Safety monitoring and early warning device for energy storage power station

Abstract

The invention discloses an energy storage power station safety monitoring and early warning device, and relates to the technical field of energy storage power station safety monitoring, the energy storage power station safety monitoring and early warning device comprises a shell and a shell cover, the shell is internally provided with a monitoring unit, a data transmission module, a storage module, a main control module and an alarm module arranged on the bottom surface of the shell; through the mutual cooperative use of the main control module and the monitoring unit, the safety in the energy storage power station is monitored, so that the fire in the energy storage power station can be alarmed in time, the fire can be extinguished in time, and the safety of the energy storage power station is ensured; a normal / abnormal voiceprint feature library of the equipment is established through Gaussian mixture model training, after real-time voiceprint data is compared with the feature library, the abnormal type is automatically recognized, and an early warning signal and a specific fault report are generated; after the execution unit receives the early warning signal, the audible and visual alarm is triggered immediately, the abnormal position, the abnormal type and the abnormal severity are transmitted to a remote monitoring center in real time through the data transmission module, and a manager can start an emergency plan synchronously.

Description

Technical Field

[0001] The present invention relates to the technical field of safety monitoring of energy storage power stations, and particularly to a safety monitoring and early warning device for energy storage power stations. Background Technique

[0002] With the rapid development of energy storage technology, energy storage power stations are increasingly widely used in the power system. However, there are many potential safety hazards during the operation of energy storage power stations, such as battery overheating, fire, explosion, and leakage of harmful gases. Once a safety accident occurs, it will not only cause huge economic losses but also may threaten the lives of personnel. Currently, most of the existing safety monitoring devices for energy storage power stations have a single function and can only monitor a single parameter, unable to detect the situation of smoke or harmful gas leakage in a timely manner. Moreover, some monitoring devices lack effective early warning and data transmission functions, cannot issue alarms in a timely manner when danger occurs, and cannot comprehensively and timely grasp the operation safety status of energy storage power stations, resulting in fires in energy storage power stations, affecting the normal operation of energy storage power stations and economic losses. The existing safety monitoring devices for energy storage power stations can only give passive alarms when the concentration of smoke and harmful gases reaches a certain level, unable to give early warnings at the budding stage of accidents, resulting in a narrow accident disposal window. In the early stage, the mechanical failure of equipment, abnormal circuits, abnormal airflows, and pipeline leaks only show subtle changes in acoustic fingerprints, and the acoustic fingerprint changes usually precede the abnormalities of physical quantities such as smoke and gas concentration. Traditional devices cannot identify them, and ultimately may trigger chain accidents. Summary of the Invention

[0003] The present invention proposes a safety monitoring and early warning device for energy storage power stations to solve the problems in the background technique.

[0004] To achieve the above object, the present invention adopts the following technical solutions: A safety monitoring and early warning device for energy storage power stations includes a housing and a housing cover. A monitoring unit, a data transmission module, a storage module, a main control module, and an alarm module provided on the bottom surface of the housing are respectively arranged inside the housing. The monitoring unit includes a temperature and humidity monitoring module, a smoke monitoring module, and a gas concentration monitoring module. The main control module is electrically connected to the temperature and humidity monitoring module, the smoke monitoring module, the gas concentration monitoring module, the data transmission module, the storage module, and the alarm module respectively. An intelligent control component is further arranged on the main control module, and the intelligent control component includes an analysis unit. An analysis unit analyzes the sound data collected by the collection unit, extracts features after preprocessing, and establishes a voiceprint feature library; extracts the feature data of the sound data detected in real time, compares it with the preset normal sound data in the voiceprint feature library to determine whether an abnormality occurs; if it is determined that an abnormality occurs, it is compared with the preset abnormal sound data in the voiceprint feature library. If it is similar to the feature data of the corresponding abnormal sound in the voiceprint feature library, it is determined that the abnormal sound is caused by the situation corresponding to the corresponding abnormal sound in the voiceprint feature library, generates a warning signal, and transmits the determined abnormal situation information and warning signal to the execution unit; if it is not similar to the feature data of all abnormal sounds in the voiceprint feature library, analyze the feature data, generate a corresponding warning signal according to the determined abnormal situation, and transmit the determined abnormal situation information and warning signal to the execution unit.

