibration / acoustic signals associated with a specific reference event among the plurality of reference vibration / acoustic signals exceeds a first threshold, the processing unit classifying the event as the specific reference event among N different reference events.

[0167] It should be noted that the segments each extending over a different part of the time window include that some segments may overlap, but do not have to overlap.

[0168] In one example, the method includes when the comparison match degree between two or more segments of one or more vibration / acoustic signals related to an event and each of the plurality of reference vibration / acoustic signals is less than the first threshold, the processing unit classifying the event as an unknown event.

[0169] The database can be accessed, for example, regarding events classified as unknown events to update the database, so that unknown events of circuit breakers in different switching devices (these circuit breakers can be of the same or different types) can be used to update the database.

[0170] In one example, the method includes when the comparison match degree between two or more segments of one or more vibration / acoustic signals related to the event and each of the plurality of reference vibration / acoustic signals is greater than a second threshold, the processing unit classifying the event as an unknown event.

[0171] The second threshold is similar to the first threshold in terms of its standard form, but its purpose is such that not all comparisons below the first threshold are classified as unknown events, otherwise it would mean that background noise would be classified as an unknown event. Thus, the second threshold is only used to provide confidence that an event (rather than just noise) has been detected, but it cannot be classified as above, and thus it is classified as an unknown event.

[0172] In one example, at least one of the plurality of segments extends over a different time period than at least one other of the plurality of segments.

[0173] The inventors recognize that the current situation can be improved. So far, only events related to the main function of the switchgear have been tracked by using specific sensors. Events such as panel interlocks, door closing / opening, racking-in / out, shutter closing / opening, etc. are all performed without any sensor-based feedback to the operator. This also applies to detecting animals (e.g., snakes, rats, cats, dogs, etc.) that may break into the switchgear and cause harm.

[0174] The inventors recognize that vibration sensors and / or acoustic sensors can be used in new ways to detect and identify the above events.

[0175] In addition, vibration sensors and / or acoustic sensors can be used in new ways to detect and identify other events, such as detecting personnel in the area near the switchgear panel or in the switchgear, and this detection is also important for improving safety. These events can have a negative impact on the operation of the substation (high voltage, medium voltage, low voltage), such as unauthorized manipulation of the switchgear, human or animal intrusion into the substation area. Now these events can be detected and identified. Additionally, environmental events such as seismic detection or strong vibrations from neighboring machines or vehicles can also be detected and identified and used for environmental history files to link certain events to certain circuit breaks and faults in the switchgear.

[0176] The new technology involves using vibration and / or acoustic sensors to detect and classify events by calculating feature-based event fingerprints and comparing them with a database. Then the events are characterized and classified according to a classification algorithm. Trigger feedback to the operator (e.g., safety-critical warnings). Thereby, the health status of the switchgear (including its components) can be calculated, and the switchgear maintenance procedures can be improved. Additionally, the new technology also reduces the risks for on-site workers, as well as damage and downtime caused by equipment failures due to unauthorized manipulation or intrusion.

[0177] The new technology solves safety-related incidents. Such incidents in electrical equipment may ultimately lead to catastrophic failures. Therefore, panel interlocks can be detected and identified. A panel interlock is a type of earthing switch that can ensure the safe disconnection of switchgear by personnel. The interlock system includes the following safety-related functions: (a) the circuit breaker must be disconnected, and (b) the three-phase busbar must be earthed through a mechanical system. The new technology can, for example, monitor, detect, and identify earthing events to ensure that the busbar is properly earthed. In addition, the closing and opening of the switchgear door can also be detected and identified. The switchgear door can only be opened after the panel interlock is successful. However, a mechanical lock failure of the door may allow the door to be opened without the panel interlock being successful. The new technology can detect and identify this event to improve the safety of on-site workers. If necessary, the detected event can be compared with other sensor data to further enhance safety. Therefore, the new technology can achieve personnel detection and identification. By detecting personnel in the area near the switchgear panel or inside the switchgear, safety can be improved by combining this event with additional data or sensors to ensure, for example, that personnel can only enter the switchgear when the switchgear itself is de-energized and the earthing switch is activated.

