Phase modifier rotating equipment fault early warning method and system based on voiceprint recognition

Abstract

The invention belongs to the technical field of phase modifier rotating equipment fault early warning, and particularly relates to a phase modifier rotating equipment fault early warning method and system based on voiceprint recognition. According to the method, the running state of the equipment can be monitored in real time, abnormity can be rapidly recognized through dynamic analysis of voiceprint information, the real-time performance and accuracy of fault detection are improved, the influence of external environments such as vibration and temperature on voiceprint features is considered in a working condition compensation mechanism, the possibility of misjudgment is reduced, and the fault detection accuracy is improved. Through the matching of the predicted voiceprint information and the historical fault voiceprint set data, the current abnormity can be found, the possible fault trend can be recognized in advance, the longer response time is provided for equipment maintenance, an early warning mechanism helps maintenance personnel to take measures before the fault occurs, the equipment is prevented from being damaged or shut down, and the maintenance efficiency is improved. The maintenance cost and the equipment downtime are reduced, the fusion of multi-source information is realized, and the depth and the breadth of data analysis are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of fault early warning for synchronous condenser rotating equipment, and particularly relates to a method and system for fault early warning of synchronous condenser rotating equipment based on voiceprint recognition. Background Art

[0002] Synchronous condenser rotating equipment is widely used in power systems, especially playing an important role in the regulation and stability of high-voltage power grids. Its main function is to adjust the phase angle of the power grid to ensure the frequency stability and load balance of the power grid. However, as the equipment usage time increases, the synchronous condenser rotating equipment is prone to various degrees of faults, resulting in a decline in equipment performance and even more serious system failures. These faults usually manifest as abnormal operation, reduced efficiency, or even shutdown of the equipment, and may seriously affect the stability and safety of the entire power grid when severe.

[0003] Traditional fault detection methods usually rely on regular inspections and manual monitoring. However, these methods often have disadvantages such as long response time and low detection accuracy, and it is difficult to detect potential problems in a timely manner before equipment faults occur. Especially when the equipment has minor faults or early abnormalities, traditional methods may not be able to effectively identify them, resulting in the failure to handle equipment faults in a timely manner, thus increasing the equipment maintenance cost and downtime.

[0004] To solve this problem, in recent years, voiceprint recognition technology, as an emerging means of equipment fault detection, has gradually been applied to the field of equipment fault early warning. Voiceprint recognition technology can identify the operating state and potential faults of equipment by analyzing the sound characteristics generated during equipment operation. Different equipment will generate specific sound patterns during operation, and the abnormal state of the equipment will cause changes in voiceprint characteristics. Therefore, through real-time monitoring of equipment sounds and voiceprint analysis, accurate assessment of equipment status can be achieved.

[0005] However, existing voiceprint recognition technologies still face some challenges in practical applications. The operating environment factors of the equipment will affect the sound signal, resulting in changes in voiceprint characteristics and affecting the accuracy of fault detection. The faults of the equipment are usually progressive, and early faults often manifest as minor sound changes, and traditional voiceprint recognition methods are difficult to accurately capture these subtle changes. Summary of the Invention

[0006] The purpose of the present invention is to provide a method and system for fault early warning of synchronous condenser rotating equipment based on voiceprint recognition, which can accurately judge the operating state of the equipment, identify possible fault trends in advance, and issue early warning signals, thereby improving the accuracy and response speed of fault detection and reducing equipment downtime and maintenance costs.

[0007] The technical solutions adopted by the present invention are specifically as follows:

[0008] A method for early warning of faults in a synchronous condenser rotating device based on voiceprint recognition, comprising:

[0009] Obtain the real-time operating sound information of the synchronous condenser rotating device, and determine whether the real-time operating sound information meets the first preset condition. If not, it is determined that the synchronous condenser rotating device is operating abnormally, and abnormal voiceprint information is obtained;

[0010] Construct an evaluation period, obtain the evaluation operating sound data within the evaluation period, and obtain voiceprint change information according to the evaluation operating sound data;

[0011] Obtain the vibration data and temperature data of the synchronous condenser rotating device during the evaluation period, and obtain the working condition compensation information of the synchronous condenser rotating device according to the vibration data and temperature data;

[0012] Obtain the predicted voiceprint information according to the voiceprint change information, the working condition compensation information and the abnormal voiceprint information;

[0013] Obtain the set data of operating fault voiceprints, obtain the voiceprint similarity data according to the predicted voiceprint information and the set data of operating fault voiceprints, determine whether the voiceprint similarity data meets the second preset condition. If so, mark it as the target voiceprint data, obtain the corresponding fault information according to the target voiceprint data, and send out an alarm message.

[0014] In a preferred solution, the step of obtaining the real-time operating sound information of the synchronous condenser rotating device, determining whether the real-time operating sound information meets the first preset condition, and if not, determining that the synchronous condenser rotating device is operating abnormally and obtaining the abnormal voiceprint information includes:

[0015] Obtain the real-time operating sound information of the synchronous condenser rotating device;

[0016] Extract the corresponding real-time operating sound vector according to the real-time operating sound information;

[0017] Obtain the operating sound threshold vector;

[0018] Determine whether the real-time operating sound vector exceeds the operating sound threshold vector;

[0019] If a certain dimension of the real-time operating sound vector exceeds the corresponding operating sound threshold vector, it is determined that the synchronous condenser rotating device is operating abnormally, and the abnormal voiceprint information in the real-time operating sound information at the time of abnormality is extracted.

