Method for monitoring acoustic phenomena in a submarine

Abstract

Method for monitoring acoustic phenomena of a submarine The present invention relates to a method for monitoring acoustic phenomena of a submarine (10) having a hull (12), at least one accelerometer (14) on the hull (12), and at least one piece of equipment (16) capable of performing dynamic movements in operation.The method comprises the following steps: - data acquisition by at least one accelerometer (14), - recording of the dynamic movements performed by at least one piece of equipment (16), - identification in the recorded dynamic movements of one or more occurrences of at least one event, each event corresponding to a type of dynamic movement for a given piece of equipment (16), each occurrence corresponding to a respective instant, and - labeling of the acquired data with at least one label, each label corresponding to an identified occurrence, each label being associated with the respective instant of the corresponding identified occurrence, each label including a field identifying the corresponding event. Figure for the abstract: Fig 1.

Description

Title of the invention: Method for monitoring acoustic phenomena in a submarine

[0001] The present invention relates to a method for monitoring acoustic phenomena of a submarine.

[0002] To date, acoustic state monitoring systems equipping submarines include hull-mounted accelerometer instrumentation and onboard systems. The onboard systems are related to a source, i.e., to equipment that is being specifically monitored.

[0003] The detection of an acoustic anomaly relies on all of these systems.

[0004] On the one hand, for each embedded system called a source sensor, a frequency template is, for example, defined to detect any anomaly of the source.

[0005] In this case, it is easy to identify the equipment responsible.

[0006] However, only equipment fitted with a dedicated sensor is monitored by this system.

[0007] On the other hand, an overall frequency template is defined on the average energy of the instrumentation on the hull.

[0008] In this case, the monitoring system is not able to locate the source of the anomaly.

[0009] This then encourages increasing the number of source sensors to monitor as much equipment as possible.

[0010] However, this complicates the system, and thus increases maintenance, cost, and the space required, while the available space is particularly limited in a submarine.

[0011] Moreover, a large amount of data is generated, and it is therefore not possible to analyze all of it.

[0012] The aim of the invention is therefore to propose a method of monitoring acoustic phenomena of a submarine which makes it possible to identify the source of an anomaly and which does not require an increased number of sensors.

[0013] To this end, the invention relates to a method for monitoring acoustic phenomena of a submarine having a hull, the submarine comprising at least one accelerometer on the hull, the submarine further comprising at least one piece of equipment, the at least one piece of equipment being capable of performing dynamic movements in operation, the method comprising the following steps:

[0014] - acquisition by at least one accelerometer of data over a period acquisition,

[0015] - recording during the acquisition time of the dynamic movements performed by at least one piece of equipment,

[0016] - identification in the recorded dynamic movements of one or more occurrences of at least one event, the event or events corresponding to a type of dynamic movement for a given piece of equipment, each occurrence corresponding to a respective instant in the acquisition duration, and

[0017] - labeling of acquired data with at least one label, each label corresponding to a respective identified occurrence, each label being associated with the respective time of the corresponding identified occurrence, each label including a field identifying the corresponding event.

[0018] Labeling the acquired data thus makes it possible to link an event to the various measurements acquired by the accelerometer mounted on the hull, for example, in order to identify it subsequently. Therefore, an anomaly can be linked to an event, and thus to its source.

[0019] According to other advantageous aspects of the invention, the method comprises one or more of the following features, taken individually or in all technically possible combinations:

[0020] - the method includes a step of defining an acoustic template for each event from the acquired labeled data;

[0021] - the acoustic template is defined for each event from the data acquired during at least one time segment, each time segment corresponding to a time interval comprising a respective instant of an identified occurrence of said event;

[0022] - each time segment begins at the respective instant and has a duration data;

[0023] - the acoustic template includes a first classifier for operation operational and a second classifier for operation with anomaly;