[0178] The new technology also solves operation-related incidents. Some of these incidents provide an indication of the health of mechanical components to the operator (e.g., push / pull), while some incidents may damage the switchgear itself (e.g., animals in the switchgear). The new technology can detect, identify, and classify the switching events of the circuit breaker. The detection, analysis, and classification of switchgear events provide health information about the circuit breaker, such as information related to the individual components of the circuit breaker. The new technology can also achieve the detection, identification, and classification of push / pull events: pushing / pulling the circuit breaker into / out of the switchgear may damage or harm the contacts of the circuit breaker busbar. Most of the time, workers or operators are not aware of the damage to the contacts. However, now by detecting and classifying such push / pull events, the operator can obtain feedback on the event and whether it was carried out as intended or in an incorrect manner, thereby improving the safety of the equipment. The new technology can also detect and identify the operation of the barrier mechanism: the barrier is part of the push / pull mechanism. If the circuit breaker is pulled out, the barrier closes and ensures protection against accidental contact with the busbar interface of the circuit breaker and the switchgear. A damaged barrier may damage the circuit breaker arm or lead to a serious accident, and the new technology can determine whether a damaged barrier has been operated. The new technology can detect and identify animals. Animals may break into the switchgear and damage the switchgear by causing a short circuit between phases or between a phase and the ground. Such a short circuit can lead to severe damage to the switchgear and a power grid collapse, and the presence of animals can be detected and identified.

[0179] The new technology also addresses environmental events, such as earthquake detection or strong vibrations from neighboring machines or vehicles, which can be detected and identified and used in environmental history files to correlate certain events with certain interruptions and malfunctions of the switch.

[0180] Therefore, the new technology can detect safety-related events, operation-related events, and environmental events before they have harmful consequences, and it is important to take action on such events.

[0181] Reference will now be made Figures 2 - 5 to describe in more detail a switchgear event detection system and a switchgear event detection method.

[0182] The new technology / development is based on three main features:

[0183] 1. Applying switchgear monitoring can improve the safety of on-site workers and equipment. Monitoring can reduce downtime caused by equipment failures, unauthorized operations, or intrusions.

[0184] 2. Using vibration or acoustic sensor(s) in combination with a processing unit to measure and detect critical events. The sensor(s) can be placed inside a compartment of the switchgear or outside the switchgear wall. More information is described elsewhere.

[0185] 3. Conducting event detection, analysis, and classification of known and unknown critical events inside and outside the switchgear. Based on the classification algorithm, the operator can obtain meaningful feedback for taking action (such as safety-critical warnings).

[0186] Figure 2 An example of a switchgear event detection system is shown, demonstrating the actual operation of the analysis algorithm: for each measurement, an event fingerprint is created (see Figure 3 ). The fingerprint consists of a set of features defined for each event. Different events can be characterized by the following non-limiting signal attributes:

[0187] · Panel interlock: ground event monitoring, pulse excitation, low amplitude

[0188] · Door closed / open: continuous with panel interlock, medium duration, low amplitude

[0189] · Human detection: medium duration event, low to medium amplitude, repeating pattern

[0190] · Circuit breaker switch event: pulse excitation, short duration event, large amplitude

[0191] · Push / pull: medium duration event, medium amplitude

[0192] · Shutter mechanism: short duration event, medium amplitude

[0193] · Animal detection: long-duration events, low amplitude, repetitive motion patterns

[0194] · Environmental events: long-duration events, low to medium amplitude

[0195] Therefore, these signal characteristics can be described and distinguished by a set of features to characterize event duration, amplitude, and repeatability. This results in a set of features:

[0196] i) Analytical features based on the time signal envelope (signal energy, peak energy, signal duration), by using time compression / stretching methods (such as dynamic time warping and / or cyclic shift) to align events for detecting temporal patterns.

