Multi-parameter calibration evaluation method and system based on cross-modal parameters

Abstract

The invention discloses a multi-parameter calibration evaluation method and system based on cross-modal parameters, and relates to the technical field of data processing. Acquiring electrical data of target power equipment in the target area, and calculating an electrical operation coefficient according to the electrical data; acquiring temperature data of the target power equipment in the target area from multiple perspectives, and fusing according to the temperature data to obtain a temperature characteristic distribution diagram; acquiring a sound wave signal set of the target power equipment through the microphone array, and fitting according to the sound wave signal set to obtain a sound field distribution diagram; calculating an equipment state coefficient according to the electrical operation coefficient, the sound field distribution diagram and the temperature characteristic distribution diagram, and evaluating the operation state of the power equipment according to the equipment state coefficient; the electrical operation, temperature distribution and sound field characteristic data of the electrical equipment in the target area are accurately obtained and analyzed to calculate the equipment state coefficient, accurate evaluation of the operation state of the electrical equipment is realized, and the accuracy and efficiency of fault early warning and state monitoring are effectively improved.

Description

Technical Field

[0001] The present invention belongs to the field of data processing technology, and in particular relates to a multi-parameter calibration evaluation method and system based on cross-modal parameters. Background Art

[0002] With the rapid development of smart grid technology, the main grid operation status assessment is increasingly dependent on multimodal data analysis. Traditional single data sources can no longer meet the needs of comprehensive and accurate monitoring of complex power grid systems. Multimodal data analysis can capture more dimensional main grid (especially main grid equipment) operation status information by integrating multiple types of data from different sensors, monitoring systems and historical records, providing data support for the main grid operation assessment.

[0003] By conducting multimodal data monitoring on the main network, mainly by collecting multimodal monitoring data and inspection data of the main network equipment in the main network, and performing fusion processing and encoding, and then using image judgment models, knowledge graphs and other technical means to perform fault identification and early warning, real-time monitoring of the main network status and fault early warning can be achieved.

[0004] The above technology has the following problems: the collected data is not comprehensive enough, and the operating status of the power equipment (main grid equipment) cannot be displayed three-dimensionally from the data aspect, resulting in inaccurate analysis results. Summary of the invention

[0005] The purpose of the present invention is to solve the problem that the collected data is not comprehensive enough and the operating status of the power equipment cannot be displayed three-dimensionally from the data aspect, resulting in inaccurate analysis results, and to propose a multi-parameter calibration evaluation method and system based on cross-modal parameters.

[0006] In a first aspect of the present invention, a multi-parameter calibration evaluation method based on cross-modal parameters is first proposed, the method comprising:

[0007] Acquire electrical data of target power equipment in the target area, and calculate the electrical operation coefficient according to the electrical data; the electrical data includes: actual power and apparent power;

[0008] Acquire temperature data of target power equipment in a target area from multiple perspectives, and obtain a temperature characteristic distribution map based on the fusion of the temperature data;

[0009] Acquire a sound wave signal set of the target power equipment through a microphone array, and obtain a sound field distribution diagram according to the sound wave signal set;

[0010] The equipment state coefficient is calculated according to the electrical operation coefficient, the sound field distribution diagram and the temperature characteristic distribution diagram, and the operation state of the electric equipment is evaluated according to the equipment state coefficient.

[0011] Optionally, obtaining a temperature characteristic distribution map according to the temperature data fusion includes:

[0012] Acquire a visible light image and an infrared image of the target power equipment, extract spatial features in the visible light image, and identify pixel features in the infrared image;

[0013] Dividing the pixel features into regional blocks according to the number of channels, converting the regional blocks into frequency spectra by discrete cosine transform, and encoding the frequency spectra to obtain a two-dimensional matrix;

[0014] Flatten the two-dimensional matrix to obtain a one-dimensional array, reshape the one-dimensional array to obtain a three-dimensional tensor, and reconstruct each three-dimensional tensor to obtain the target three-dimensional tensor;

[0015] Downsampling the target three-dimensional tensor to obtain a high-frequency signal and a low-frequency signal, and fusing the high-frequency signal and the low-frequency signal to obtain a target frequency feature;

[0016] The target frequency feature is input into the machine learning model to obtain a weight map, and the spatial feature is added to the residual path to obtain a temperature feature distribution map.