[0020] In a preferred solution, the step of constructing an evaluation period, obtaining the evaluation operating sound data within the evaluation period, and obtaining voiceprint change information according to the evaluation operating sound data includes:

[0021] Construct an evaluation period;

[0022] Obtain the operating sound data of the synchronous condenser rotating equipment during the evaluation period and mark it as the evaluated operating sound data;

[0023] Convert the evaluated operating sound data into corresponding multiple evaluated operating voiceprint vectors, and each evaluated operating voiceprint vector represents the sound characteristics at a certain time point or time period during the evaluation period;

[0024] Obtain the voiceprint change value according to the multiple evaluated operating voiceprint vectors and mark it as the voiceprint change information;

[0025] The voiceprint change value is calculated as;

[0026] , where, represents the voiceprint change value, i represents the number of the multiple evaluated operating voiceprint vectors, i = 2, 3, 4…n, n represents the number of the evaluated operating voiceprint vectors, represents the i-th evaluated operating voiceprint vector, represents the (i - 1)-th evaluated operating voiceprint vector.

[0027] In a preferred solution, the steps of constructing the evaluation period include:

[0028] Obtain the time node when it is determined that the synchronous condenser rotating equipment operates abnormally and mark it as the start time;

[0029] Convert the real-time operating sound information at the time of the abnormality into the corresponding real-time operating sound vector;

[0030] Obtain the duration table, where the duration table includes multiple operating sound vectors and the evaluation duration corresponding to each operating sound vector;

[0031] Obtain the corresponding evaluation duration from the duration table according to the real-time operating sound vector;

[0032] Obtain the end time according to the start time and the evaluation duration;

[0033] Obtain the evaluation period according to the start time and the end time.

[0034] In a preferred solution, the steps of obtaining the vibration data and temperature data of the synchronous condenser rotating equipment during the evaluation period and obtaining the working condition compensation information of the synchronous condenser rotating equipment according to the vibration data and temperature data include:

[0035] Obtain the vibration data and temperature data of the synchronous condenser rotating equipment during the evaluation period;

[0036] Convert the vibration data into multiple vibration vectors;

[0037] Convert the temperature data into multiple temperature values, representing the equipment temperature states at different time points;

[0038] Calculate the working condition compensation value based on multiple vibration vectors and temperature values, and mark it as working condition compensation information;

[0039] The working condition compensation value is calculated as:

[0040] ,

[0041] In the formula, represents the working condition compensation value, b represents the number of multiple temperature values, b = 1, 2, 3…m, represents the b-th temperature value, m represents the number of temperature values, d represents the number of multiple vibration vectors, d = 1, 2, 3…u, represents the d-th vibration vector, and u represents the number of vibration vectors.

[0042] In a preferred solution, the steps of obtaining predicted voiceprint information based on voiceprint change information, working condition compensation information, and abnormal voiceprint information include:

[0043] Obtain the corresponding abnormal voiceprint vector according to the abnormal voiceprint information;

[0044] Obtain the voiceprint prediction value according to the voiceprint change information, working condition compensation information, and abnormal voiceprint vector;

[0045] Obtain a prediction table, where the prediction table includes multiple voiceprint prediction intervals and the corresponding predicted voiceprint information for each voiceprint prediction interval;

[0046] Obtain the corresponding predicted voiceprint information from the prediction table according to the voiceprint prediction interval where the voiceprint prediction value is located;

[0047] The voiceprint prediction value is calculated as:

[0048] , in the formula, Y represents the voiceprint prediction value, represents the voiceprint change value, represents the working condition compensation value, and C represents the abnormal voiceprint vector.

[0049] In a preferred solution, the steps of obtaining the running fault voiceprint set data, obtaining the voiceprint similarity data according to the predicted voiceprint information and the running fault voiceprint set data, determining whether the voiceprint similarity data meets the second preset condition, if it meets, marking it as the target voiceprint data, obtaining the corresponding fault information according to the target voiceprint data, and sending out an alarm message include:

[0050] Obtain the running fault voiceprint set data; the running fault voiceprint set data contains the voiceprint feature records of known faults during the operation of the equipment, and each record includes the voiceprint feature value, the fault category, and the corresponding working condition information;

[0051] Convert the data of the running fault voiceprint set into corresponding multiple running fault voiceprint vectors;

[0052] Extract the corresponding predicted voiceprint vector according to the predicted voiceprint information;

[0053] Calculate and obtain the voiceprint similarity value between each running fault voiceprint vector and the predicted voiceprint vector;

[0054] Summarize multiple voiceprint similarity values, and mark the summary result as voiceprint similarity data;

[0055] Judge whether the voiceprint similarity data meets the second preset condition. If it meets, mark it as target voiceprint data, obtain the corresponding fault information according to the target voiceprint data, and send an alarm message;

[0056] The calculation of the voiceprint similarity value is as follows:

[0057] , where, represents the voiceprint similarity value, A represents the predicted voiceprint vector, and B represents the running fault voiceprint vector.

[0058] In a preferred solution, the step of judging whether the voiceprint similarity data meets the second preset condition. If it meets, mark it as target voiceprint data, obtain the corresponding fault information according to the target voiceprint data, and send an alarm message includes:

[0059] Sort the multiple voiceprint similarity values in the voiceprint similarity data from largest to smallest, and mark the sorting result as the voiceprint sorting table;

[0060] Obtain the voiceprint similarity value corresponding to the upper limit value in the voiceprint sorting table, and mark it as the upper limit voiceprint similarity value;

[0061] Obtain the running fault voiceprint corresponding to the upper limit voiceprint similarity value, and mark it as target voiceprint data;

[0062] Obtain the corresponding fault information according to the target voiceprint data, and send an alarm message; the fault information includes the fault type, fault cause and treatment suggestions.