[0197] ii) Frequency and / or time-frequency characteristics, such as specific coefficients or energies from Fourier transform, wavelet transform, or cepstral coefficients.

[0198] iii) Excitation frequency estimation by autocorrelation.

[0199] iv) Additionally, the original measurement signal itself can also be used as a feature.

[0200] By using an overlapping filter bank for segmentation, these features and their combinations can be extracted from the time signal (see Figure 4 ). The filter length can be adjusted according to individual features. A parallel implementation of multiple filter banks with different filter lengths can also be utilized. In this case, the filter bank is used for time signal segmentation and / or time windowing.

[0201] For each new segment (see Figure 4 , **highlighted in braces), the selected features are calculated. These features form the fingerprint of the signal, which can then be compared with a database. Optionally, additional steps can be implemented to implement a trigger method for each attribute / feature of the fingerprint (see Figure 2 ). The method includes, for example, detection thresholds, deterministic or statistical methods. By using this event trigger, event classification is triggered only when any feature causes a trigger. This can be implemented as a warning to the operator. If this step is omitted, any fingerprint is classified. For event classification, any of the following methods / method combinations are used: feature threshold, feature limit, similarity measure, correlation coefficient, machine learning algorithms (such as SVM, etc.).

[0202] If an event cannot be recognized, it is classified as "unknown". Such events have a feedback loop that triggers human interaction to manually classify these events. The newly classified events are then fed back into the fingerprint database. If an unknown event has been manually assigned to an existing category, retraining of the detection algorithm is triggered.

[0203] Therefore, events such as switching, panel interlocking, and push - pull can be classified as healthy events, while deviation events will be classified as "unknown". If a deviation event has not been recognized and stored in the event database, it can be inspected by the operator and fed back into the event database after successful manual classification.

[0204] Overview

[0205] Therefore, new technologies or R & D involve using vibration or acoustic sensors (one or more) in combination with a processing unit to measure and detect critical events. Different sensor technologies for vibration and acoustic sensors are listed below. All sensors are placed inside or on the switchgear. The sensors can be placed in one or more compartments of the switchgear or outside the housing of the switchgear. Figure 5 Some examples of sensor placement inside or outside the switchgear are shown. These sensors or some of them can be placed on or attached to components of the switchgear (such as circuit breakers, earthing switches, two / three - position switches, etc.).

[0206] For each new sensor measurement / event detection, a fingerprint is created. The fingerprint consists of several features to characterize the important attributes of the signal. The fingerprint is then compared with a database that stores typical fingerprints of common safety - critical events. The event is then characterized and classified based on a classification algorithm. A feedback to the operator is triggered (e.g., a safety severe warning).

[0207] In the case where an event cannot be associated with a single event (unknown event), a general alarm is triggered, and then human interaction can be used to identify the event and update the fingerprint database. The initial fingerprint database is created using data from experiments / laboratories and / or on - site data of the switchgear and data of the switchgear itself being monitored. Continuously updating and manually classifying unknown events can continuously improve the system.

[0208] Benefits include:

[0209] 1. Improved safety of on - site workers.

[0210] 2. Improved safety of the equipment.

[0211] Exemplary list of possible vibration sensors:

[0212] Piezoelectric accelerometer

[0213] High-impedance accelerometer

[0214] MEMS capacitive accelerometer

[0215] Strain gauge sensor

[0216] Eddy current or capacitive displacement sensor

[0217] Laser vibrometer

[0218] Gyroscope

[0219] Electromagnetic induction vibration sensor

[0220] Ferrari sensor

[0221] Bulk micromachined capacitive

[0222] Bulk micromachined piezoresistive

[0223] Capacitive spring mass system base DC responsive

[0224] Electromechanical servo (servo force balance)

[0225] Laser accelerometer

[0226] Modal tuning impact hammer

[0227] Optical pendulous integrating gyro accelerometer (PIGA)

[0228] Quantum (rubidium atomic cloud, laser cooling)