[0017] Optionally, the steps of deploying the microphone array include:

[0018] The target microphone is taken as the central node, and the microphones are deployed at preset intervals with the central node as the center to obtain a target circular array; the circular array includes multiple nodes;

[0019] The central node is used as an endpoint, and the nodes on the circular array are used as directions to determine the node positioning according to a preset interval to obtain a multi-ring array; the multi-ring array includes a plurality of circular arrays;

[0020] Determine a rotation reference point according to the target circular ring array, and rotate each circular ring array according to P=(i-1)θ according to the rotation reference point to obtain a calibration multi-ring array; P is the rotation angle, θ is the preset rotation angle, and i is the number of rings;

[0021] Sound is collected according to the calibrated multi-ring array to obtain a sound data set, and the sound data set is deployed in three-dimensional space according to the spatial characteristics to obtain a sound field distribution map.

[0022] Optionally, the device state coefficient calculated according to the electrical operation coefficient, the sound field distribution diagram, and the temperature characteristic distribution diagram includes:

[0023] Extracting an abnormal sound data set and abnormal temperature data according to the sound field distribution map and the temperature characteristic distribution map; the abnormal sound data set includes the abnormal sound data sets of each circular array;

[0024] Calculate the equipment status coefficient according to the electrical operation coefficient, the abnormal sound data set and the abnormal temperature data;

[0025] Equipment status coefficient calculation formula:

[0026]

[0027] Among them, F is the equipment status coefficient, e is the natural constant, D is the electrical operation coefficient, S i is the abnormal sound data set of the ith ring, n is the total number of rings, T j Abnormal temperature data in the temperature characteristic distribution diagram, m is the number of abnormal temperature areas, α, β i and γ are constant proportional coefficients and are not zero, β i The value decreases as the number of rings increases.

[0028] Optionally, evaluating the operating state of the power equipment according to the equipment state coefficient includes:

[0029] If the equipment status coefficient is greater than the status threshold, the target power equipment is judged to be in an abnormal state, the acquisition frequency of multi-parameter data is increased, and a fault alarm is issued;

[0030] If the device status coefficient is not greater than the status threshold, it is determined that the target power device is in normal status, and a normal data acquisition frequency is maintained and continuous monitoring is performed.

[0031] In a second aspect of the present invention, a multi-parameter calibration and evaluation system based on cross-modal parameters is proposed, including: an electrical operation coefficient module, a temperature characteristic distribution diagram module, an acoustic field distribution diagram module and an equipment state coefficient module:

[0032] The electrical operation coefficient module is used to obtain electrical data of target power equipment in the target area and calculate the electrical operation coefficient according to the electrical data; the electrical data includes: actual power and apparent power;

[0033] The temperature characteristic distribution diagram module is used to obtain temperature data of target power equipment in a target area from multiple perspectives, and obtain a temperature characteristic distribution diagram based on the fusion of the temperature data;

[0034] The sound field distribution diagram module is used to obtain a sound wave signal set of the target power equipment through a microphone array, and obtain a sound field distribution diagram by fitting the sound wave signal set;

[0035] The equipment state coefficient module is used to calculate the equipment state coefficient according to the electrical operation coefficient, the sound field distribution diagram and the temperature characteristic distribution diagram, and evaluate the operation state of the power equipment according to the equipment state coefficient.

[0036] Optionally, the temperature feature distribution diagram module includes: a feature extraction module, a data conversion module, a data reconstruction module, a feature fusion module and a temperature feature module:

[0037] The feature extraction module is used to obtain a visible light image and an infrared image of the target power equipment, extract spatial features in the visible light image, and identify pixel features of the infrared image;

[0038] The data conversion module is used to divide the pixel features according to the number of channels to obtain regional blocks, convert the regional blocks into frequency spectra through discrete cosine transformation, and encode the frequency spectrum to obtain a two-dimensional matrix;

[0039] The data reconstruction module is used to flatten the two-dimensional matrix to obtain a one-dimensional array, reshape the one-dimensional array to obtain a three-dimensional tensor, and reconstruct each three-dimensional tensor to obtain a target three-dimensional tensor;

[0040] The feature fusion module is used to downsample the target three-dimensional tensor to obtain a high-frequency signal and a low-frequency signal, and perform feature fusion on the high-frequency signal and the low-frequency signal to obtain a target frequency feature;

[0041] The temperature feature module is used to input the target frequency feature into the machine learning model to obtain a weight map, and add the spatial feature to the residual path to obtain a temperature feature distribution map.