[0005] Preferably, the temperature and humidity monitoring module uses a high-precision temperature and humidity sensor to monitor the temperature and humidity data inside the energy storage power station in real time and transmit the data to the main control module. The smoke monitoring module uses an infrared beam smoke detector. The gas concentration monitoring module includes a variety of gas sensors, which are respectively used to monitor the concentrations of harmful gases such as oxygen, hydrogen, and carbon monoxide.

[0006] Preferably, the data transmission module uses a 5G communication module or a wireless network module, which can transmit the monitoring data processed by the main control module to the remote monitoring center in real time, facilitating the management personnel to master the operation safety status of the energy storage power station anytime and anywhere.

[0007] Preferably, the storage module uses a large-capacity memory card to store the historical data detected for subsequent data analysis and fault troubleshooting.

[0008] Preferably, the alarm module includes an audible and visual alarm symmetrically arranged on the bottom surface of the housing.

[0009] Preferably, the intelligent control component further includes a collection unit and an execution unit. The collection unit collects the sound data at the corresponding equipment position in the energy storage power station and transmits the collected sound data to the analysis unit. The execution unit receives the warning signal transmitted by the analysis unit, issues an audible and visual alarm through the audible and visual alarm, and transmits the abnormal situation information and abnormal position information to the system platform for display through the data transmission module.

[0010] Preferably, the preprocessing steps of the analysis unit for the sound data are as follows: S1: Set the collected sound signal as , is an integer, and the unit impulse response of the filter is , it is derived that the sound signal obtained after the filtering operation is , is the index used to traverse the input signal during the convolution process ; S2: Obtain the quantization interval, quantization level , and quantization interval of the quantization level. Then, for the amplified sound signal, the digital sound signal obtained through digital conversion is . .

[0011] Preferably, the feature extraction steps of the analysis unit for sound data are as follows: Q1: Obtain the digital sound signal within a set time period, divide it into short time periods with a length of . The short-time energy of the digital sound signal in the th short time period is , and the zero-crossing rate of the digital sound signal in the th short time period is , is the th digital sound signal in the th short time period; Q2: Process the sound signal using a first-order high-pass filter to obtain the pre-emphasized sound signal , is the pre-emphasis coefficient; divide the pre-emphasized sound signal into several frames, and then apply the Hamming window function to each frame of the signal to obtain ; perform a fast Fourier transform on each windowed frame of the signal, then design a set of triangular filters according to the Mel frequency scale, convert the linear frequency axis to the Mel frequency axis, filter the signal spectrum obtained after the fast Fourier transform to obtain the Mel spectrum, and then perform a discrete cosine transform on the Mel spectrum to obtain the Mel frequency cepstral coefficients , , is the number of selected Mel frequency cepstral coefficients; Q3: Obtain the sound data under different working conditions at the positions of the fan, transformer, and battery pack, extract the features of short-time energy, zero-crossing rate, and Mel frequency cepstral coefficients from the obtained sound data, label the extracted features according to the working condition information corresponding to the sound data, train the labeled input feature data through the Gaussian mixture model GMM to obtain the rules of the corresponding voiceprint features under the corresponding working condition information, and store the trained model parameters, voiceprint feature data, and their labeling information to form a voiceprint feature library.

[0012] Preferably, the analysis unit makes the following abnormal judgment on the sound data: P1: Obtain the sound data detected by the voiceprint sensor in real time, extract the features of short-time energy, zero-crossing rate, and Mel-frequency cepstral coefficients, and compare the extracted feature data with the feature data corresponding to the preset normal working conditions in the voiceprint feature library; calculate the Euclidean distance data between the two feature data , and the distance data is compared with the preset distance threshold . If , it is determined that the sound data is abnormal; P2: After determining the sound abnormality, compare the three-item feature data corresponding to the sound abnormality with the feature data corresponding to the corresponding faults included in the voiceprint feature library. If the Euclidean distance value between the two is less than the preset distance threshold , it is determined that the abnormal conditions corresponding to the two feature data are the same, a warning signal is generated, and the determined abnormal condition information and warning signal are transmitted to the execution unit; P3: If no similar feature data is obtained after comparison with the feature data corresponding to the corresponding faults included in the voiceprint feature library, analyze the feature data; if the change rates of the three-item feature data all exceed the preset change rate thresholds of the corresponding items at the same time, it is determined that a foreign object has entered, a warning signal is generated, and the determined abnormal condition information and warning signal are transmitted to the execution unit; if the change form of the short-time energy data of the feature data is an increase, the growth rate of the zero-crossing rate exceeds the preset growth rate, and the Mel-frequency cepstral coefficient exceeds the preset normal range, it is determined that the abnormal sound is caused by the increased wear of the equipment components, a warning signal is generated according to the position of the voiceprint sensor, and the abnormal position, the determined abnormal condition information, and the warning signal are transmitted to the execution unit.