[0229] Seat accelerometer

[0230] Shear mode accelerometer

[0231] Surface acoustic wave (SAW)

[0232] Surface micromachined capacitive (MEMS)

[0233] Thermal (sub-micron CMOS process)

[0234] Triaxial flexible anode vacuum diode

[0235] Potentiometric

[0236] LVDT type accelerometer

[0237] Inclinometer

[0238] Exemplary list of possible acoustic sensors:

[0239] MEMS microphone

[0240] Capacitive microphone

[0241] Moving coil microphone

[0242] Multi-mode microphone

[0243] Carbon fiber microphone

[0244] External polarization, pre-polarization and Microphone

[0245] Laser acoustic microphone

[0246] Stereo microphone

[0247] Piezoelectric microphone

[0248] Magnetic microphone

[0249] Piezoelectric microphone

[0250] Space surface microphone

[0251] Surface microphone

[0252] Acoustic array microphone

[0253] In another exemplary embodiment, a computer program or a computer program component is provided, which is characterized in that it is configured to execute the method steps of the method according to any one of the foregoing embodiments on a suitable processor or system.

[0254] Therefore, the computer program component can be stored on a computer unit, which can also be part of an embodiment. The computing unit can be configured to execute or induce the execution of the steps of the above method. In addition, it can be configured to operate the components of the above system. The computing unit can be configured to automatically operate and / or execute the commands of the user. The computer program can be loaded into the working memory of the data processor. Therefore, the data processor can be equipped to execute the method according to any one of the foregoing embodiments.

[0255] This exemplary embodiment of the present invention covers computer programs that use the present invention from the beginning and computer programs that transform existing programs into programs using the present invention through updates.

[0256] Furthermore, the computer program component may be able to provide all the steps required for the program to complete the exemplary embodiments of the method described above.

[0257] According to another exemplary embodiment of the present invention, a computer-readable medium is provided, such as a CD-ROM, a USB stick, etc., wherein a computer program component is stored on the computer-readable medium, and the computer program component is described in the previous section.

[0258] A computer program may be stored and / or distributed on a suitable medium, such as an optical storage medium or a solid-state medium provided together with other hardware or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

[0259] However, a computer program may also be presented via a network such as the World Wide Web and may be downloaded into the working memory of a data processor from such a network. According to yet another exemplary embodiment of the present invention, there is provided a medium for making a computer program component available for download, the computer program component being arranged to execute the method according to any one of the foregoing-described embodiments of the present invention.

Claims

1. A switchgear event detection system, comprising: - one or more vibration / acoustic sensors (10); - a processing unit (20); and - an output unit (30). Wherein, the one or more vibration / acoustic sensors are configured to be mounted in and / or on the switchgear; Wherein, the one or more vibration / acoustic sensors are configured to acquire one or more vibration / acoustic signals related to events inside and / or outside the switchgear and near the switchgear, and wherein the one or more vibration / acoustic signals extend over a time window; Wherein, the one or more vibration / acoustic sensors are configured to provide the one or more vibration / acoustic signals to the processing unit; Wherein, the processing unit is configured to compare at least a portion of the one or more vibration / acoustic signals with a plurality of reference vibration / acoustic signals stored in a database (40), wherein the plurality of reference vibration / acoustic signals are related to N different reference events, wherein information about the N different reference events is stored in the database, and wherein each of the N different reference events is associated with a different one or more of the plurality of reference vibration / acoustic signals; Wherein, the processing unit is configured to classify the event as the specific reference event when the degree of comparison match between at least a portion of the one or more vibration / acoustic signals related to the event and one or more of the plurality of reference vibration / acoustic signals associated with a specific reference event among the N different reference events exceeds a first threshold; and Wherein, the output unit is configured to output information about the event, including using the information of the specific reference event stored in the database.

2. The system according to claim 1, wherein, The N different reference events include at least one reference event inside and / or outside the reference switchgear; and / or, the N different reference events include reference events inside and / or outside the switchgear.

