Method for generating diagnostic information of facility on basis of acoustic data

Abstract

A method for generating diagnostic information of a facility on the basis of acoustic data according to the present invention is a method in which a facility diagnostic device generates diagnostic information of a facility on the basis of acoustic data, and comprises the steps of: acquiring raw data related to acoustic data generated from the facility; generating a spectrogram image on the basis of the raw data, wherein the spectrogram image is image information including a first axis in a time domain, a second axis in a frequency domain, and a color indicating the strength; generating first conversion data by normalizing the spectrogram image with respect to a frequency domain; generating second conversion data by axially synthesizing the first conversion data with respect to a synthesis time section; and generating third conversion data by extracting the second conversion data with respect to a frequency band of interest, wherein in the step of generating the second conversion data, second configuration information related to the synthesis time section is configured, and the second configuration information is determined according to characteristic information of the facility.

Description

A method for generating diagnostic information for equipment based on acoustic data.

[0001] The present invention relates to a technology for determining the condition of equipment based on acoustic data of the equipment.

[0002] Recently, many factory facilities have been experiencing unexpected breakdowns, leading to downtime. A single equipment failure can threaten worker safety, and if a particular piece of equipment operates in a dynamic manner with other equipment, it can disrupt the entire system. Therefore, facility operators must assess the expected lifespan of specific equipment and take steps to address these issues before a failure occurs.

[0003] However, because it's difficult to visually assess the condition of numerous factory facilities, analyzing acoustic data generated by the equipment can be used to determine their condition. However, using the acoustic data directly from the equipment can result in a massive data volume that can overload the server.

[0004] Accordingly, there is a growing demand for not using the acoustic data generated from the equipment as is, but for processing it to partially extract only the necessary portions of the acoustic data and to clearly understand the condition of the equipment based on this.

[0005] The purpose of the present invention is to extract only necessary data from acoustic data generated from equipment and to generate data capable of identifying equipment diagnostic information based on a small amount of data.

[0006] The purpose of the present invention is to minimize data loss when extracting necessary information from acoustic data and to generate diagnostic information by comparing data extracted by an artificial intelligence model with pre-stored reference data.

[0007] A method for generating equipment diagnostic information based on acoustic data of the present invention comprises a step of: obtaining raw data related to acoustic data generated from the equipment by a equipment diagnostic device; generating a spectrogram image based on the raw data, wherein the spectrogram image is image information composed of a first axis in the time domain, a second axis in the frequency domain, and a color indicating intensity; performing normalization of the spectrogram image for the frequency domain to generate first converted data; performing axis synthesis for a synthetic time interval on the first converted data to generate second converted data; and performing data extraction for a frequency band of interest on the second converted data to generate third converted data, wherein in the step of generating the second converted data, second setting information related to the synthetic time interval is set, and the second setting information is determined according to characteristic information of the equipment.

[0008] In one embodiment of the present invention, the equipment is operated according to a repeating operation cycle, and the time interval information may correspond to a time interval of at least a portion of the operation cycle.

[0009] In one embodiment of the present invention, in the step of generating the spectrogram image, first setting information related to a window size for a time domain of the raw data is set, and the first setting information can be determined according to the characteristic information.

[0010] In one embodiment of the present invention, in the step of generating the third conversion data, third setting information for the frequency band of interest is set, and the third setting information can be determined according to the characteristic information.

[0011] In one embodiment of the present invention, the method may further include a step of reducing the length of at least one of a time axis and a frequency axis of the first converted data, which is performed after the step of generating the first converted data and before the step of generating the second converted data.

[0012] In one embodiment of the present invention, the step of generating the spectrogram image includes the step of performing STFT (Short Time Fourier Transform) on the raw data, and in the step of performing STFT, the first setting information may be set.

[0013] In one embodiment of the present invention, the characteristic information may include at least one of the type, material, operation type, operation time, number of operations, and operation cycle of the equipment.

[0014] In one embodiment of the present invention, the characteristic information may include at least one of temperature information, vibration information, and humidity information for the equipment.

[0015] In one embodiment of the present invention, in the step of generating the second conversion data, the second conversion data can be synthesized as an average intensity of the intensity for the synthesis time interval.

[0016] In one embodiment of the present invention, the method may further include a step of generating diagnostic information of the equipment by comparing the second conversion data with reference data that can determine whether the equipment is faulty by an artificial intelligence model.

[0017] The present invention has the advantage of extracting only necessary data from acoustic data generated from equipment and generating data that can identify equipment diagnostic information based on a small amount of data.

[0018] The present invention has the advantage of minimizing data loss when extracting necessary information from acoustic data and generating diagnostic information by comparing data extracted by an artificial intelligence model with pre-stored reference data.

[0019] FIG. 1 is a diagram illustrating an example of a network environment according to one embodiment of the present invention.

[0020] FIG. 2 is a flowchart illustrating a method by which the equipment diagnostic device of the present invention generates equipment diagnostic information based on acoustic data.

[0021] FIG. 3 is a diagram illustrating a method for generating a spectrogram image based on raw data by a facility diagnostic device according to one embodiment of the present invention.

[0022] FIG. 4 is a diagram for explaining a method in which a facility diagnostic device according to one embodiment of the present invention performs normalization on a spectrogram image to generate first transformed data.