[0013] In summary, due to the adoption of the above technical solutions, the beneficial effects of the present invention are: Through the coordinated use of the main control module and the monitoring unit, the safety inside the energy storage power station is monitored, so that a fire in the energy storage power station can be alarmed in time, facilitating timely fire extinguishing and ensuring the safe operation of the energy storage power station; By training with the Gaussian mixture model (GMM) to establish a normal / abnormal voiceprint feature library of the equipment, after comparing the real-time voiceprint data with the feature library, the abnormal type is automatically identified, and a warning signal and a specific fault report are generated; after receiving the warning signal, the execution unit immediately triggers an audible and visual alarm, and transmits the abnormal position, type, and severity to the remote monitoring center in real time through the data transmission module, and the management personnel can synchronously start the emergency plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 Shows the overall structural schematic diagram provided by the embodiment of the present invention; Figure 2 Shows an exploded view provided according to an embodiment of the present invention; Figure 3 Shows a system flow chart provided according to an embodiment of the present invention.

[0015] Legend description: 1. Housing; 2. Housing cover; 3. Acoustic-optic alarm; 4. Monitoring unit; 5. Data transmission module; 6. Storage module; 7. Main control module. Specific implementation manners

[0016] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present invention.

[0017] Please refer to Figures 1 - 3 , the present invention provides a technical solution: An energy storage power station safety monitoring and early warning device includes a housing 1 and a housing cover 2. Inside the housing 1, there are respectively a monitoring unit 4, a data transmission module 5, a storage module 6, a main control module 7 and an alarm module arranged on the bottom surface of the housing 1; wherein, through the mutual cooperation of the main control module 7 and the monitoring unit 4, the safety inside the energy storage power station is monitored, so that a fire in the energy storage power station can be alarmed in time, thereby facilitating timely fire extinguishing and ensuring the safety of the energy storage power station; The data transmission module 5 transmits the monitoring data and alarm information to the remote monitoring center in real time. The management personnel can clearly understand the safety status of the energy storage power station through the system platform of the monitoring center and make corresponding processing decisions in time; The storage module 6 continuously stores the monitoring data in chronological order. When the storage space is insufficient, the earliest data is automatically overwritten; through the analysis of historical data, the operation rules of the energy storage power station can be summarized, potential safety hazards can be discovered in advance, and the safety management level of the energy storage power station can be further improved; The monitoring unit 4 includes a temperature and humidity monitoring module, a smoke monitoring module and a gas concentration monitoring module; The main control module 7 is electrically connected to the temperature and humidity monitoring module, the smoke monitoring module, the gas concentration monitoring module, the data transmission module 5, the storage module 6 and the alarm module respectively.

[0018] In the present invention, the temperature and humidity monitoring module adopts a high-precision temperature and humidity sensor, which is used to monitor the temperature and humidity data inside the energy storage power station in real time and transmit the data to the main control module 7; moreover, the temperature and humidity sensor has the characteristics of high sensitivity and accurate measurement, and can quickly respond to the changes in the environmental temperature and humidity; The smoke monitoring module adopts an infrared beam smoke detector; through the principle of infrared beam occlusion, the smoke concentration in the energy storage power station is monitored. Once the detected smoke concentration exceeds the preset threshold, the signal is transmitted to the main control module 7; The gas concentration monitoring module includes a variety of gas sensors, which are respectively used to monitor the concentrations of harmful gases such as oxygen, hydrogen, and carbon monoxide; these gas sensors can accurately detect the concentration changes of the corresponding gases and timely feedback the data to the main control module 7; In the present invention, the data transmission module 5 adopts a 5G communication module or a wireless network module, which can transmit the monitoring data processed by the main control module 7 to the remote monitoring center in real time, facilitating the management personnel to master the operation safety status of the energy storage power station anytime and anywhere; meanwhile, the data transmission module also supports the function of resuming interrupted transmission to ensure the integrity of data transmission.