[0042] Optionally, the sound field distribution diagram module includes: an array deployment module, a multi-ring array module, an array rotation module and a sound feature module:

[0043] The array deployment module is used to use the target microphone as a central node, and deploy the microphones at preset intervals with the central node as the center of the circle to obtain a target circular array; the circular array includes multiple nodes;

[0044] The multi-ring array module is used to use the central node as an endpoint and the nodes on the circular array as directions to determine the node positioning according to a preset interval to obtain a multi-ring array; the multi-ring array includes a plurality of circular arrays;

[0045] The array rotation module is used to determine a rotation reference point according to the target circular ring array, and rotate each circular ring array according to P=(i-1)θ according to the rotation reference point to obtain a calibration multi-ring array; P is a rotation angle, θ is a preset rotation angle, and i is the number of rings;

[0046] The sound feature module is used to collect sound according to the calibrated multi-ring array to obtain a sound data set, and to perform three-dimensional spatial deployment of the sound data set according to the spatial feature to obtain a sound field distribution map.

[0047] Optionally, the device state coefficient module includes: a data extraction module and a device state coefficient acquisition module:

[0048] The data extraction module is used to extract abnormal sound data sets and abnormal temperature data according to the sound field distribution map and the temperature characteristic distribution map; the abnormal sound data sets include abnormal sound data sets of each circular array;

[0049] The equipment state coefficient acquisition module is used to calculate the equipment state coefficient according to the electrical operation coefficient, the abnormal sound data set and the abnormal temperature data;

[0050] Equipment status coefficient calculation formula:

[0051]

[0052] Among them, F is the equipment status coefficient, e is the natural constant, D is the electrical operation coefficient, S i is the abnormal sound data set of the ith ring, n is the total number of rings, T j Abnormal temperature data in the temperature characteristic distribution diagram, m is the number of abnormal temperature areas, α, β i and γ are constant proportional coefficients and are not zero, β i The value decreases as the number of rings increases.

[0053] Optionally, the device state coefficient module further includes:

[0054] The first judgment module is used to determine that the state of the target power equipment is abnormal, increase the acquisition frequency of multi-parameter data and issue a fault alarm if the equipment state coefficient is greater than the state threshold;

[0055] The second judgment module is used to determine that the state of the target power equipment is normal if the equipment state coefficient is not greater than the state threshold, maintain a normal data acquisition frequency and perform continuous monitoring.

[0056] Beneficial effects of the present invention:

[0057] The present invention proposes a multi-parameter calibration and evaluation method based on cross-modal parameters, which obtains electrical data of target power equipment in a target area, and calculates an electrical operation coefficient according to the electrical data; obtains temperature data of target power equipment in the target area from multiple perspectives, and obtains a temperature characteristic distribution diagram according to temperature data fusion; obtains a sound wave signal set of the target power equipment through a microphone array, and obtains a sound field distribution diagram according to the sound wave signal set fitting; calculates an equipment state coefficient according to the electrical operation coefficient, the sound field distribution diagram and the temperature characteristic distribution diagram, and evaluates the operating state of the power equipment according to the equipment state coefficient; accurately obtains and analyzes the electrical operation, temperature distribution and sound field characteristic data of each power equipment in the target area to calculate the equipment state coefficient, thereby achieving accurate evaluation of the operating state of the power equipment, and effectively improving the accuracy and efficiency of fault warning and status monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The present invention will be further described below in conjunction with the accompanying drawings.

[0059] Figure 1 A flowchart of a multi-parameter calibration evaluation method based on cross-modal parameters is provided for an embodiment of the present invention;

[0060] Figure 2 A schematic diagram of another multi-parameter calibration evaluation method based on cross-modal parameters is provided for an embodiment of the present invention;

[0061] Figure 3 A structural schematic diagram of a multi-parameter calibration and evaluation system based on cross-modal parameters is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. The term "and / or" herein is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the present invention, the description of "first", "second", etc. is only used for descriptive purposes, and cannot be understood as indicating or implying its relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" can explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0063] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of the present invention.