[0023] FIG. 5 is a diagram for explaining a method in which a facility diagnostic device according to one embodiment of the present invention performs condensation synthesis on resized data to generate second converted data.

[0024] FIG. 6 is a diagram illustrating a method for generating third conversion data by extracting data for a frequency band of interest from second conversion data by a facility diagnostic device according to one embodiment of the present invention.

[0025] FIG. 7 is a diagram for explaining that the first setting information (window size) is set differently depending on the characteristic information of the equipment according to one embodiment of the present invention.

[0026] FIG. 8 is a diagram for explaining that the second setting information (synthesis time interval) is set differently depending on the characteristic information of the equipment according to one embodiment of the present invention.

[0027] FIG. 9 is a diagram for explaining that third setting information (frequency band of interest) is set differently depending on the characteristic information of the equipment according to one embodiment of the present invention.

[0028] FIG. 10 is a diagram illustrating a method for generating diagnostic information of a facility by comparing third conversion data and reference data according to one embodiment of the present invention.

[0029] Hereinafter, embodiments disclosed in this specification will be described in detail with reference to the attached drawings. Regardless of the drawing reference numerals, identical or similar components will be assigned the same reference numerals, and redundant descriptions thereof will be omitted. Furthermore, when describing embodiments disclosed in this specification, if a detailed description of a related known technology is judged to obscure the gist of the embodiments disclosed in this specification, the detailed description thereof will be omitted.

[0030] Terms that include ordinal numbers, such as first, second, etc., may be used to describe various components, but the components are not limited by these terms. These terms are used solely to distinguish one component from another.

[0031] Singular expressions include plural expressions unless the context clearly indicates otherwise.

[0032] In this application, each step described may be performed regardless of the listed order, except in cases where a special causal relationship requires that the steps be performed in the listed order.

[0033] In this application, terms such as “include” or “have” are intended to specify the presence of a feature, number, step, operation, component, part or combination thereof described in the specification, but should be understood not to exclude in advance the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts or combinations thereof.

[0034]

[0035] Hereinafter, the present invention will be described with reference to the attached drawings.

[0036] FIG. 1 is a diagram illustrating an example of a network environment according to one embodiment of the present invention.

[0037] The equipment diagnostic device (10) may include a communication unit (11), an input unit (12), an output unit (13), a memory (14), a sensor unit (15), and a processor (16).

[0038] The communication unit (11) can communicate with the server (20) or other terminals in a wired / wireless manner. For example, the communication unit (11) can be connected to the server (20) to provide diagnostic information of the equipment.

[0039] The input unit (12) can receive various information through the user's manipulation and input actions. Such an input unit (12) can be a touch screen module, a keyboard, a mouse, a button, a camera, a stylus, a microphone, a sensor, etc.

[0040] The equipment diagnostic device (10) can receive user interaction input through the input unit (12). Interaction refers to the user manipulating the input unit to input information reflecting the user's selection or intention into the terminal. For example, the interaction may be a touch of a touchscreen, a click of a mouse, a typing on a keyboard, a sound input from a microphone, an image capture from a camera, or a motion recognition from a motion sensor. The user can input first, second, and third setting information, etc. through the input unit (12).

[0041] The output unit (13) can output various information. The output unit (13) can be a display device, a speaker, a vibration generating device, a tactile generating device, etc. In some cases, the output unit (13) can be a device (e.g., a Bluetooth earphone) that is connected to a user terminal (20) via wired or wireless communication (e.g., short-range wireless communication such as Bluetooth) to receive and output a signal. The equipment diagnostic device (10) can output equipment diagnostic information based on raw data acquired through the input unit (12) or the sensor unit (14).

[0042] The sensor unit (14) can collect (or detect) various information. The sensor unit (14) includes various sensors, microphones, and audio devices for collecting sound data. The sensor unit (14) can be configured to be attached to the equipment diagnosis device (10) or included therein. The equipment diagnosis device (10) can collect equipment sound data (raw data) through the sensor unit (14). In addition, the equipment diagnosis device (10) can collect equipment temperature information, vibration information, humidity information, etc. through the sensor unit (14).

[0043] The memory (15) functions as a storage medium and can store a number of application programs run by the equipment diagnostic device (10), data for the operation of the equipment diagnostic device (10), and commands. This memory (15) can be provided in the form of various storage devices such as ROM, RAM, flash drive, hard drive, etc., or in the form of web storage.

[0044] In one embodiment, the memory (15) may store acoustic data collected from the sensor unit (15). In another example, the memory (15) may store reference data input from a user.

[0045] The processor (16) can generate diagnostic information of the equipment by controlling the overall operation of the communication unit (11), input unit (12), output unit (13), sensor unit (14), and memory (15).

[0046] In the present invention, the equipment diagnosis device (10) can obtain raw data related to acoustic data generated from the equipment, and generate a spectrogram image based on the raw data. Here, the spectrogram image may be image information composed of a first axis in the time domain, a second axis in the frequency domain, and a color indicating intensity. The equipment diagnosis device (10) can perform normalization of the spectrogram image for the frequency domain to generate first transformed data, perform axis synthesis for the first transformed data for a synthesized time interval to generate second transformed data, and perform data extraction for a frequency band of interest on the second transformed data to generate third transformed data. Here, when generating the spectrogram image, first setting information related to a window size for the time domain of the raw data is set, and the first setting information can be determined according to characteristic information of the equipment.