[0063] The present invention also provides a synchronous condenser rotating equipment fault warning system based on voiceprint recognition for the above-mentioned synchronous condenser rotating equipment fault warning method based on voiceprint recognition, including:

[0064] A sound judgment module, configured to obtain the real-time running sound information of the synchronous condenser rotating equipment, judge whether the real-time running sound information meets the first preset condition. If it does not meet, determine that the synchronous condenser rotating equipment is operating abnormally, and obtain abnormal voiceprint information;

[0065] A voiceprint change module, configured to construct an evaluation period, obtain evaluation operation voice data within the evaluation period, and obtain voiceprint change information according to the evaluation operation voice data;

[0066] A working condition compensation module, configured to obtain vibration data and temperature data of the synchronous condenser rotating equipment during the evaluation period, and obtain working condition compensation information of the synchronous condenser rotating equipment according to the vibration data and the temperature data;

[0067] A predicted voiceprint module, configured to obtain predicted voiceprint information according to the voiceprint change information, the working condition compensation information, and the abnormal voiceprint information;

[0068] A voiceprint early warning module, configured to obtain operation fault voiceprint set data, obtain voiceprint similarity data according to the predicted voiceprint information and the operation fault voiceprint set data, determine whether the voiceprint similarity data meets a second preset condition, if so, mark it as target voiceprint data, obtain corresponding fault information according to the target voiceprint data, and send out an alarm message.

[0069] And, a synchronous condenser rotating equipment fault early warning terminal based on voiceprint recognition, including:

[0070] One or more processors;

[0071] A storage device, on which one or more programs are stored;

[0072] When the one or more programs are executed by the one or more processors, the one or more processors implement a synchronous condenser rotating equipment fault early warning method based on voiceprint recognition.

[0073] The technical effects achieved by the present invention are:

[0074] The present invention can monitor the operation state of the equipment in real time, quickly identify abnormalities through dynamic analysis of voiceprint information, improve the real-time performance and accuracy of fault detection. The working condition compensation mechanism takes into account the influence of external environments such as vibration and temperature on voiceprint characteristics, reduces the possibility of misjudgment. By matching the predicted voiceprint information and the historical fault voiceprint set data, not only can the current abnormalities be discovered, but also the possible fault trends can be identified in advance, providing a longer response time for equipment maintenance. The early warning mechanism helps maintenance personnel take measures before the fault occurs, avoid equipment damage or shutdown, reduce the maintenance cost and equipment downtime, realize the fusion of multi-source information, and improve the depth and breadth of data analysis. Description of the Drawings

[0075] Figure 1 is a flowchart of a synchronous condenser rotating equipment fault early warning method provided by an embodiment of the present invention;

[0076] Figure 2It is a module diagram of a fault warning system for a synchronous condenser rotating device based on voiceprint recognition provided by an embodiment of the present invention. Specific embodiments

[0077] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of the specific embodiments of the present invention will be made in conjunction with the accompanying drawings of the specification.

[0078] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0079] Secondly, the so-called "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation manner of the present invention. The phrase "in a preferred embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that excludes other embodiments.

[0080] Thirdly, the present invention will be described in detail in conjunction with the schematic diagrams. When describing the embodiments of the present invention in detail, for the sake of illustration, the schematic diagrams are only examples and should not limit the scope of protection of the present invention herein.

[0081] The first embodiment of the present invention provides a method for fault warning of a synchronous condenser rotating device based on voiceprint recognition. Refer to Figure 1 , including:

[0082] S1. Obtain the real-time operating sound information of the synchronous condenser rotating device, and determine whether the real-time operating sound information meets the first preset condition. If not, it is determined that the synchronous condenser rotating device is operating abnormally, and abnormal voiceprint information is obtained;

[0083] S2. Construct an evaluation period, obtain the evaluation operating sound data within the evaluation period, and obtain voiceprint change information based on the evaluation operating sound data;

[0084] S3. Obtain the vibration data and temperature data of the synchronous condenser rotating device during the evaluation period, and obtain the working condition compensation information of the synchronous condenser rotating device based on the vibration data and temperature data;

[0085] S4. Obtain the predicted voiceprint information based on the voiceprint change information, the working condition compensation information, and the abnormal voiceprint information;

[0086] S5. Obtain the running fault voiceprint set data, obtain the voiceprint similarity data based on the predicted voiceprint information and the running fault voiceprint set data, determine whether the voiceprint similarity data meets the second preset condition. If it meets, mark it as the target voiceprint data, obtain the corresponding fault information based on the target voiceprint data, and send an alarm message.

[0087] In the above steps S1 to S5, by obtaining the real-time running sound information of the synchronous condenser rotating equipment and determining whether it meets the first preset condition (such as the frequency range, energy distribution, specific eigenvalue, etc. of the running sound), the abnormal running state can be quickly identified. If the running sound information does not meet the first preset condition, it is determined that the equipment is running abnormally, and the abnormal voiceprint information is extracted. An evaluation period is constructed to obtain the running sound data within this period, and the voiceprint change information is obtained. By obtaining the equipment vibration data and temperature data, combined with the voiceprint change information, the working condition compensation information is generated to reduce the interference of external factors on the voiceprint analysis and improve the accuracy of fault prediction. Combining the voiceprint change information, the working condition compensation information and the abnormal voiceprint information, the predicted voiceprint information is generated. The predicted voiceprint information can be understood as the possible future running sound characteristics of the equipment. By matching with the running fault voiceprint set data, it is determined whether the predicted voiceprint information has a high similarity with the historical fault voiceprint characteristics (the voiceprint similarity data meets the second preset condition). If the matching is successful, the target voiceprint data is marked, and an alarm is issued in combination with the fault information to prompt the maintenance personnel to take measures. It can monitor the running state of the equipment in real time and quickly identify abnormalities through the dynamic analysis of the voiceprint information, improving the real-time performance and accuracy of fault detection. The working condition compensation mechanism considers the influence of external environments such as vibration and temperature on the voiceprint characteristics, reducing the possibility of misjudgment. By matching the predicted voiceprint information and the historical fault voiceprint set data, not only the current abnormalities can be found, but also the possible fault trends can be identified in advance, providing a longer response time for equipment maintenance. The early warning mechanism helps the maintenance personnel take measures before the fault occurs, avoiding equipment damage or shutdown, reducing the maintenance cost and equipment downtime, realizing the integration of multi-source information, and improving the depth and breadth of data analysis.