3. The system according to any one of claims 1 to 2, wherein The database is contained within the processing unit.

4. The system according to any one of claims 1 to 2, wherein The database is separate from the processing unit.

5. The system according to any one of claims 1 - 4, wherein, The processing unit is configured to classify the event as an unknown event when the degree of comparison match between at least a portion of the one or more vibration / acoustic signals related to the event and each of the plurality of reference vibration / acoustic signals is less than the first threshold.

6. The system according to claim 5, wherein, The processing unit is configured to classify the event as an unknown event when the degree of comparison match between at least a portion of the one or more vibration / acoustic signals related to the event and each of the plurality of reference vibration / acoustic signals is greater than a second threshold.

7. The system according to any one of claims 5-6, wherein, The processing unit is configured to control the output unit to output an alarm when the event has been classified as an unknown event.

8. The system according to any one of claims 5 - 7, wherein, The processing unit is configured to control the output unit to output information related to the event when the event has been classified as an unknown event, wherein the system includes an input unit (50), and wherein the input unit is configured to enable an operator to input information about the event, wherein the information about the event classifies the event, and wherein the processing unit is configured to update the database using one or more vibration / acoustic signals related to the event, and the one or more vibration / acoustic signals are stored as new one or more reference vibration / acoustic signals together with the information about the event that classifies the event.

9. The system according to any one of claims 1-8, wherein, The processing unit is configured to cut each of one or more vibration / acoustic signals related to the event into a plurality of segments, each segment continuing on a different part of the time window, and wherein the processing unit is configured to compare each segment of the one or more vibration / acoustic signals with a plurality of reference vibration / acoustic signals stored in the database (40), and wherein the processing unit is configured to classify the event as the specific reference event among the N different reference events when the comparison matching degree between two or more segments of the one or more vibration / acoustic signals related to the event and one or more reference vibration / acoustic signals among the plurality of reference vibration / acoustic signals associated with the specific reference event exceeds the first threshold.

10. The system according to claim 9 when dependent on any one of claims 5 - 8, wherein the processing unit is configured to classify the event as an unknown event when the comparison matching degree between two or more segments of one or more vibration / acoustic signals related to the event and each of the plurality of reference vibration / acoustic signals is less than the first threshold.

11. The system according to claim 10, wherein, The processing unit is configured to classify the event as an unknown event when the comparison matching degree between two or more segments of one or more vibration / acoustic signals related to the event and each of the plurality of reference vibration / acoustic signals is greater than the second threshold.

12. The system according to any one of claims 9-11, wherein, At least one of the plurality of segments continues in a time period different from at least one other of the plurality of segments.

13. A switching device, comprising the system according to any one of claims 1 - 12.

14. A method for detecting a switching device event, comprising: - obtaining, by one or more vibration / acoustic sensors installed in and / or on the switching device, one or more vibration / acoustic signals related to an event inside and / or outside the switching device and in the vicinity of the switching device, and wherein the one or more vibration / acoustic signals continue over a time window; - providing the one or more vibration / acoustic signals to a processing unit by the one or more vibration / acoustic sensors; - Comparing at least a portion of the one or more vibration / acoustic signals with a plurality of reference vibration / acoustic signals stored in a database (40) by the processing unit, wherein the plurality of reference vibration / acoustic signals are associated with N different reference events, and wherein information about the N different reference events is stored in the database, and wherein each of the N different reference events is associated with a different one or more of the plurality of reference vibration / acoustic signals; - Classifying by the processing unit, the processing unit being configured to classify the event as the specific reference event when a comparison match between at least a portion of the one or more vibration / acoustic signals associated with the event and one or more of the plurality of reference vibration / acoustic signals associated with a specific reference event among the N different reference events exceeds a first threshold; and - Outputting information about the event by an output unit, including using the information of the specific reference event stored in the database.

15. A computer program component for controlling a system, the computer program component being configured to perform the method of claim 14 when executed by a processor.

Images