[0047] Here, acoustic data refers to data related to the sounds generated by equipment during operation. This data may include sounds generated by equipment during operation and surrounding noise. This acoustic data may include data such as amplitude, frequency, time, and waveform.

[0048] Here, raw data refers to data that represents acoustic data in the form of waves. Raw data may be data that represents acoustic data in a time-amplitude graph. The equipment diagnostic device (10) contains a large amount of collected acoustic data, which may require a large amount of storage space.

[0049] Here, normalization refers to assigning values ​​corresponding to the intensities of a spectrogram image within a specific range when converting a spectrogram into an image. Specifically, the values ​​corresponding to the intensities of a spectrogram image can be adjusted to fall within a range between a minimum and maximum value. For example, when converting raw data into a spectrogram image, all values ​​corresponding to intensities can be assigned a value between 0 (minimum value) and 255 (maximum value).

[0050] The equipment diagnostic device (10) can perform quantization while performing normalization. Quantization refers to the process of approximating the values ​​assigned to intensity values ​​to specific values ​​when converting raw data into a spectrogram image. Specifically, when converting a spectrogram into an image, all values ​​assigned to intensity values ​​can be approximated to integer values.

[0051] Here, the first transformation data refers to data in which the spectrogram image is normalized. The first transformation data can be assigned integer values ​​from 0 to 255 to all intensity values ​​through normalization. In this case, the first transformation data can utilize existing image processing techniques. Specifically, image processing techniques are generally optimized for processing integer values ​​from 0 to 255. Therefore, if the first transformation data is generated through normalization and quantization in which any value from 0 to 255 is assigned to the intensity, existing image processing techniques can be used for the first transformation data.

[0052] Here, the second transformation data refers to data synthesized from the first transformation data along the time axis. Due to the time-axis synthesis, the second transformation data can only display values ​​for frequency and intensity. This can contain information similar to the compressed result of performing a Fast Fourier Transform (FFT) on raw data.

[0053] Time-synthesizing the first transformation data generates compressed data, which can be processed. Specifically, performing an FFT on the entire raw data transforms the time-domain data into the frequency domain. In this case, the capacity of the raw data and the capacity of the output data may be identical. However, by transforming the time-domain data into the frequency domain through time-synthesizing the first transformation data, the data capacity can be reduced by using the average value of the intensity data on the time axis.

[0054] Here, the frequency band of interest refers to the frequency band the user wishes to analyze. For example, if the user wishes to analyze the ultrasonic range, data below 20 kHz may be removed.

[0055] Here, the third transformation data refers to data extracted from the second transformation data, including only data within the frequency band of interest. When the equipment diagnostic device (10) extracts and analyzes only data within the region of interest (ROI), the data capacity to be analyzed can be reduced. Furthermore, it offers the advantage of enabling more precise comparisons than comparing all data.

[0056] Here, the window size refers to the length of time used to analyze data over a given time interval when converting raw data into a spectrogram image. A smaller window size results in higher frequency resolution, but lower temporal resolution. Conversely, a larger window size results in lower frequency resolution, but higher temporal resolution.

[0057] When the window size is set equal to the total time span of the spectrogram image, results similar to those obtained by performing an FFT on the entire raw data described above can be obtained. However, when the window size is divided into time intervals based on the equipment's characteristics, results similar to those obtained by an FFT may not be obtained. In such cases, the window size can be set for each time interval.

[0058] Here, characteristic information refers to information related to the unique characteristics of equipment, such as its type, purpose, and operating method. For example, if the equipment is a press machine, characteristic information may include the sound, time, intensity, and cycle of the press's operation.

[0059]

[0060] The server (20) may be implemented as a computer device or multiple computer devices that provide commands, codes, files, content, services, etc. The server (20) may be a server (20) that can transmit and receive information by communicating with the equipment diagnostic device (10) through a network.

[0061] The server (20) may include a processor (21), memory (22), and communication unit (23).

[0062] The processor (21) controls the overall operation of the memory (22) and the communication unit (23) to provide the equipment diagnostic device (10) with the function of determining equipment diagnostic information. The function of determining equipment diagnostic information provides the equipment diagnostic device (10) with the function of determining whether or not a specific equipment is faulty.

[0063] The memory (22) functions as a storage medium and can store a plurality of application programs running on the server (20), data for the operation of the server (20), and commands. In one embodiment, the memory (22) can store diagnostic information for a specific facility received from the facility diagnostic device (10).

[0064] This memory (22) may be provided in the form of various storage devices such as hardware, ROM, RAM, flash drive, hard drive, etc., or may be provided in the form of web storage.

[0065] The communication unit (23) can communicate with the equipment diagnostic device (10) via a network in a wired / wireless manner. For example, the server (20) of the present invention can receive diagnostic information on a specific equipment from the equipment diagnostic device (10) via the communication unit (23).

[0066]

[0067] Hereinafter, with reference to FIGS. 2 to 10, one embodiment of a method for generating equipment diagnostic information based on acoustic data of the equipment diagnostic device (10) of the present invention will be described.