[0064] The embodiment of the present invention provides a multi-parameter calibration evaluation method based on cross-modal parameters. Figure 1 , Figure 1 A flowchart of a multi-parameter calibration evaluation method based on cross-modal parameters provided by an embodiment of the present invention. The method comprises the following steps:

[0065] S101, obtaining electrical data of target power equipment in a target area, and calculating an electrical operation coefficient based on the electrical data;

[0066] S102, acquiring temperature data of target power equipment in a target area from multiple perspectives, and obtaining a temperature characteristic distribution map based on temperature data fusion;

[0067] S103, acquiring a sound wave signal set of the target power equipment through a microphone array, and obtaining a sound field distribution diagram according to the sound wave signal set;

[0068] S104, calculating the equipment state coefficient according to the electrical operation coefficient, the sound field distribution diagram and the temperature characteristic distribution diagram, and evaluating the operation state of the power equipment according to the equipment state coefficient;

[0069] Electrical data include: real power and apparent power;

[0070] A multi-parameter calibration and evaluation method based on cross-modal parameters is provided based on an embodiment of the present invention. By accurately acquiring and analyzing the electrical operation, temperature distribution and acoustic field characteristic data of each power equipment in the target area, the equipment state coefficient is calculated, thereby achieving an accurate evaluation of the operating state of the power equipment, and effectively improving the accuracy and efficiency of fault warning and status monitoring.

[0071] In one implementation, in the multi-parameter calibration evaluation of cross-modal parameters, cross-modality refers to a process or system that involves or integrates multiple different types of information modalities or sensory channels. The "modality" here usually refers to different forms of information flow, such as temperature, sound, and load of electrical equipment, etc. The multi-parameter calibration evaluation of cross-modal parameters means that during the evaluation process, it not only relies on data from a single modality, but also comprehensively processes and analyzes data from different modalities to reduce the limitations of a single data and provide a more comprehensive and accurate data source for the evaluation.

[0072] In one implementation, the power factor is an important parameter for measuring the power factor of a circuit or device. It indicates the ratio of actual power to apparent power. Because the main network equipment is AC, in AC circuits, due to the phase difference between current and voltage, the actual power is often less than the apparent power. The larger the power factor, the higher the power factor of the circuit or device, and the better the energy efficiency; actual power refers to the power actually consumed by the circuit or device, which is indirectly obtained by measuring the input power and output power of the device, and the input or output power of the circuit or device is directly measured using special instruments such as power meters or power analyzers; apparent power refers to the product of the effective value of voltage and the effective value of current in the circuit or device, without considering the phase difference between them. It indicates the maximum power that the circuit or device may provide or the maximum active power that it may consume; use instruments such as power meters or power quality analyzers to directly measure the apparent power of the circuit or device.

[0073] In one implementation, the temperature data of each power equipment in the target area is obtained from multiple perspectives. The temperature is measured by using infrared radiation from the object, which has the advantage of non-contact and can avoid direct contact with the object being measured. Infrared thermometers and infrared thermal imagers are two common infrared temperature measurement tools. Infrared thermal imagers can obtain temperature distribution images of target objects and are suitable for measuring the temperature of equipment with large areas or complex shapes. The data is transmitted to the background host for display, storage and analysis. This method has the advantages of easy installation, strong real-time performance and high measurement accuracy. It is suitable for temperature monitoring of electrical equipment in the main network equipment. The target area refers to the area where the main network equipment is located, such as transformers, generators, etc.

[0074] In one implementation, the process of comprehensively evaluating the operating status of power equipment involves combining multiple parameters such as the electrical operating coefficient, the acoustic field distribution diagram, and the temperature characteristic distribution diagram. By comprehensively monitoring and analyzing the changes in these parameters, potential failures and problems of the equipment can be discovered in a timely manner, thereby improving the accuracy and reliability of the evaluation results.

[0075] In one embodiment, step S102 includes:

[0076] Obtain visible light images and infrared images of target power equipment, extract spatial features in visible light images, and identify pixel features in infrared images;

[0077] The pixel features are divided into regional blocks according to the number of channels, the regional blocks are converted into frequency spectra through discrete cosine transformation, and the frequency spectrum is encoded to obtain a two-dimensional matrix;

[0078] Flatten the two-dimensional matrix to obtain a one-dimensional array, reshape the one-dimensional array to obtain a three-dimensional tensor, and reconstruct each three-dimensional tensor to obtain the target three-dimensional tensor;

[0079] Down-sample the target three-dimensional tensor to obtain high-frequency signals and low-frequency signals, and fuse the high-frequency signals and low-frequency signals to obtain the target frequency features;

[0080] The target frequency features are input into the machine learning model to obtain the weight map, and the spatial features are added to the residual path to obtain the temperature feature distribution map.