[0024] - the method includes the steps of: providing a data record acquired by at least one accelerometer during an observation period, detection of at least one acoustic template in the data recording, identification of the event corresponding to the detected acoustic template, and identification of the equipment corresponding to the event;

[0025] - the method includes a step of emitting an alert signal in case of detection of the second classifier of the detected acoustic template;

[0026] - the method includes a step of displaying an alert and / or issuing a visual and / or audible alert in case of an alert signal being emitted;

[0027] - the step of identifying the occurrence(s) of at least one event in the Recorded dynamic movements include:

[0028] - the calculation of a speed corresponding to the recorded dynamic movements,

[0029] - the determination of a function, called a state function, such that, at each instant of the acquisition time is equal to a first value if the speed is greater than a given positive threshold, to a second value if the speed is less than a given negative threshold (the given negative threshold being less than the given positive threshold), and to a third value otherwise, and

[0030] - the detection of rising and / or falling edges of the state function; and / or

[0031] - the step of identifying the occurrence(s) of at least one event in the recorded dynamic movements further includes the selection of fronts selected from among the rising and / or falling fronts detected, such that the selected fronts are temporally separated from the other detected fronts by at least a given time interval of isolation.

[0032] The invention further relates to a computer program product comprising software instructions which, when executed by a computer, implement the identification and labeling steps of the monitoring process as defined above.

[0033] The invention will become clearer upon reading the following description, given solely by way of non-limiting example, and made with reference to the drawings in which:

[0034] [Fig-1] [Fig.1] is a schematic view of a submarine in accordance with the invention,

[0035] [Fig.2] [Fig.2] is a schematic view of the steps of an example of a monitoring method according to the invention,

[0036] [Fig.3] [Fig.3] is an example of time-dependent functions, during the acquisition time, for the identification step,

[0037] [Fig.4] [Fig.4] is a schematic view of additional steps according to a second example of a monitoring method according to the invention, and

[0038] [Fig.5] [Fig.5] is a schematic view of additional steps according to a third example of a monitoring method according to the invention.

[0039] The invention relates to a method for monitoring acoustic phenomena of a submarine 10, for example a nuclear submarine.

[0040] Submarine 10 has a hull 12.

[0041] The submarine 10 includes at least one accelerometer 14 on the hull 12.

[0042] The accelerometer is capable of perceiving acoustic phenomena at the level of the hull 12.

[0043] Submarine 10 further includes at least one piece of equipment 16.

[0044] In operation, at least one piece of equipment 16 is capable of performing dynamic movements, that is to say that in operation, at least one part composing the equipment 16 is capable of moving.

[0045] At least one piece of equipment 16 is arranged inside the submarine 10.

[0046] The at least one piece of equipment 16 comprises, for example, by way of illustration, a bar, the bar being capable of angular movement during operation.

[0047] The invention relates to a method for monitoring 100 acoustic phenomena of the submarine 10.

[0048] The monitoring process comprises the following steps:

[0049] - acquisition 102 by at least one accelerometer 14 of data for a duration acquisition,

[0050] - recording 104 during the duration of the acquisition of dynamic movements carried out by at least one piece of equipment 16,

[0051] - identification 106 in the recorded dynamic movements of one or occurrences of at least one event, and

[0052] - labeling 108 of the acquired data with at least one label.

[0053] During acquisition 102, the data measured by at least one accelerometer during the acquisition time are acquired, and recorded here.

[0054] The acquired data correspond to acoustic phenomena at the level of the hull.

[0055] During recording 104, the dynamic movements carried out by at least one piece of equipment 16 in parallel with the acquisition 102, i.e. at the same time, are recorded.

[0056] In the example of a bar, the angular movement(s) made by the bar during the acquisition 102 are, for example, recorded.

[0057] For the remainder, when the bar moves angularly in one direction, the movement is said to be positive, and if it is in the opposite direction, the movement is said to be negative.

[0058] During identification 106, one or more occurrences of at least one event are identified, for example, by an identification module.