[0068] FIG. 2 is a flowchart for explaining a method in which the equipment diagnostic device (10) of the present invention generates equipment diagnostic information based on acoustic data.

[0069] In step (201), the equipment diagnostic device (10) acquires raw data related to acoustic data generated from the equipment.

[0070] The equipment diagnostic device (10) can acquire acoustic data generated from the equipment. The acoustic data may include sounds generated from the equipment to be measured and sounds generated from equipment other than the equipment to be measured or from the surrounding environment.

[0071] When the equipment diagnostic device (10) acquires acoustic data, the equipment diagnostic device (10) can generate raw data related to the acoustic data. Specifically, the acoustic data can include data such as intensity, frequency, time, and period. The equipment diagnostic device (10) can generate raw data including a time-amplitude graph based on the acoustic data.

[0072] Raw data may include multiple sections. Specifically, the equipment may operate according to a repeating operation cycle. The operation cycle may include a first section, a second section, and a third section according to characteristic information. Here, the characteristic information may be information related to the unique characteristics of the equipment according to the type, purpose, and operating method of the equipment. Specifically, the characteristic information may include at least one of the type, material, operating type, operating time, number of operations, and operating cycle of the equipment. In addition, the characteristic information may include at least one of temperature information, vibration information, and humidity information regarding the equipment.

[0073]

[0074] In step (203), the equipment diagnostic device (10) generates a spectrogram image based on raw data.

[0075] The equipment diagnostic device (10) can generate a spectrogram image by performing a Short Time Fourier Transform (STFT) on raw data. When the equipment diagnostic device (10) generates a spectrogram image, it can set first setting information related to the window size for the time domain of the raw data.

[0076] The first configuration information can be set in various ways. For example, the first configuration information can be set according to the operating cycle. For example, if a specific device operates according to a repetitive operating cycle, the first configuration information can be set to a time interval corresponding to the operating cycle.

[0077] As another example, the first configuration information may be set based on at least a portion of the time interval of the operating cycle. For example, if a particular facility includes a first, second, and third intervals in its operating cycle, the first configuration information may be set to a time corresponding to each interval. In this case, the first configuration information may be set differently based on the time interval information of the first, second, and third intervals.

[0078]

[0079] In step (205), the equipment diagnostic device (10) performs normalization of the spectrogram image in the frequency domain to generate first transformed data.

[0080] The equipment diagnostic device (10) can perform normalization on the spectrogram image. The equipment diagnostic device (10) can perform quantization along with normalization. Specifically, when the equipment diagnostic device (10) normalizes the spectrogram image, each element of the spectrogram image can be assigned a value within a certain range. In this case, direct comparison with other spectrogram images can be easily made. For example, each element of the spectrogram image can be assigned a value between 0 and 255.

[0081] Normalized spectrogram images can be useful when analyzing data with varying sizes and amplitudes. Specifically, because signal amplitudes fluctuate significantly in noisy environments, normalization can maintain a consistent dynamic range.

[0082] The intensity value of the spectrogram image may be assigned a value including a decimal point. When the equipment diagnostic device (10) performs quantization along with normalization on the spectrogram image, the intensity value of the spectrogram may be assigned an integer value of any one of 0 to 255. Here, the data for which the spectrogram image is normalized corresponds to the first conversion data.

[0083] When the equipment diagnostic device (10) performs normalization, each element of the spectrogram is assigned an integer value within a certain range, making data processing and analysis easier. Furthermore, the normalized spectrogram (first transformed data) can be visually expressed more clearly. Normalization allows for the main patterns of the spectrum to be more clearly revealed, making it easier to understand the characteristics of the data.

[0084] The equipment diagnostic device (10) can reduce the length of at least one of the time axis and frequency axis of the first conversion data. Specifically, the equipment diagnostic device (10) can regard the first conversion data as an image and apply image processing technology. This can adjust the capacity of the first conversion data or extract a specific portion.

[0085] The equipment diagnostic device (10) can enlarge or reduce the first conversion data to a desired size. In this process, the length of at least one of the time axis and the frequency axis can be enlarged or reduced. If necessary, additional image processing techniques can be used to remove noise and adjust clarity.

[0086]

[0087] In step (207), the equipment diagnostic device (10) performs condensed synthesis on the first conversion data for a synthesis time interval to generate second conversion data.

[0088] The equipment diagnostic device (10) can perform condensation synthesis on the first conversion data for a time interval to generate second conversion data that includes only frequency and intensity values. Specifically, when performing condensation synthesis, the equipment diagnostic device (10) can synthesize intensity as an average value of intensity for the synthesis time interval.

[0089] When condensation synthesis is performed on the first transformation data, second transformation data can be generated. The second transformation data can have a result value similar to the result of performing FFT on a spectrogram image. In addition, when time-sequencing the first transformation data, data can be efficiently compressed for storage and processing. Specifically, when FFT is performed on all raw data, time-domain data can be converted into frequency domain, so that the capacity of the input data (raw data) and the capacity of the output data can be the same. However, when time-sequencing the first transformation data is converted into frequency domain, the data capacity can be reduced by using the average value of the intensity data on the time axis.