[0088] In a preferred embodiment, the steps of obtaining the real-time running sound information of the synchronous condenser rotating equipment, determining whether the real-time running sound information meets the first preset condition, and if not, determining that the synchronous condenser rotating equipment is running abnormally and obtaining the abnormal voiceprint information include:

[0089] S101. Obtain the real-time running sound information of the synchronous condenser rotating equipment;

[0090] S102. Extract the corresponding real-time running sound vector according to the real-time running sound information;

[0091] S103. Obtain the running sound threshold vector;

[0092] S104. Determine whether the real-time operating sound vector exceeds the operating sound threshold vector;

[0093] If a certain dimension of the real-time operating sound vector exceeds the corresponding operating sound threshold vector, it is determined that the rotating device of the synchronous condenser operates abnormally, and the abnormal voiceprint information in the real-time operating sound information when it is determined that the rotating device of the synchronous condenser operates abnormally is extracted.

[0094] In the above steps S101 to S104, the operating sound information of the rotating device of the synchronous condenser is obtained in real time through sensors. These sound data usually exist in the form of waveforms or spectra. The obtained sound information is processed and converted into a real-time operating sound vector. The sound vector is a digital representation method, which may include the frequency distribution, energy amplitude, time-domain characteristics, etc. of the sound. According to the normal operating state data or historical operating data of the device, an operating sound threshold vector is predefined, and the real-time operating sound vector is compared with the operating sound threshold vector. If some dimensions of the real-time operating sound vector exceed the corresponding threshold range (for example, the energy of a certain frequency component increases abnormally), it is determined that the device operates abnormally, and the sound feature information at the time of abnormality, that is, the abnormal voiceprint information, is extracted. These abnormal voiceprint information contains the sound characteristics when the device operates abnormally. In this embodiment, the sound characteristics can be analyzed in a quantitative form, avoiding the difficulty of directly processing complex sound signals, improving the accuracy and efficiency of the analysis, being able to flexibly adapt to the operating characteristics of different devices or different working conditions, reducing misjudgment and missed judgment, and quickly discovering abnormal changes in the sound characteristics, so as to quickly determine whether the operating state of the device is abnormal.

[0095] In a preferred embodiment, the steps of constructing an evaluation period, obtaining the evaluation operating sound data within the evaluation period, and obtaining the voiceprint change information according to the evaluation operating sound data include:

[0096] S201. Construct an evaluation period;

[0097] S202. Obtain the operating sound data of the rotating device of the synchronous condenser within the evaluation period and mark it as the evaluation operating sound data;

[0098] S203. Convert the evaluation operating sound data into corresponding multiple evaluation operating voiceprint vectors;

[0099] S204. Obtain the voiceprint change value according to the multiple evaluation operating voiceprint vectors and mark it as the voiceprint change information.

[0100] In the above steps S201 to S204, according to the operating characteristics and monitoring requirements of the synchronous condenser rotating equipment, an evaluation period is set. The evaluation period can be a fixed time interval (such as 10 minutes or 1 hour), or it can be dynamically adjusted according to the equipment operating state. During the evaluation period, the operating sound data of the synchronous condenser rotating equipment is continuously collected and marked as evaluation operating sound data. The evaluation operating sound data is processed and converted into multiple evaluation operating voiceprint vectors. Each voiceprint vector represents the sound characteristics (such as spectral characteristics, time-domain characteristics, etc.) at a certain time point or time period within the evaluation period. The voiceprint change value is calculated through multiple evaluation operating voiceprint vectors. The calculation formula of the voiceprint change value is , where represents the voiceprint change value, i represents the number of multiple evaluation operating voiceprint vectors, i = 2, 3, 4... n, represents the i-th evaluation operating voiceprint vector, represents the (i - 1)-th evaluation operating voiceprint vector. In this embodiment, the dynamic changes of the equipment operating state within a certain time range can be captured, avoiding misjudgment caused by data fluctuations at a single moment, and accurately reflecting the change trend of the equipment operating sound characteristics, providing a reliable basis for fault prediction and operating state evaluation. The flexible setting of the evaluation period can adapt to the monitoring requirements of different equipment and different operating scenarios. For example, for equipment with frequent operation changes, a shorter evaluation period can be set; for equipment with stable operation, a longer evaluation period can be set.

[0101] In a preferred embodiment, the steps of constructing the evaluation period include:

[0102] S2011. Obtain the time node when it is determined that the synchronous condenser rotating equipment operates abnormally and mark it as the start time;

[0103] S2012. Convert the real-time operating sound information at the time of the abnormality into the corresponding real-time operating sound vector;

[0104] S2013. Obtain a duration table, where the duration table includes multiple operating sound vectors and the evaluation duration corresponding to each operating sound vector;

[0105] S2014. Obtain the corresponding evaluation duration from the duration table according to the real-time operating sound vector;

[0106] S2015. Obtain the end time according to the start time and the evaluation duration;

[0107] S2016. Obtain the evaluation period according to the start time and the end time.