[0059] The event or each event corresponds to a type of dynamic movement for a given piece of equipment.

[0060] Events are, for example, predetermined and provided to the identification module.

[0061] Alternatively, the events are determined by an electronic module by analyzing the acquired data, for example, by searching for patterns.

[0062] In the example of a given bar, the events are, for example, the following: absence of bar movement, transition from a fixed bar angle to a positive angle movement, transition from a fixed bar angle to a movement negative angle, transition from a positive bar movement to a stop or hold of the bar, transition from a negative bar movement to a stop or hold of the bar, positive bar movement, and negative bar movement.

[0063] Here, at least one occurrence of at least one event is identified.

[0064] Each occurrence corresponds to a respective instant in the acquisition duration.

[0065] The identification 106 of the occurrence(s) of at least one event in the recorded dynamic movements includes, for example:

[0066] - the calculation of a speed corresponding to the recorded dynamic movements,

[0067] - the determination of a function, called a state function, such that, at each instant of the acquisition time is equal to a first value if the speed is greater than a given positive threshold, to a second value if the speed is less than a given negative threshold (the given negative threshold being less than the given positive threshold), and to a third value otherwise, and

[0068] - the detection of rising and / or falling edges of the state function.

[0069] An example of functions during the acquisition time is shown in [Fig.3] to illustrate this.

[0070] The function fl corresponds to the function of the dynamic movements recorded during the acquisition period.

[0071] For example, in the case of a bar, the function fl corresponds to the angle of the bar.

[0072] The function f2 corresponds to the velocity corresponding to the dynamic movements.

[0073] For example, in the case of a bar, the function f2 corresponds to the angular velocity of the bar.

[0074] The function f3 corresponds to an example of the state function defined above.

[0075] The negative threshold given here is equal to the opposite of the positive threshold given.

[0076] The given positive threshold and / or the given negative threshold are, for example, adjustable.

[0077] The second value is, for example, equal to the opposite of the first value.

[0078] The third value is, for example, zero.

[0079] The function f3 has several rising and falling fronts.

[0080] Identification 106 further includes, for example, the selection of selected fronts from among the rising and / or falling detected fronts, such that the selected fronts are temporally distant from the other detected fronts by at least a given isolation time interval.

[0081] The given isolation time interval is, for example, adjustable.

[0082] This makes it possible to identify isolated events.

[0083] The function f4 corresponds to the selected fronts during the acquisition time.

[0084] During identification 106, the occurrence or occurrences of at least one event are, for example, identified among a stationary state or the detected rising and / or falling fronts of the state function.

[0085] More particularly here, during identification 106, the occurrence or occurrences of at least one event are, for example, identified as the selected rising and / or falling fronts of the state function.

[0086] In the example shown in [Fig.5], five occurrences of events are identified: a transition from a maintained bar angle to a positive angle movement, a transition from a positive bar movement to a bar stop / hold, a transition from a maintained bar angle to a positive angle movement, a transition from a maintained bar angle to a positive angle movement, and then a transition from a positive bar movement to a bar stop / hold.

[0087] Then, during the labeling 108, the acquired data are labeled with at least one label, for example by a labeling module.

[0088] The identification module transmits to the labeling module the identified occurrence(s), the respective times of each identified occurrence, and the identified event for each occurrence.

[0089] During labeling 108, each label corresponds to a respective identified occurrence.

[0090] More specifically, for each identified occurrence, a label is added to the acquired data.

[0091] Each label is associated with the respective instant of the corresponding identified occurrence.

[0092] Each label includes a field identifying the corresponding event.

[0093] For example, said field is equal to a value, the value being associated with the corresponding event.

[0094] Thus, in the illustrative example, five labels are added by the labeling module to the acquired data, at the same times as the event occurrences are identified. Each label specifies whether it is a transition from a maintained bar angle to a positive angle movement, or a transition from a positive bar movement to a stop / hold of the bar, for example by using a corresponding code.