[0090] When the equipment diagnostic device (10) generates the second conversion data, second configuration information related to the synthetic time interval can be set. Here, the second configuration information can be determined according to the characteristic information of the equipment. For example, the operation cycle of the first equipment can include a first interval, a second interval, and a third interval. If the data required by the user is data regarding the first interval, the user can set information regarding the time that includes at least a portion of the first interval as the second configuration information. In this case, the equipment diagnostic device (10) can generate second conversion data related to the first interval.

[0091] If a user requires all data contained in the first, second, and third sections, the user can set secondary configuration information for each section. In this case, the secondary configuration information for each section may be set differently.

[0092]

[0093] In step (209), the equipment diagnostic device (10) performs data extraction for the second conversion data in the frequency band of interest to generate third conversion data.

[0094] When the equipment diagnostic device (10) generates the third conversion data, third configuration information for the frequency band of interest may be set. Here, the third configuration information may be determined based on characteristic information. For example, if the frequency band of interest is 20 kHz or higher, which is an ultrasonic range, the equipment diagnostic device (10) may generate the third conversion data by extracting only data above 20 kHz.

[0095] In these cases, only the data required for analysis is partially extracted, which reduces data volume. Furthermore, by comparing only the relevant portions, a clearer data comparison is possible.

[0096] The first, second, and third configuration information described above can be set in various ways depending on the characteristic information of the equipment. For example, the configuration information can be input by a user or administrator of the equipment diagnostic device. Specifically, the equipment diagnostic device (10) can request the user to input the configuration information in the steps of generating a spectrogram image (201), generating second conversion data (207), and generating third conversion data (209).

[0097] As another example, configuration information can be set in a manner that synchronizes with characteristic information. Specifically, the server (20) can receive characteristic information for a specific facility from a user. If the server (20) receives directly synchronizable information, such as operating time, number of operations, and operating cycle, among the characteristic information, the server (20) can synchronize the configuration information with the received information.

[0098] As another example, configuration information can be set based on characteristic information and existing data. Specifically, when the server (20) receives characteristic information related to the type, material, and operating mode of the equipment from the user, the server (20) can determine configuration information based on existing data that is identical or similar to the received characteristic information.

[0099] In some cases, when the server (20) receives only some of the setting information from the user, the server (20) can determine the remaining setting information that was not input based on the setting information and existing data.

[0100]

[0101] In step (211), the equipment diagnostic device (10) compares the third conversion data and reference data by an artificial intelligence model to generate equipment diagnostic information.

[0102] The equipment diagnostic device (10) can compare the third conversion data with reference data through an artificial intelligence model. The artificial intelligence model can store reference data for each equipment. For example, the artificial intelligence model can store reference data related to normal, warning, and failure status for the first, second, and third equipment.

[0103] Reference data can be modified or newly generated based on the equipment's configuration information. Specifically, the AI ​​model may include reference data for a specific equipment. The AI ​​model can identify the equipment's time interval information, first, second, and third configuration information, and generate corresponding reference data.

[0104] When the equipment diagnostic device (10) generates the third conversion data, the equipment diagnostic device (10) can identify the equipment corresponding to the third conversion data. Thereafter, the equipment diagnostic device (10) can compare the third conversion data with the reference data of the corresponding equipment to generate diagnostic information.

[0105] The equipment diagnostic device (10) can provide diagnostic information to the terminal of the equipment user or manager. In some cases, the equipment diagnostic device (10) can provide third conversion data and reference data along with the diagnostic information.

[0106]

[0107] Referring to FIGS. 7 to 10 below, an example of a method in which a facility diagnostic device (10) generates facility diagnostic information by comparing third conversion data and reference data will be described.

[0108] FIG. 7 is a diagram for explaining that the first setting information (window size) is set differently depending on the characteristic information of the equipment according to one embodiment of the present invention.

[0109] Referring to Figure 7, the equipment can operate according to a repeating operation cycle. The operation cycle period of each equipment can be set differently based on the equipment's operation. As illustrated in Figure 7, the first equipment and the third equipment can have different operation cycle periods set according to the equipment's operating time.

[0110] The multiple time interval information included in the operation cycle may be set differently based on the operation of the equipment in the operation cycle. For example, the first operation cycle of the first equipment may include a first interval (711), a second interval (713), and a third interval (715) depending on the operation of the first equipment. However, the first operation cycle of the third equipment may include a first interval (731) and a second interval (733) depending on the operation of the third equipment.

[0111] Multiple time interval information may be set differently based on the operation of the equipment. Specifically, the first equipment and the second equipment may each include a first interval (711, 721), a second interval (713, 723), and a third interval (715, 725). The first interval (711), the second interval (713), and the third interval (715) of the first equipment may have the same operation time for each interval, and thus may have the same time interval. However, the first interval (721), the second interval (723), and the third interval (725) of the second equipment may have different operation times for each interval, and thus may have different time intervals, as illustrated in FIG. 7.

[0112] The first configuration information of a facility may be set differently depending on the facility's characteristic information. The characteristic information may include multiple time interval information divided based on the facility's operation. For example, the first facility may include a first interval (711), a second interval (713), and a third interval (715).