[0108] In the above steps S2011 to S2016, when the synchronous condenser rotating device is determined to be operating abnormally, the time node of the occurrence of the abnormality is recorded and marked as the start time of the evaluation period. When the abnormality occurs, the corresponding real-time operating sound information is obtained and converted into a real-time operating sound vector. A duration table is pre-constructed. The duration table contains multiple operating sound interval vectors (intervals representing different operating sound characteristics) and their corresponding evaluation durations. Different sound characteristics may correspond to different evaluation durations to adapt to the complexity of the device operating state. For example, abnormal high-frequency sounds may require a shorter evaluation duration, while low-frequency fluctuations may require longer observation time. According to the real-time operating sound vector, the corresponding evaluation duration is matched from the duration table. Through the dynamic matching of the sound characteristics and the duration table, the length of the evaluation period can be flexibly adjusted according to the actual situation, avoiding the inapplicability that may be brought by a fixed duration. The start time and the evaluation duration are added to calculate the end time of the evaluation period. The determination of the end time ensures that the evaluation period covers the critical time period after the occurrence of the abnormality, providing a guarantee for a comprehensive analysis of the abnormal characteristics. According to the start time and the end time, the time range of the evaluation period is finally determined. In this embodiment, the initial state of the device abnormality and its subsequent change process can be captured. Through the flexible matching of the evaluation duration, the time waste or data insufficiency that may be caused by a fixed evaluation period is avoided, and the utilization efficiency of the monitoring resources is improved.

[0109] In a preferred embodiment, the steps of obtaining the vibration data and temperature data of the synchronous condenser rotating device during the evaluation period and obtaining the operating condition compensation information of the synchronous condenser rotating device according to the vibration data and temperature data include:

[0110] S301. Obtain the vibration data and temperature data of the synchronous condenser rotating device during the evaluation period;

[0111] S302. Convert the vibration data into corresponding multiple vibration vectors;

[0112] S303. Convert the temperature data into corresponding multiple temperature values;

[0113] S304. Obtain an operating condition compensation value according to the multiple vibration vectors and temperature values, and mark it as the operating condition compensation information.

[0114] In the above steps S301 to S304, vibration data and temperature data of the synchronous condenser rotating equipment are continuously collected during the evaluation period. The vibration data reflects the mechanical vibration characteristics during equipment operation, and the temperature data reflects the thermal condition characteristics during equipment operation. The two together constitute the basic information of the equipment operation state. The collected vibration data is converted into multiple vibration vectors. The vibration vector is a quantitative representation of the vibration characteristics, including information such as frequency, amplitude, and acceleration. The temperature data collected during the evaluation period is converted into multiple temperature values, representing the equipment temperature state at different time points. According to the vibration vectors and temperature values, combined with a preset working condition model or compensation algorithm, a working condition compensation value is calculated. The calculation formula of the working condition compensation value is , where represents the working condition compensation value, b represents the number of multiple temperature values, b = 1, 2, 3…m, represents the b-th temperature value, m represents the number of temperature values, d represents the number of multiple vibration vectors, d = 1, 2, 3…u, represents the d-th vibration vector, u represents the number of vibration vectors. The working condition compensation value is marked as working condition compensation information, which can comprehensively reflect the operation state of the synchronous condenser rotating equipment. In this embodiment, the accuracy and reliability of the working condition compensation information are improved. The working condition compensation information can effectively eliminate the interference of vibration and temperature on the acoustic fingerprint characteristics of the equipment, provide a more pure sound characteristic for the analysis of acoustic fingerprint changes, and improve the accuracy of fault prediction.

[0115] In a preferred embodiment, the steps of obtaining predicted acoustic fingerprint information according to the acoustic fingerprint change information, working condition compensation information, and abnormal acoustic fingerprint information include:

[0116] S401. Obtain the corresponding abnormal acoustic fingerprint vector according to the abnormal acoustic fingerprint information;

[0117] S402. Obtain the acoustic fingerprint prediction value according to the acoustic fingerprint change information, working condition compensation information, and abnormal acoustic fingerprint vector;

[0118] S403. Obtain a prediction table, where the prediction table includes multiple acoustic fingerprint prediction interval values and the corresponding predicted acoustic fingerprint information for each acoustic fingerprint prediction interval value;

[0119] S404. Obtain the corresponding predicted acoustic fingerprint information from the prediction table according to the acoustic fingerprint prediction interval where the acoustic fingerprint prediction value is located.

[0120] In the above steps S401 to S404, the abnormal acoustic fingerprint vector is a quantitative representation of the abnormal operation sound characteristics, including information such as spectral characteristics and time-domain characteristics, and is used to characterize the acoustic fingerprint characteristics when the equipment operates abnormally. Combining the acoustic fingerprint change information, working condition compensation information, and abnormal acoustic fingerprint vector, the acoustic fingerprint prediction value is calculated. The calculation formula of the acoustic fingerprint prediction value is , where Y represents the acoustic fingerprint prediction value, Represented as the voiceprint change value, It is represented as the working condition compensation value, C is represented as the abnormal voiceprint vector, and the voiceprint prediction value comprehensively considers the current state and historical change trend of the equipment, reflects the possible voiceprint characteristics of the equipment in the future, and presets a prediction table, which contains multiple voiceprint prediction interval values. Each interval value corresponds to a predicted voiceprint information. The predicted voiceprint information may include the category, state label or risk level of future voiceprint characteristics, which is used to support fault warning. According to the calculated voiceprint prediction value, the voiceprint prediction interval value in the prediction table is matched to obtain the corresponding predicted voiceprint information, which realizes a comprehensive evaluation of the equipment operation status. The prediction result is more accurate and reliable, and can dynamically predict the future operation voiceprint characteristics of the equipment.