[0095] The labeling of the acquired data thus makes it possible to link an event to the different measurements acquired by the accelerometer arranged on the hull.

[0096] In a particular embodiment, shown partially in [Fig.4], the method further includes a step of defining an acoustic template 110 for each event from the acquired labeled data, after the labeling step 108, for example by a definition module.

[0097] The labeling module transmits the labeled data to the definition module.

[0098] The steps described opposite [Fig. 2] are, for example, repeated, the whole labeled data being used during the definition step.

[0099] The acoustic template is defined for each event from the data acquired during at least one time segment, each time segment corresponding to a time interval comprising a respective instant of an identified occurrence of said event.

[0100] The respective instant here corresponds to the instant of the rising or falling front detected.

[0101] More particularly here, each time segment begins at the respective instant and has a given duration.

[0102] The given duration is, for example, predetermined according to the event.

[0103] Alternatively, the given duration is adjustable.

[0104] For each event, the acoustic template is defined from the acquired labeled data corresponding to each labeled occurrence of the event, more particularly including if an anomaly is detected.

[0105] In the example of the bar, the detected anomaly is, for example, that the movement of the bar was not carried out according to a predefined dynamic.

[0106] The anomaly is, for example, detected for the occurrence by comparing the labeled data with a model of accelerometer data measured during operational or normal functioning, more specifically during the time segment. When the labeled data do not conform to the model, an anomaly is detected.

[0107] The model is, for example, predetermined.

[0108] In a particular embodiment, the acoustic template includes a first classifier for operational functioning and a second classifier for functioning with anomaly.

[0109] The first classifier includes, for example, a first template determined from the acquired labeled data corresponding to all occurrences where no anomaly is detected.

[0110] The second classifier includes at least one second template determined from the acquired labeled data corresponding to all occurrences where an anomaly is detected.

[0111] Each of the first and second templates includes, for example, a signal envelope corresponding to the labeled data during the corresponding time segments. In a particular embodiment, the extreme acquired data are excluded from the definition of the envelope.

[0112] Alternatively, each of the first and second templates is determined by artificial intelligence, for example using a neural network.

[0113] The neural network comprises an ordered succession of layers of neurons, each of which takes its inputs from the outputs of the previous layer.

[0114] More precisely, each layer comprises neurons taking their inputs from the outputs of the neurons of the previous layer, or from the input variables for the first layer.

[0115] Alternatively, more complex neural network structures can be envisaged with a layer that can be linked to a layer further away than the immediately preceding layer.

[0116] Each neuron is also associated with an operation, that is to say a type of processing, to be carried out by said neuron within the corresponding processing layer.

[0117] Each layer is connected to the other layers by a plurality of synapses. A synaptic weight is associated with each synapse, and each synapse forms a link between two neurons. It is often a real number, which takes both positive and negative values. In some cases, the synaptic weight is a complex number.

[0118] Each neuron is designed to perform a weighted sum of the value(s) received from the neurons of the preceding layer, each value being multiplied by the respective synaptic weight of each synapse, or connection, between said neuron and the neurons of the preceding layer, and then to apply an activation function, typically a non-linear function, to said weighted sum, and to deliver at the output of said neuron, in particular to the neurons of the next layer connected to it, the value resulting from the application of the activation function. The activation function introduces non-linearity into the processing performed by each neuron. The sigmoid function, the hyperbolic tangent function, and the Heaviside function are examples of activation functions.

[0119] As an optional complement, each neuron is also capable of applying, in addition, a multiplicative factor, also called bias, to the output of the activation function, and the value delivered at the output of said neuron is then the product of the bias value and the value from the activation function.

[0120] A convolutional neural network is also sometimes called a convolutional neural network or by the acronym CNN, which refers to the English term "Convolutional Neural Networks".

[0121] In a convolutional neural network, each neuron in the same layer has exactly the same connection pattern as its neighboring neurons, but at different input positions. The connection pattern is called the convolutional kernel or, more commonly, the "kernel" in reference to the corresponding English term.