[0081] In one implementation, in visible light images, spatial features refer to spatial information such as the shape, size, and position of objects in the image; pixel features in infrared images usually refer to pixel values, which reflect the temperature information on the surface of the object; discrete cosine transform (DCT) is a mathematical transformation method used to convert signals from the spatial domain to the frequency domain, and is often used in image compression and signal processing; the frequency spectrum is the result of DCT conversion, which represents the component intensity of the signal at different frequencies; downsampling is used to decompose the target three-dimensional tensor into high-frequency and low-frequency signals; weight maps in machine learning models, weight maps usually indicate the degree of importance or influence of the model on each input feature; residual path: in deep learning, the residual path is a structure used to retain input information and help the network better learn deep features, and is used to add spatial features to the final temperature feature distribution map.

[0082] In one implementation, the machine learning model is used to extract features from different modal data, build a correlation model between modalities, capture the spatiotemporal characteristics and internal connections of each state parameter, and automatically identify and model the complex relationship between multimodal data, thereby enhancing the understanding and analysis of the device status. On this basis, the study designs a collaborative calibration algorithm based on machine learning technology, and uses the extracted correlation model to achieve automatic calibration for the deviation of each modal data, thereby improving the consistency and accuracy of the data.

[0083] In one implementation, flattening refers to the process of converting multidimensional data (such as a two-dimensional matrix) into a one-dimensional array. The two-dimensional matrix is ​​obtained by encoding the frequency spectrum, and the flattening operation converts it into a one-dimensional array. The flattening operation helps to simplify the data structure for subsequent processing; reshaping refers to the process of reorganizing a one-dimensional array into multidimensional data (for example, a three-dimensional tensor). After flattening the two-dimensional matrix into a one-dimensional array, it needs to be reshaped into a three-dimensional tensor for subsequent feature fusion and machine learning processing. The reshaping operation retains the information of the original data, but changes its structure to meet different processing requirements. Reconstruction refers to the process of combining or adjusting multiple three-dimensional tensors to generate a target three-dimensional tensor. By integrating multiple source data to form new data with specific structures or attributes, the reconstruction operation may involve weighted averaging, splicing, or other forms of combination of multiple three-dimensional tensors to generate the final target three-dimensional tensor. The target three-dimensional tensor will be used in subsequent downsampling and feature fusion steps.

[0084] In one implementation, the spatial features of the visible light image of the power equipment are combined with the pixel features of the infrared image, and the high-frequency and low-frequency features of the temperature distribution of the power equipment are extracted from the infrared image by using methods such as discrete cosine transform, frequency spectrum coding, three-dimensional tensor reshaping and reconstruction, and feature fusion. Then, a weight map is generated through a machine learning model and fused with the spatial features of the visible light image in the residual path. Finally, an accurate temperature feature distribution map is obtained, which can significantly improve the accuracy and efficiency of thermal fault detection and temperature monitoring of power equipment.

[0085] In one embodiment, step S103 includes:

[0086] The target microphone is taken as the central node, and the microphones are deployed at preset intervals with the central node as the center to obtain a target circular array; the circular array includes multiple nodes;

[0087] The central node is taken as the endpoint, and the nodes on the circular array are taken as the direction to determine the node positioning according to the preset interval to obtain a multi-ring array; the multi-ring array includes multiple circular arrays;

[0088] Determine the rotation reference point according to the target circular ring array, and rotate each circular ring array according to P=(i-1)θ according to the rotation reference point to obtain a calibration multi-ring array; P is the rotation angle, θ is the preset rotation angle, and i is the number of rings;

[0089] A sound dataset is obtained by collecting sound according to the calibrated multi-ring array, and a sound field distribution map is obtained by deploying the sound dataset in three-dimensional space according to spatial characteristics.

[0090] In one implementation, a multi-ring microphone array is constructed with the target power equipment as the center, and the array is calibrated using preset intervals and rotation angles to achieve accurate acquisition of sound features. Combined with spatial features, this technology can construct an accurate three-dimensional sound field distribution map, significantly improving the accuracy of sound feature acquisition and providing powerful spatial features for subsequent applications such as sound recognition, positioning, and environmental monitoring.

[0091] In one implementation, see Figure 2 , Figure 2A schematic diagram of a multi-parameter calibration and evaluation method based on cross-modal parameters provided by an embodiment of the present invention; wherein the hollow circles are midline nodes, and the solid circles are nodes; according to the rotation reference point, each circular array is rotated according to P=(i-1)θ to obtain a calibrated multi-ring array, that is, the circular array closest to the central node is used as the rotation reference point, and the subsequent circular array is positioned with the rotation reference point and rotated according to the rotation angle determined by the number of rings, θ is a preset rotation angle, for example: θ is 5°; the circular array is obtained by a circle fitted by a microphone, so there are multiple nodes (microphones) on the circular array, with the central node as the endpoint, and the node is used as the direction to emit outward in the form of a ray (two points and one line), and the node positioning on the ray is determined according to a preset interval, and the preset interval is related to the distance of the central node (that is, the actual distance between the microphone and the power equipment, for example: 10 cm), and the multi-ring array is obtained by fitting nodes with the same interval.