[0121] In a preferred embodiment, the steps of obtaining a set of operation fault voiceprint data, obtaining voiceprint similarity data according to the predicted voiceprint information and the set of operation fault voiceprint data, determining whether the voiceprint similarity data meets a second preset condition, and if so, marking it as target voiceprint data, obtaining corresponding fault information according to the target voiceprint data, and issuing an alarm message include:

[0122] S501, obtaining operation fault voiceprint collection data;

[0123] S502, converting the operation fault voiceprint set data into a corresponding plurality of operation fault voiceprint vectors;

[0124] S503, extracting a corresponding predicted voiceprint vector according to the predicted voiceprint information;

[0125] S504, obtaining a voiceprint similarity value between each running fault voiceprint vector and the predicted voiceprint vector;

[0126] S505, summarizing multiple voiceprint similarity values, and marking the summary result as voiceprint similarity data;

[0127] S506: Determine whether the voiceprint similarity data meets the second preset condition. If so, mark it as target voiceprint data, obtain corresponding fault information according to the target voiceprint data, and issue an alarm.

[0128] As in the above steps S501 to S506, the operation fault voiceprint collection data is extracted from the fault database, the collection data contains the voiceprint feature records of known faults during the operation of the equipment, each record includes the voiceprint feature value, the fault category and the corresponding working condition information, the operation fault voiceprint collection data is converted into multiple operation fault voiceprint vectors, the vector represents the feature value of each fault voiceprint, the predicted voiceprint vector is extracted according to the predicted voiceprint information, the predicted voiceprint vector is a quantitative representation of the future operation voiceprint characteristics of the equipment, the similarity value between each operation fault voiceprint vector and the predicted voiceprint vector is calculated, and the calculation formula of the voiceprint similarity value is: , where is represented as the voiceprint similarity value, A is represented as the predicted voiceprint vector, B is represented as the operating fault voiceprint vector. Multiple similarity values are aggregated, and the aggregation result is marked as voiceprint similarity data. It is judged whether the voiceprint similarity data meets the second preset condition (such as selecting the largest value from multiple values). If it meets the condition, it is marked as the target voiceprint data, and the fault information corresponding to the target voiceprint data is extracted from the fault database to generate an alarm message, indicating the possible fault type and recommended handling measures of the device. In this embodiment, the possible fault type corresponding to the predicted voiceprint can be accurately located, the accuracy of fault warning is improved, complex calculations can be completed in a short time, real-time fault warning is realized, dynamic prediction of possible future faults of the device is achieved, and forward-looking guidance is provided for operation and maintenance.

[0129] In a preferred embodiment, the steps of judging whether the voiceprint similarity data meets the second preset condition, if it meets, marking it as the target voiceprint data, obtaining the corresponding fault information according to the target voiceprint data, and sending out an alarm message include:

[0130] S5061. Sort the multiple voiceprint similarity values in the voiceprint similarity data in descending order, and mark the sorting result as the voiceprint sorting table;

[0131] S5062. Obtain the voiceprint similarity value corresponding to the upper limit value in the voiceprint sorting table, and mark it as the upper limit voiceprint similarity value;

[0132] S5063. Obtain the operating fault voiceprint corresponding to the upper limit voiceprint similarity value, and mark it as the target voiceprint data;

[0133] S5064. Obtain the corresponding fault information according to the target voiceprint data, and send out an alarm message.

[0134] In the above steps S5061 to S5064, the multiple voiceprint similarity values in the voiceprint similarity data are sorted in descending order to generate a voiceprint sorting table, and the highest voiceprint similarity value is extracted from the voiceprint sorting table and marked as the upper limit voiceprint similarity value. This value represents the maximum matching degree between the predicted voiceprint vector and the operating fault voiceprint vector. According to the upper limit voiceprint similarity value, the corresponding operating fault voiceprint data is located and marked as the target voiceprint data. The corresponding fault information, including the fault type, possible cause and handling suggestions, is extracted from the target voiceprint data, and an alarm message is sent to notify the operation and maintenance personnel of the possible fault risk and countermeasures, ensuring that the most likely fault type can be quickly focused on, shortening the warning response time, reducing the processing of irrelevant voiceprint data, and reducing the misjudgment rate.

[0135] Based on the same inventive concept, the second embodiment of the present invention further provides a fault warning system for a synchronous condenser rotating device based on voiceprint recognition, which is used for the above-mentioned fault warning method for a synchronous condenser rotating device based on voiceprint recognition. Refer to Figure 2 , the system includes:

[0136] A sound judgment module, configured to obtain real-time operation sound information of the synchronous condenser rotating device, judge whether the real-time operation sound information meets a first preset condition. If not, it is determined that the synchronous condenser rotating device is operating abnormally, and abnormal voiceprint information is obtained;

[0137] A voiceprint change module, configured to construct an evaluation period, obtain evaluation operation sound data within the evaluation period, and obtain voiceprint change information according to the evaluation operation sound data;

[0138] A working condition compensation module, configured to obtain vibration data and temperature data of the synchronous condenser rotating device during the evaluation period, and obtain working condition compensation information of the synchronous condenser rotating device according to the vibration data and the temperature data;

[0139] A predicted voiceprint module, configured to obtain predicted voiceprint information according to the voiceprint change information, the working condition compensation information, and the abnormal voiceprint information;

[0140] A voiceprint warning module, configured to obtain operation fault voiceprint set data, obtain voiceprint similarity data according to the predicted voiceprint information and the operation fault voiceprint set data, judge whether the voiceprint similarity data meets a second preset condition. If so, it is marked as target voiceprint data, and corresponding fault information is obtained according to the target voiceprint data, and an alarm message is sent.