[0113] The first setting information may be set differently for each section depending on the characteristics of the equipment. For example, if the time required for data analysis is 1 second depending on the operation type of the first equipment in the first section (711), the first setting information for the first section may be set to 1 second. However, if the time required for data analysis is 0.5 seconds depending on the operation type of the first equipment in the second section (713), the first setting information for the second section may be set to 0.5 seconds.

[0114] Depending on the first configuration information of the raw data (710, 720, 730), the frequency resolution of the spectrogram image may change. Specifically, a larger window size (first configuration information) may include more frequency intervals, thereby increasing the frequency resolution. Conversely, a smaller window size may include a relatively smaller number of frequency intervals, thereby decreasing the frequency resolution.

[0115]

[0116] FIG. 8 is a diagram for explaining that the second setting information (synthesis time interval) is set differently depending on the characteristic information of the equipment according to one embodiment of the present invention.

[0117] Referring to FIG. 8, the second setting information (811, 821, 831) of the first, second, and third equipment may be set differently. Specifically, the second setting information (811, 821, 831) may be set according to the equipment's characteristic information. The characteristic information may include at least one of the equipment type, operation type, operation time, number of operations, and operation cycle.

[0118] For example, the characteristic information may be related to the operating time (operation cycle) of the equipment. In this case, the second configuration information (811, 821, 831) may be set differently depending on the operating time of the first equipment, the operating time of the second equipment, and the operating time of the third equipment. For example, if the operating time (operation cycle) of the first equipment is 10 seconds, the operating time (operation cycle) of the second equipment is 5 seconds, and the operating time (operation cycle) of the third equipment is 20 seconds, each of the second configuration information (811, 821, 823) may be set differently.

[0119] As another example, the characteristic information may be related to the operating type of the equipment. For example, the second configuration information (811, 821, 831) may be set differently for the pressure, load, RPM, intensity, power, operating time, and cycle of the first, second, and third equipment, respectively.

[0120] The second configuration information may be set as a time interval that includes the operating time (operation cycle) of the equipment in the resized data (810, 820, 830). Here, the resized data (810, 820, 830) may have a time axis that is compressed compared to the actual time due to STFT, normalization, and resizing. In this case, if the operating time (operation cycle) of the first equipment is 10 seconds, the second configuration information (811) of the first equipment may be a compressed time (0 to 1 second) rather than 0 to 10 seconds.

[0121] The equipment diagnostic device (10) can perform time-synthesis on the resized data (810, 820, 830) based on the second setting information (811, 821, 831) to generate second conversion data.

[0122]

[0123] FIG. 9 is a diagram for explaining that third setting information (frequency band of interest) is set differently depending on the characteristic information of the equipment according to one embodiment of the present invention.

[0124] Referring to FIG. 9, the third configuration information (911, 921, 931) of the first, second, and third equipment may be set differently. Specifically, the third configuration information (911, 921, 931) may be set according to the equipment's characteristic information. The characteristic information may include at least one of the equipment type, operation type, operation time, number of operations, and operation cycle.

[0125] For example, characteristic information may be related to the type of equipment. For example, the first equipment may be a welding machine, the second equipment may be a press machine, and the third equipment may be a grinder. In this case, the frequency band required for stability testing may be a mid-frequency band for the first equipment, a low-frequency band for the second equipment, and a high-frequency band for the third equipment. In this case, the third configuration information (911, 921, 931) may be set differently.

[0126] As another example, characteristic information may be related to the operating type of the equipment. For example, the third configuration information (911, 921, 931) may be set differently for the pressure, load, RPM, intensity, power, operating time, and cycle of the first, second, and third equipment, respectively.

[0127] The equipment diagnostic device (10) can extract data included in the frequency range of interest of the second conversion data (910, 920, 930) based on the third setting information (911, 921, 931) to generate third conversion data.

[0128]

[0129] FIG. 10 is a diagram illustrating a method for generating diagnostic information of a facility by comparing third conversion data and reference data according to one embodiment of the present invention.

[0130] Referring to FIG. 10, the equipment diagnostic device (10) or server (20) can compare the third conversion data with reference data. Specifically, the reference data may be image data corresponding to the third conversion data.

[0131] The equipment diagnostic device (10) or server (20) can compare the third conversion data with the reference data through an artificial intelligence model. Specifically, the equipment diagnostic device (10) or server (20) can store multiple reference data. Here, the reference data may be different images for each equipment. For example, the equipment diagnostic device (10) can store reference data of different images for the first equipment, the second equipment, and the third equipment, respectively.

[0132] Reference data can contain a variety of information. For example, reference images can include normal images, warning images, and fault images. Reference data is provided in image form, which has the advantage of allowing users to easily determine the equipment status with their own eyes, rather than through a facility diagnostic device (10).

[0133] Reference data can be generated or modified by an AI model. Specifically, if there is no reference data corresponding to the third transformation data generated based on the first, second, and third configuration information input by the user or administrator, the AI ​​model can modify or create existing reference data to match the configuration information.

[0134] The AI ​​model may generate or modify reference data corresponding to the third transformation data only when the change in reference data has a minimal impact on the diagnosis of equipment condition. If the change significantly impacts the diagnosis of equipment condition, the AI ​​model may request additional input for the relevant reference data without modifying or generating the reference data.