[0122] A fully connected layer of neurons is a layer in which the neurons of said layer are each connected to all the neurons of the preceding layer.

[0123] Such a type of layer is more often referred to by the English term "fully connected", and sometimes designated by the name "dense layer".

[0124] In a particular embodiment, partially shown in [Fig. 5], the process further comprises the steps of:

[0125] - supplying 112 of a record of data acquired by at least one accelerometer on the hull of a submarine during an observation period,

[0126] - detection 114 of at least one acoustic template in the data recording,

[0127] - identification of event 116 corresponding to the detected acoustic template, and

[0128] - identification of equipment 118 corresponding to the event.

[0129] The accelerometer that acquired the recording is arranged similarly on the sub- marine than for the acquisition stage.

[0130] The detection and identification of the event and the equipment are, for example, carried out by an analysis module.

[0131] The definition module transmits the acoustic template defined for each event to the analysis module.

[0132] In addition, during the supply step, the record is supplied to the analysis module.

[0133] During the detection step, the data recording is analyzed so as to detect occurrences of acoustic templates.

[0134] More specifically, the entire recording data is compared to the different acoustic templates, an acoustic template being detected when the recording data during a time segment corresponds to said acoustic template.

[0135] When an acoustic template is detected, the event corresponding to said acoustic template is identified, then the corresponding equipment.

[0136] For example, the detected acoustic template corresponds to the first classifier for the event "transition from a maintained bar angle to a positive angle movement", therefore corresponding to the bar considered during the steps described opposite [Fig.2],

[0137] This makes it possible, from an accelerometer on the hull, to determine the occurrence of an event, here with or without anomaly, of equipment inside the submarine, without requiring equipment dedicated to monitoring said equipment.

[0138] Thus, after the steps of acquiring and recording dynamic movements, the monitoring of the corresponding equipment is likely to be carried out only with at least one accelerometer on the hull.

[0139] This then makes it possible to limit the source sensor in operation of the submarine.

[0140] In a particular embodiment, the method further includes a step of emitting an alert signal 120 in the event of detection of the second classifier of the detected acoustic template, for example by an alert module.

[0141] The method then includes, for example, a step of displaying an alert, for example by a display device, and / or of emitting a visual and / or audible alert, for example by an alarm device, 122 in the event of the emission of an alert signal.

[0142] The display and / or alarm device is arranged here in the submarine.

[0143] In the event of an alert being displayed, the display device shows, for example, the presence of an anomaly, and advantageously the event and the equipment identified.

[0144] This allows a person to be alerted as soon as it is detected, in order to allow for possible intervention depending on the anomaly.

[0145] Identifying the equipment concerned allows for rapid intervention on the corresponding equipment.

[0146] The identification and labeling steps, and where appropriate the steps of defining an acoustic template, and possibly of detecting at least one acoustic template, of identifying the event and of identifying the equipment, of the monitoring process as described above are suitable for implementation by an electronic device.

[0147] The electronic device includes the identification module; and the labeling module connected to the identification module.

[0148] As an optional addition, the electronic device includes the definition module connected to the labeling module, the analysis module connected to the definition module; and the alert module connected to the analysis module.

[0149] The electronic device includes an information processing unit formed for example of a memory and a processor associated with the memory.

[0150] The identification module and the labeling module, as well as, optionally, the definition module, the analysis module, and the alert module, are each implemented as a software program, or a software component, executable by the processor. The memory of the electronic device is then capable of storing identification software and labeling software, as well as, optionally, definition software, analysis software, and alert software. The processor is then capable of executing each of these software programs.

[0151] In an alternative not shown, the identification module and the labeling module, as well as the optional definition module, the analysis module and the alert module, are each implemented as a programmable logic component, such as an FPGA (Field Programmable Gate Array). or even an integrated circuit, such as an ASIC (from the English Application Specifies Integrated Circuit).