[0141] As described above, the sound judgment module obtains the real-time operation sound information of the synchronous condenser rotating device, compares the real-time sound information with the preset conditions to judge whether the device operation state is abnormal. If it is abnormal, the abnormal voiceprint information is extracted. When the device is abnormal, the voiceprint change module constructs a specific evaluation period, obtains the sound data within the evaluation period, extracts multiple voiceprint feature vectors, and calculates the voiceprint change value. The working condition compensation module obtains the vibration data and temperature data within the evaluation period, generates the working condition compensation information according to the vibration and temperature data to correct the influence of the external environment change on the voiceprint features. The predicted voiceprint module fuses the voiceprint change information, the working condition compensation information and the abnormal voiceprint information, calculates the predicted voiceprint value, and obtains the predicted voiceprint information by looking up the prediction table. The voiceprint warning module obtains the set data of the device operation fault voiceprint, matches the predicted voiceprint information with the fault voiceprint set, calculates the voiceprint similarity value, summarizes the voiceprint similarity data, and judges whether it meets the second preset condition. If it meets, it is marked as the target voiceprint data, the corresponding fault information is extracted, and an alarm message is sent to notify the operation and maintenance personnel to take measures. It can obtain the sound data and working condition information of the device in real time, quickly judge the operation state and potential faults of the device through the collaborative work of multiple modules, can dynamically adapt to the device operation state and external environment changes, improve the accuracy of fault judgment, can analyze the operation state of the device from multiple dimensions, and provide a more comprehensive fault warning.

[0142] Moreover, a synchronous condenser rotating device fault warning terminal based on voiceprint recognition includes:

[0143] One or more processors;

[0144] A storage device on which one or more programs are stored;

[0145] When the one or more programs are executed by the one or more processors, the one or more processors implement the synchronous condenser rotating device fault warning method based on voiceprint recognition.

[0146] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0147] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices produce means for implementing the functions specified in the flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0148] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in the flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0149] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in the flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0150] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: the specific embodiments of the present invention can still be modified or equivalently replaced, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. A method for early warning of faults in a synchronous condenser rotating device based on voiceprint recognition, characterized in that, Including: Obtain the real-time operating sound information of the synchronous condenser rotating equipment, determine whether the real-time operating sound information meets the first preset condition. If not, determine that the synchronous condenser rotating equipment is operating abnormally, and obtain abnormal voiceprint information; Construct an evaluation period, obtain the evaluation operating sound data within the evaluation period, and obtain voiceprint change information based on the evaluation operating sound data; Obtain the vibration data and temperature data of the synchronous condenser rotating equipment during the evaluation period, and obtain the working condition compensation information of the synchronous condenser rotating equipment based on the vibration data and temperature data; Obtain predicted voiceprint information based on the voiceprint change information, working condition compensation information, and abnormal voiceprint information; Obtain the operating fault voiceprint set data, obtain voiceprint similarity data based on the predicted voiceprint information and the operating fault voiceprint set data, determine whether the voiceprint similarity data meets the second preset condition. If so, mark it as target voiceprint data, obtain the corresponding fault information based on the target voiceprint data, and send an alarm message.

2. The fault warning method for a synchronous condenser rotating device based on voiceprint recognition according to claim 1, characterized in that The step of obtaining the real-time operating sound information of the synchronous condenser rotating equipment, determining whether the real-time operating sound information meets the first preset condition. If not, determining that the synchronous condenser rotating equipment is operating abnormally, and obtaining abnormal voiceprint information includes: Obtain the real-time operating sound information of the synchronous condenser rotating equipment; Extract the corresponding real-time operating sound vector based on the real-time operating sound information; Obtain the operating sound threshold vector; Determine whether the real-time operating sound vector exceeds the operating sound threshold vector; If a certain dimension of the real-time operating sound vector exceeds the corresponding operating sound threshold vector, determine that the synchronous condenser rotating equipment is operating abnormally, and extract the abnormal voiceprint information in the real-time operating sound information at the time of abnormality.

3. The fault warning method for a synchronous condenser rotating device based on voiceprint recognition according to claim 1, wherein, The step of constructing an evaluation period, obtaining the evaluation operating sound data within the evaluation period, and obtaining voiceprint change information based on the evaluation operating sound data includes: Construct an evaluation period; Obtain the operating sound data of the synchronous condenser rotating equipment within the evaluation period, and mark it as evaluation operating sound data; Convert the evaluation operating sound data into corresponding multiple evaluation operating voiceprint vectors, and each evaluation operating voiceprint vector represents the sound characteristics at a certain time point or time period within the evaluation period; Obtain the voiceprint change value based on the multiple evaluation operating voiceprint vectors, and mark it as voiceprint change information; The voiceprint change value is calculated as; , where represents the voiceprint change value, i represents the number of multiple evaluated running voiceprint vectors, i = 2, 3, 4... n, and n represents the number of evaluated running voiceprint vectors, represents the i-th evaluated running voiceprint vector, represents the (i - 1)-th evaluated running voiceprint vector.

4. The fault warning method for a synchronous condenser rotating device based on voiceprint recognition according to claim 3, wherein The step of constructing the evaluation period includes: Obtain the time node when it is determined that the synchronous condenser rotating equipment is operating abnormally, and mark it as the start time; Convert the real-time operating sound information at the time of abnormality into the corresponding real-time operating sound vector; Obtain a duration table, where the duration table includes multiple operating sound vectors and the evaluation duration corresponding to each operating sound vector; Obtain the corresponding evaluation duration from the duration table based on the real-time operating sound vector; Obtain the end time based on the start time and the evaluation duration; Obtain the evaluation period based on the start time and the end time.