[0019] In the present invention, the storage module 6 adopts a large-capacity memory card, which is used to store the historical monitoring data for subsequent data analysis and fault troubleshooting; moreover, the storage module 6 has an automatic overwrite function. When the storage space is insufficient, the earliest data is automatically overwritten to facilitate persistent data storage.

[0020] In the present invention, the alarm module includes a sound and light alarm 3 symmetrically arranged on the bottom surface of the housing 1; when the main control module 7 analyzes and processes the received data, when a certain item of data exceeds the preset safety threshold, the main control module 7 immediately controls the alarm module to start; the sound and light alarm 3 emits strong flashing lights and loud alarm sounds to attract the attention of on-site staff.

[0021] The main control module 7 is also provided with an intelligent control component, and the intelligent control component includes a collection unit, an analysis unit, and an execution unit; The collection unit collects the sound data at the corresponding equipment position in the energy storage power station and transmits the collected sound data to the analysis unit; An analysis unit analyzes the voice data collected by the collection unit, extracts features after preprocessing, and establishes a voiceprint feature library; extracts the feature data of the voice data detected in real time, compares it with the preset normal voice data in the voiceprint feature library, and determines whether an abnormality occurs; if it is determined that an abnormality occurs, it is compared with the preset abnormal voice data in the voiceprint feature library. If it is similar to the feature data of the corresponding abnormal voice in the voiceprint feature library, it is determined that the abnormal voice is caused by the situation corresponding to the corresponding abnormal voice in the voiceprint feature library, generates a warning signal, and transmits the determined abnormal situation information and warning signal to the execution unit; if it is not similar to the feature data of all abnormal voices in the voiceprint feature library, analyze the feature data, generate a corresponding warning signal according to the determined abnormal situation, and transmit the determined abnormal situation information and warning signal to the execution unit; An execution unit receives the warning signal transmitted by the analysis unit, performs a warning operation through an audible and visual alarm 3, and transmits the abnormal situation information and abnormal position information to the system platform for display through a data transmission module 5; Install voiceprint sensors at positions close to the fan, transformer, and battery pack in the energy storage power station to collect voice data during the operation of the equipment; Filter the collected voice signal through a filter, and set the collected voice signal as , is an integer, and the unit impulse response of the filter is , and it is deduced that the voice signal obtained after the filtering operation is , is the index used to traverse the input signal during the convolution process; Amplify the voice signal obtained after the filtering operation to obtain the voice signal ; to ensure that the original analog signal is recovered without distortion from the sampling signal, make the sampling frequency , is the highest frequency of the analog signal, is the sampling time interval; obtain the quantization interval , quantization level and quantization interval , then for the amplified voice signal, the digital voice signal obtained through digital conversion;

[0022] Obtain the digital voice signal within a set time period, divide it into short time periods with a length of , and the short-time energy of the digital voice signal in the th short time period , and the zero-crossing rate of the digital voice signal in the th short time period , the nth digital audio signal in the nth short time period; Process the audio signal using a first-order high-pass filter to obtain a pre-emphasized audio signal , where is the pre-emphasis coefficient; Divide the pre-emphasized audio signal into several frames, and then apply a Hamming window function to each frame of the signal to obtain ; Perform a fast Fourier transform on each windowed frame of the signal, then design a set of triangular filters according to the Mel frequency scale, convert the linear frequency axis to the Mel frequency axis, filter the signal spectrum obtained after the fast Fourier transform to obtain the Mel spectrum, and then perform a discrete cosine transform on the Mel spectrum to obtain Mel frequency cepstral coefficients , , where is the number of selected Mel frequency cepstral coefficients; Obtain the sound data under different working conditions at the positions of the fan, transformer, and battery pack, and extract the short-time energy, zero-crossing rate, and Mel frequency cepstral coefficient features of the obtained sound data. According to the working condition information corresponding to the sound data, label the extracted features. Train the labeled input feature data through a Gaussian mixture model GMM to obtain the rules of the corresponding voiceprint features under the corresponding working condition information, and store the trained model parameters, voiceprint feature data, and their annotation information to form a voiceprint feature library.