[0152] When the electronic device is implemented in the form of one or more software programs, that is, in the form of a computer program, also called a computer program product, it is further capable of being stored on a computer-readable medium, not shown, more particularly a non-transient medium. A computer-readable medium is, for example, a medium capable of storing electronic instructions and being connected to a bus of a computer system. By way of example, a readable medium is an optical disc, a magneto-optical disc, ROM, RAM, any type of non-volatile memory (for example, FLASH or NVRAM), or a magnetic card. A computer program comprising software instructions is then stored on the readable medium.

[0153] The invention further relates to a computer program product comprising software instructions which, when executed by a computer, implement the steps of identification and labeling, and where appropriate the steps of defining an acoustic template, and possibly of detecting at least one acoustic template, of identifying the event and of identifying the equipment, of the monitoring process as described above.

Claims

Demands

1. A method for monitoring (100) acoustic phenomena of a submarine (10) having a hull (12), the submarine (10) comprising at least one accelerometer (14) on the hull (12), the submarine (10) further comprising at least one piece of equipment (16), the at least one piece of equipment (16) being capable of performing dynamic movements during operation, the method comprising the following steps: - acquisition (102) by the at least one accelerometer (14) of data during an acquisition period, - recording (104) during the acquisition period of the dynamic movements performed by the at least one piece of equipment (16), - identification (106) in the recorded dynamic movements of one or more occurrences of at least one event, the event or each event corresponding to a type of dynamic movement for a given piece of equipment (16), each occurrence corresponding to a respective instant of the acquisition period,and - labeling (108) of the acquired data with at least one label, each label corresponding to a respective identified occurrence, each label being associated with the respective time of the corresponding identified occurrence, each label comprising a field identifying the corresponding event.

2. A monitoring method according to claim 1, comprising a step of defining an acoustic template (110) for each event from the acquired labeled data.

3. A monitoring method according to claim 2, wherein the acoustic template is defined for each event from data acquired during at least one time segment, each time segment corresponding to a time interval comprising a respective instant of an identified occurrence of said event.

4. A monitoring method according to claim 3, wherein each time segment starts at the respective instant and has a given duration.

5. A monitoring method according to any one of claims 2 to 4, wherein the acoustic template comprises a first classifier for operational functioning and a second classifier for functioning with anomaly.

6. A monitoring method according to any one of claims 2 to 5, comprising steps of: providing a recording (112) of data acquired by at least one accelerometer during an observation period, detecting at least one acoustic template (114) in the data recording, identifying the event (116) corresponding to the detected acoustic template, and identifying the equipment (118) corresponding to the event.

7. A monitoring method according to claims 5 and 6, comprising a step of emitting (120) an alert signal in case of detection of the second classifier of the detected acoustic template.

8. A monitoring method according to claim 7, comprising a step of displaying an alert and / or emitting a visual and / or audible alert (122) in the event of the emission of an alert signal.

9. A monitoring method according to any one of claims 1 to 8, wherein the identification step (106) of the occurrence or occurrences of at least one event in the recorded dynamic movements comprises: - the calculation of a velocity corresponding to the recorded dynamic movements, - the determination of a function, called the state function, such that, at each instant of the acquisition time, it is equal to a first value if the velocity is greater than a given positive threshold, to a second value if the velocity is less than a given negative threshold, the given negative threshold being less than the given positive threshold, and to a third value otherwise, and - the detection of the rising and / or falling edges of the state function.

10. A monitoring method according to claim 9, wherein the identification step (106) of the occurrence or occurrences of at least one event in the recorded dynamic movements further comprises the selection of selected fronts from among the rising and / or falling fronts detected, such that the selected fronts are temporally separated from the other detected fronts by at least a given isolation time interval.

11. Computer program product comprising software instructions which, when executed by a computer, implement the identification and labeling steps of the monitoring process (100) according to any one of claims 1 to 10.

Images