5. The fault warning method for a synchronous condenser rotating device based on voiceprint recognition according to claim 1, characterized in that, The step of obtaining the vibration data and temperature data of the synchronous condenser rotating equipment during the evaluation period, and obtaining the working condition compensation information of the synchronous condenser rotating equipment based on the vibration data and temperature data includes: Obtain the vibration data and temperature data of the synchronous condenser rotating equipment during the evaluation period; Convert the vibration data into multiple vibration vectors; Convert the temperature data into multiple temperature values, representing the equipment temperature status at different time points; Calculate the operating condition compensation value based on the multiple vibration vectors and temperature values, and mark it as operating condition compensation information; The operating condition compensation value is calculated as: ， Wherein, represents the working condition compensation value, b represents the serial number of multiple temperature values, b = 1, 2, 3…m, represents the b-th temperature value, m represents the number of temperature values, d represents the serial number of multiple vibration vectors, d = 1, 2, 3…u, represents the d-th vibration vector, and u represents the number of vibration vectors.

6. The fault warning method for a synchronous condenser rotating device based on voiceprint recognition according to claim 5, wherein, The step of obtaining the predicted voiceprint information according to the voiceprint change information, operating condition compensation information and abnormal voiceprint information includes: Obtain the corresponding abnormal voiceprint vector according to the abnormal voiceprint information; Obtain the voiceprint prediction value according to the voiceprint change information, operating condition compensation information and abnormal voiceprint vector; Obtain a prediction table, where the prediction table includes multiple voiceprint prediction intervals and the corresponding predicted voiceprint information for each voiceprint prediction interval; Obtain the corresponding predicted voiceprint information from the prediction table according to the voiceprint prediction interval where the voiceprint prediction value is located; The voiceprint prediction value is calculated as: , where Y represents the voiceprint prediction value, represents the voiceprint change value, represents the working condition compensation value, and C represents the abnormal voiceprint vector.

7. The fault warning method for a synchronous condenser rotating device based on voiceprint recognition according to claim 6, characterized in that The step of obtaining the set of running fault voiceprint data, obtaining the voiceprint similarity data according to the predicted voiceprint information and the set of running fault voiceprint data, judging whether the voiceprint similarity data meets the second preset condition, if it meets, marking it as the target voiceprint data, obtaining the corresponding fault information according to the target voiceprint data, and sending an alarm message includes: Obtain the set of running fault voiceprint data; the set of running fault voiceprint data contains the voiceprint feature records of known faults during the operation of the equipment, and each record includes a voiceprint feature value, a fault category and the corresponding operating condition information; Convert the set of running fault voiceprint data into the corresponding multiple running fault voiceprint vectors; Extract the corresponding predicted voiceprint vector according to the predicted voiceprint information; Calculate and obtain the voiceprint similarity value between each running fault voiceprint vector and the predicted voiceprint vector; Summarize the multiple voiceprint similarity values and mark the summary result as voiceprint similarity data; Judge whether the voiceprint similarity data meets the second preset condition, if it meets, mark it as the target voiceprint data, obtain the corresponding fault information according to the target voiceprint data, and send an alarm message; The voiceprint similarity value is calculated as: , where represents the voiceprint similarity value, A represents the predicted voiceprint vector, and B represents the voiceprint vector of the running fault.

8. The method for early warning of the failure of the synchronous condenser rotating equipment based on voiceprint recognition according to claim 7, characterized in that, The step of judging whether the voiceprint similarity data meets the second preset condition, if it meets, marking it as the target voiceprint data, obtaining the corresponding fault information according to the target voiceprint data, and sending an alarm message includes: Sort the multiple voiceprint similarity values in the voiceprint similarity data in descending order, and mark the sorting result as the voiceprint sorting table; Obtain the voiceprint similarity value corresponding to the upper limit value in the voiceprint sorting table, and mark it as the upper limit voiceprint similarity value; Obtain the running fault voiceprint corresponding to the upper limit voiceprint similarity value, and mark it as the target voiceprint data; Obtain the corresponding fault information according to the target voiceprint data, and send an alarm message; the fault information includes the fault type, fault cause and treatment suggestions.

9. A synchronous condenser rotating equipment fault warning system based on voiceprint recognition, which is applied to the synchronous condenser rotating equipment fault warning method based on voiceprint recognition according to any one of claims 1 to 8, and is characterized in that, Includes: A sound judgment module, configured to obtain the real-time operation sound information of the synchronous condenser rotating equipment, judge whether the real-time operation sound information meets the first preset condition, if not, determine that the synchronous condenser rotating equipment is operating abnormally, and obtain abnormal voiceprint information; A voiceprint change module, configured to construct an evaluation period, obtain evaluation operation voice data within the evaluation period, and obtain voiceprint change information according to the evaluation operation voice data; A working condition compensation module, configured to obtain vibration data and temperature data of the synchronous condenser rotating equipment during the evaluation period, and calculate and obtain the working condition compensation information of the synchronous condenser rotating equipment according to the vibration data and the temperature data; A predicted voiceprint module, configured to obtain predicted voiceprint information according to the voiceprint change information, the working condition compensation information, and the abnormal voiceprint information; A voiceprint warning module, configured to obtain operation fault voiceprint set data, obtain voiceprint similarity data according to the predicted voiceprint information and the operation fault voiceprint set data, determine whether the voiceprint similarity data meets a second preset condition, if so, mark it as target voiceprint data, obtain corresponding fault information according to the target voiceprint data, and send out a warning message.

10. A fault warning terminal for a synchronous condenser rotating device based on voiceprint recognition, characterized in that, Including: One or more processors; A storage device, on which one or more programs are stored; When the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the method for warning of faults in a synchronous condenser rotating equipment based on voiceprint recognition according to any one of claims 1 to 8.

Images