[0135] The AI ​​model can generate equipment diagnostic information by comparing previously learned normal reference data and fault reference data with the third-converted data. Specifically, if the third-converted data matches or is similar to the normal reference data, the AI ​​model can determine that the third-converted data is in a normal state.

[0136] Additionally, if the third transformation data matches or is similar to the failure reference data, the artificial intelligence model can determine that the status of the third transformation data is a failure.

[0137] Additionally, if the third transformation data is not similar to both the normal reference data and the faulty reference data, the artificial intelligence model can determine the status of the third transformation data as a warning status.

[0138] Once an AI model generates diagnostic information for a facility, it can provide it to a user or administrator terminal. In some cases, the AI ​​model may only provide diagnostic information if the facility is in a warning or fault condition.

[0139]

[0140] The technical features disclosed in each embodiment of the present invention are not limited to that embodiment, and, unless they are mutually incompatible, the technical features disclosed in each embodiment may be combined and applied to different embodiments.

[0141] Therefore, although each embodiment focuses on its own technical features, each technical feature can be applied in combination with each other as long as they are not mutually incompatible.

[0142] The present invention is not limited to the above-described embodiments and the attached drawings, and various modifications and variations are possible within the scope of those skilled in the art. Therefore, the scope of the present invention should be defined not only by the claims of this specification but also by equivalents thereof.

[0143] Referring to FIGS. 3 to 6 below, an example of a method in which the equipment diagnostic device (10) generates third conversion data based on raw data will be described.

[0144] FIG. 3 is a diagram for explaining a method for generating a spectrogram image (320) based on raw data (310) by an equipment diagnostic device (10) according to one embodiment of the present invention.

[0145] Referring to FIG. 3, raw data (310) may be waveform data composed of a first axis in the time domain and a second axis in the amplitude domain. A spectrogram image (320) may be image information composed of a first axis in the time domain, a second axis in the frequency domain, and a color representing intensity.

[0146] The equipment diagnostic device (10) can perform STFT on raw data (310) to generate a spectrogram image (320). When the raw data (310) is converted into a spectrogram image (320), frequency information can be compressed. When the equipment diagnostic device (10) performs STFT on raw data (310), first setting information related to a window size (311) can be set.

[0147] The first configuration information may be determined based on the equipment's characteristic information. For example, if the equipment is a press machine, it may relate to the pressure, load, material, operating time, and cycle of the press machine. The determination of the first configuration information based on the equipment's characteristic information is described in detail in Fig. 7.

[0148]

[0149] FIG. 4 is a diagram for explaining a method in which an equipment diagnostic device (10) according to one embodiment of the present invention performs normalization on a spectrogram image (410) to generate first transformed data (420).

[0150] Referring to FIG. 4, the equipment diagnostic device (10) can normalize the spectrogram image (410) and perform gamma correction and resizing using image processing technology.

[0151] The equipment diagnostic device (10) can perform normalization on the spectrogram image (410). The equipment diagnostic device (10) can perform quantization together with normalization. The spectrogram image (410) can be composed of two-dimensional matrix values ​​including decimal point values. In this case, a lot of calculations may be required to analyze the spectrogram image (410). However, since the main information of the spectrogram image (410) is contained in integer values, even when the spectrogram image (410) is normalized, calculations can be performed without a large error in analyzing the data.

[0152] Specifically, the equipment diagnostic device (10) can assign intensity values ​​included in the spectrogram image (410) to integer values ​​between 0 and 255 through quantization. As illustrated in FIG. 4, the equipment diagnostic device (10) can normalize the spectrogram image (410) to generate first conversion data (420). The first conversion data (420) can be assigned intensity values ​​as integer values ​​between 0 and 255. In this case, the first conversion data (420) can utilize existing image processing technology.

[0153] The equipment diagnostic device (10) can perform gamma correction and resizing, which are conventional image processing technologies. Specifically, the equipment diagnostic device (10) can perform gamma correction. The equipment diagnostic device (10) can convert the first conversion data (420) through gamma correction to better distinguish between light and dark. When gamma correction is performed on the first conversion data (420), the first conversion data (420) can be nonlinearly converted and corrected to data (430) that can better distinguish between light and dark areas.

[0154] In addition, the equipment diagnostic device (10) can perform resizing. Specifically, the equipment diagnostic device (10) can perform resizing on the first converted data (420) or the data (430) on which gamma correction has been performed. The equipment diagnostic device (10) can change the size to include only the desired data (440) by performing resizing on the time axis or frequency axis. When the equipment diagnostic device (10) performs resizing, the resolution of the resized data (440) on the time axis or frequency axis can change.

[0155] For example, resizing the time axis for data (430) to a smaller interval may increase the resolution of intensity values ​​(colors), and resizing to a larger interval may decrease the resolution of intensity values ​​(colors). Additionally, resizing the frequency axis for data (430) to a smaller interval may increase the resolution of intensity values ​​(colors), and resizing to a larger interval may decrease the resolution of intensity values ​​(colors).

[0156]

[0157] FIG. 5 is a diagram for explaining a method in which a facility diagnostic device (10) according to one embodiment of the present invention performs condensation synthesis on resized data (510) to generate second converted data (521, 523).

[0158] Referring to FIG. 5, the equipment diagnostic device (10) can perform time axis synthesis on resized data (510). Specifically, the data (510) can be image information composed of a first axis in the time domain, a second axis in the frequency domain, and a color (pixel) indicating intensity.