[0023] Then obtain the sound data detected in real time by the voiceprint sensor, extract the short-time energy, zero-crossing rate, and Mel frequency cepstral coefficient features, and compare the extracted feature data with the feature data corresponding to the preset normal working condition in the voiceprint feature library; Calculate the Euclidean distance data between the two feature data, and compare the distance data with the preset distance threshold . If , it is determined that the sound data is abnormal; After determining that the sound is abnormal, compare the three feature data corresponding to the sound abnormality with the corresponding feature data of the corresponding fault included in the voiceprint feature library. If the Euclidean distance value between the two is less than the preset distance threshold , it is determined that the abnormal situations corresponding to the two feature data are the same, generate a warning signal, and transmit the determined abnormal situation information and warning signal to the execution unit; If, when comparing with the corresponding feature data of the corresponding faults stored in the voiceprint feature library, no similar feature data is obtained, the feature data is analyzed; if the change rates of all three items of feature data simultaneously exceed the preset change rate thresholds of the corresponding items, it is determined that a foreign object has entered, a warning signal is generated, and the determined abnormal situation information and warning signal are transmitted to the execution unit; if the change form of the short-time energy data of the feature data is an increase, the growth rate of the zero-crossing rate exceeds the preset growth rate, and the Mel frequency cepstral coefficient exceeds the preset normal range, it is determined that the abnormal sound is caused by the aggravated wear of the equipment components, a warning signal is generated according to the position of the voiceprint sensor, and the abnormal position, the determined abnormal situation information and the warning signal are transmitted to the execution unit (the above are two examples of abnormal feature data analysis).

[0024] Working principle: When in use, when a fire occurs in the energy storage power station, the smoke monitoring module and the gas concentration monitoring module in the monitoring unit 4 collect information at this time, and then transmit it to the main control module 7 through the data transmission module 5. After analysis by the main control module 7, the signal is stored in the storage module 6. At the same time, by analyzing the information, when a certain data exceeds the preset safety threshold, the main control module 7 immediately controls the alarm module to start; the audible and visual alarm 3 emits strong flashing lights and loud alarm sounds to attract the attention of on-site staff, playing the role of safety monitoring of the energy storage power station.

[0025] The above description of the embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A safety monitoring and early warning device for an energy storage power station, comprising a housing (1) and a housing cover (2), characterized in that, A monitoring unit (4), a data transmission module (5), a storage module (6), a main control module (7), and an alarm module provided on the bottom surface of the housing (1) are respectively provided inside the housing (1); The monitoring unit (4) includes a temperature and humidity monitoring module, a smoke monitoring module, and a gas concentration monitoring module; The main control module (7) is electrically connected to the temperature and humidity monitoring module, the smoke monitoring module, the gas concentration monitoring module, the data transmission module (5), the storage module (6), and the alarm module respectively; A smart control component is further provided on the main control module (7), and the smart control component includes an analysis unit; The analysis unit analyzes the voice data collected by the collection unit, extracts features after preprocessing, and establishes a voiceprint feature library; extracts the feature data of the voice data detected in real time, and compares it with the preset normal voice data in the voiceprint feature library to determine whether an abnormality occurs; If it is determined that an abnormality occurs, it is compared with the preset abnormal voice data in the voiceprint feature library. If it is similar to the feature data of the corresponding abnormal voice in the voiceprint feature library, it is determined that the abnormal voice is caused by the situation corresponding to the corresponding abnormal voice in the voiceprint feature library, generates a warning signal, and transmits the determined abnormal situation information and warning signal to the execution unit.

2. The safety monitoring and early warning device for an energy storage power station according to claim 1, wherein The temperature and humidity monitoring module uses a high-precision temperature and humidity sensor to continuously monitor the temperature and humidity data inside the energy storage power station and transmit the data to the main control module (7); The smoke monitoring module uses an infrared beam smoke detector; The gas concentration monitoring module includes a variety of gas sensors, which are respectively used to monitor the concentrations of harmful gases such as oxygen, hydrogen, and carbon monoxide.

3. The safety monitoring and early warning device for an energy storage power station according to claim 2, characterized in that, The data transmission module (5) uses a 5G communication module or a wireless network module, and can transmit the monitoring data processed by the main control module (7) to the remote monitoring center in real time, facilitating the management personnel to grasp the operation safety status of the energy storage power station at any time and anywhere.

4. The safety monitoring and early warning device for an energy storage power station according to claim 3, wherein The storage module (6) uses a large-capacity memory card to store the historical data detected for subsequent data analysis and fault troubleshooting.

5. The safety monitoring and early warning device for an energy storage power station according to claim 4, characterized in that, The alarm module includes a sound and light alarm (3) symmetrically arranged on the bottom surface of the housing (1).