[0159] The equipment diagnostic device (10) can perform axial synthesis on the data (510) based on the second setting information. The equipment diagnostic device (10) can convert the data (510) into an image composed of a second axis in the frequency domain and a color (pixel) representing intensity through axial synthesis. When the data (510) is synthesized in the time axis direction, the second transformed data (521, 523) can contain the same information as the effect of performing FFT on the data (510). When the data (510) is synthesized in the time axis direction, there is an advantage in that the data capacity can be reduced compared to when FFT is performed on the data (510).

[0160] When the equipment diagnostic device (10) generates the second conversion data (521, 523), second configuration information related to the synthesis time interval may be set. The second configuration information may be set differently for each interval. For example, as illustrated in FIG. 5, the first synthesis time interval and the second synthesis time interval may be set differently.

[0161] The second configuration information may be determined based on the equipment's characteristic information. The determination of the second configuration information based on the equipment's characteristic information is described in detail in Fig. 8.

[0162]

[0163] FIG. 6 is a diagram for explaining a method in which a facility diagnostic device (10) according to one embodiment of the present invention extracts data for a frequency band of interest (611) from second conversion data (610) to generate third conversion data (620).

[0164] Referring to FIG. 6, the equipment diagnostic device (10) can extract data for a frequency band of interest from the second conversion data (610) to generate third conversion data (620). Specifically, the second conversion data (610) can be image information composed of a second axis in the frequency domain and a color (pixel) representing intensity.

[0165] The equipment diagnostic device (10) can generate third conversion data (620) based on the third configuration information. The equipment diagnostic device (10) can extract only the data necessary for data analysis through data extraction. In this case, compared to analyzing the second conversion data (610), the third conversion data (620) has the advantage of reducing the data capacity.

[0166] The equipment diagnostic device (10) can perform data extraction for the frequency band of interest (611) to generate third conversion data (620). When generating the third conversion data (620), third setting information related to the frequency band of interest (611) can be set. For example, the third setting information (611) can be an ultrasonic band (20 kHz or more).

[0167] The third configuration information may be determined based on the equipment's characteristic information. The determination of the third configuration information based on the equipment's characteristic information is described in detail in Fig. 9.

Claims

1. In a method for generating equipment diagnostic information based on acoustic data by an equipment diagnostic device, A step of acquiring raw data related to acoustic data generated from the above equipment; A step of generating a spectrogram image based on the above raw data, wherein the spectrogram image is image information composed of a first axis in the time domain, a second axis in the frequency domain, and a color representing intensity; A step of generating first transformed data by performing normalization of the above spectrogram image in the frequency domain; A step of generating second transformation data by performing condensation synthesis on the first transformation data for a synthesis time interval; and A step of generating third transformed data by performing data extraction for a frequency band of interest from the second transformed data, In the step of generating the second conversion data, second setting information related to the synthetic time interval is set, The above second setting information is determined according to the characteristic information of the equipment. A method for generating diagnostic information for a facility based on acoustic data.

2. In paragraph 1, The above characteristic information includes multiple time interval information divided based on the operation of the equipment, The above second setting information is set differently depending on the time interval information. A method for generating diagnostic information for a facility based on acoustic data.

3. In paragraph 2, The above equipment operates according to a repetitive operation cycle, The above time interval information corresponds to a time interval of at least a part of the above operation cycle. A method for generating diagnostic information for a facility based on acoustic data.

4. In paragraph 1, In the step of generating the above spectrogram image, first setting information related to the window size for the time domain of the raw data is set, The above first setting information is determined according to the above characteristic information. A method for generating diagnostic information for a facility based on acoustic data.

5. In paragraph 1, In the step of generating the third conversion data, third setting information for the frequency band of interest is set, The above third setting information is determined according to the above characteristic information. A method for generating diagnostic information for a facility based on acoustic data.

6. In paragraph 1, which is performed after the step of generating the first conversion data and before the step of generating the second conversion data, Further comprising a step of reducing the length of at least one of the time axis and the frequency axis of the first conversion data. A method for generating diagnostic information for a facility based on acoustic data.

7. In paragraph 6, The step of generating the above spectrogram image is: Comprising a step of performing STFT (Short Time Fourier Transform) on the above raw data, In the step of performing the above STFT, the first setting information is set A method for generating diagnostic information for a facility based on acoustic data.

8. In paragraph 1, The above characteristic information includes at least one of the type, material, operation type, operation time, number of operations, and operation cycle of the equipment. A method for generating diagnostic information for a facility based on acoustic data.

9. In paragraph 1, The above characteristic information includes at least one of temperature information, vibration information, and humidity information for the equipment. A method for generating diagnostic information for a facility based on acoustic data.

10. In paragraph 1, In the step of generating the second conversion data, the second conversion data is synthesized as the average intensity of the intensity for the synthesis time interval. A method for generating diagnostic information for a facility based on acoustic data.

11. In paragraph 1, It further includes a step of generating diagnostic information of the equipment by comparing the second conversion data with reference data that can determine whether the equipment is faulty by an artificial intelligence model. A method for generating diagnostic information for a facility based on acoustic data.

Images