6. The safety monitoring and early warning device for an energy storage power station according to claim 1, characterized in that, The smart control component further includes a collection unit and an execution unit; The collection unit collects the voice data at the position of the corresponding equipment in the energy storage power station and transmits the collected voice data to the analysis unit; The execution unit receives the warning signal transmitted by the analysis unit, emits a sound and light alarm through the sound and light alarm (3), and transmits the abnormal situation information and the abnormal position information to the system platform for display through the data transmission module (5).

7. The safety monitoring and early warning device for an energy storage power station according to claim 1, characterized in that, The preprocessing steps of the voice data by the analysis unit are as follows: S1: Set the collected sound signal as , is an integer, the unit impulse response of the filter is , and it is deduced that the sound signal obtained after the filtering operation is , is the index used to traverse the input signal during the convolution process; S2: Obtain the quantization interval of the quantization level , the number of quantization levels and the quantization interval . Then, for the amplified sound signal, the digital sound signal obtained through digital conversion .

8. The safety monitoring and early warning device for an energy storage power station according to claim 1, characterized in that, The feature extraction steps of the voice data by the analysis unit are as follows: Q1: Obtain a digital audio signal within a set time period, and divide it into short time periods with a length of ; the short-time energy of the digital audio signal in the -th short time period, and the zero-crossing rate of the digital audio signal in the -th short time period; is the -th digital audio signal in the -th short time period. Q2: Process the sound signal using a first-order high-pass filter to obtain the pre-emphasized sound signal , is the pre-emphasis coefficient; divide the pre-emphasized sound signal into several frames, and then apply the Hamming window function to each frame of the signal to obtain ; perform a fast Fourier transform on each windowed frame of the signal, then design a set of triangular filters according to the Mel frequency scale, convert the linear frequency axis to the Mel frequency axis, filter the signal spectrum obtained after the fast Fourier transform to obtain the Mel spectrum, and then perform a discrete cosine transform on the Mel spectrum to obtain the Mel frequency cepstral coefficients , , is the number of selected Mel frequency cepstral coefficients; Q3: Acquire sound data under different working conditions at the locations of fans, transformers and battery packs, and extract the short-time energy, zero-crossing rate and Mel-frequency cepstral coefficient features of the acquired sound data. According to the working condition information corresponding to the sound data, annotate the extracted features with the corresponding working condition information. Train the annotated input feature data through the Gaussian mixture model GMM to obtain the rules of the corresponding voiceprint features under the corresponding working condition information. Store the trained model parameters, voiceprint feature data and their annotated information to form a voiceprint feature library.

9. The safety monitoring and early warning device for an energy storage power station according to claim 1, wherein, The analysis unit determines the abnormality of the sound data as follows: P1: Obtain the sound data detected in real time by the voiceprint sensor, extract the features of short-time energy, zero-crossing rate, and Mel-frequency cepstral coefficients, and compare the extracted feature data with the feature data corresponding to the preset normal working conditions in the voiceprint feature library; calculate the Euclidean distance data between the two feature data , and use the distance data to compare with the preset distance threshold . If , it is determined that the sound data is abnormal; P2: After determining that the sound is abnormal, compare the three pieces of characteristic data corresponding to the abnormal sound with the corresponding characteristic data of the corresponding fault included in the voiceprint characteristic library. If the Euclidean distance value between the corresponding characteristic data of the corresponding fault is less than the preset distance threshold , it is determined that the abnormal situations corresponding to the two pieces of characteristic data are the same, a warning signal is generated, and the determined abnormal situation information and the warning signal are transmitted to the execution unit; P3: If no similar feature data is obtained when compared with the corresponding feature data of the corresponding fault included in the voiceprint feature library, the feature data is analyzed; if the change rates of the three feature data all exceed the preset change rate thresholds of the corresponding items at the same time, it is determined that a foreign object has entered, a warning signal is generated, and the determined abnormal situation information and the warning signal are transmitted to the execution unit; if the short-term energy data of the feature data changes in the form of increase, the growth rate of the zero-crossing rate exceeds the preset growth rate, and the Mel frequency cepstrum coefficient exceeds the preset normal range, it is determined that the sound is abnormal due to the increased wear of equipment parts, and a warning signal is generated according to the position of the voiceprint sensor, and the abnormal position, the determined abnormal situation information and the warning signal are transmitted to the execution unit.

Images