[0092] In one embodiment, step S104 includes:

[0093] According to the sound field distribution map and the temperature characteristic distribution map, an abnormal sound data set and abnormal temperature data are extracted; the abnormal sound data set includes abnormal sound data sets of each circular array;

[0094] The equipment status coefficient is calculated based on the electrical operation coefficient, the abnormal sound data set and the abnormal temperature data;

[0095] Equipment status coefficient calculation formula:

[0096]

[0097] Among them, F is the equipment status coefficient, e is the natural constant, D is the electrical operation coefficient, S i is the abnormal sound data set of the ith ring, n is the total number of rings, T j Abnormal temperature data in the temperature characteristic distribution diagram, m is the number of abnormal temperature areas, α, β i and γ are constant proportional coefficients and are not zero, β i The value decreases as the number of rings increases.

[0098] In one implementation, abnormal temperature data and abnormal sound data sets of the circular array are extracted according to the sound field distribution map and the temperature characteristic distribution map, that is, the abnormal sound data sets of each circular array are extracted from the sound field distribution map, and the data in each abnormal sound data set is composed of the data provided by each node, β i is the constant proportional coefficient of the abnormal sound data set, and the β corresponding to each abnormal sound data set (circular array) i All different, β i The value of β decreases as the number of rings increases, that is, i(the value of the constant proportional coefficient) is directly proportional to i (the number of rings). When i increases, β i Also increases accordingly, (for example: when the number of rings i is 1, β i is 0.40; the number of rings i is 2, β i is 0.30; the number of rings i is 3, β i is 0.20, the number of rings i is 4, β i =0.05) where i reflects the distance between the circular array and the central node, that is, the distance between the microphone array and the power equipment. The farther the distance, the greater the sound attenuation and the more interference will be received. The weight β i It will also be reduced; the abnormal sound data set includes the abnormal sound data sets of each circular array, and each circular array corresponds to an abnormal sound data set. When the sound data collected by the nodes in the circular array is abnormal, that is, sound data = sound intensity × duration, the sound data is greater than the sound threshold, then the sound data is determined to be abnormal sound data.

[0099] In one implementation, in the temperature characteristic distribution diagram, there will be multiple temperature areas. The area with abnormal temperature, that is, the area with the highest central temperature, is identified. If there is an obvious temperature gradient between the area with the highest temperature and the surrounding temperature, the area is regarded as one temperature area; if there are two areas with temperatures of 55°C and 65°C respectively, there is an obvious temperature gradient between the two areas, and the temperatures of the two areas are both greater than the temperature threshold, then the above two temperature areas are determined to be temperature abnormal areas; similarly, if the temperatures of two temperature areas are the same, and the temperatures of the two areas are both greater than the temperature threshold, then the above two temperature areas are determined to be temperature abnormal areas.

[0100] In one embodiment, step S104 further includes:

[0101] If the equipment status coefficient is greater than the status threshold, the target power equipment is judged to be in an abnormal state, the acquisition frequency of multi-parameter data is increased, and a fault alarm is issued;

[0102] If the equipment status coefficient is not greater than the status threshold, the target power equipment is judged to be in normal status, and the normal data acquisition frequency is maintained and continuous monitoring is performed.

[0103] In one implementation, when the equipment status coefficient exceeds a preset status threshold, the system can quickly determine that the equipment status is abnormal and immediately increase the frequency of data acquisition. Immediate response helps to timely discover and handle potential faults, prevent further deterioration of faults, and avoid equipment damage or downtime. When the equipment status is normal, the system maintains a normal data acquisition frequency, which neither wastes resources by collecting data too frequently nor misses important information due to insufficient data. Balancing the use of resources helps reduce monitoring costs and improve overall operational efficiency. Timely discovery and handling of equipment failures can reduce the risk of safety accidents caused by equipment failures.

[0104] Based on the same inventive concept, the embodiment of the present invention also provides a multi-parameter calibration evaluation system based on cross-modal parameters. Figure 3 , Figure 3 A schematic diagram of the structure of a multi-parameter calibration and evaluation system based on cross-modal parameters provided by an embodiment of the present invention includes: an electrical operation coefficient module, a temperature characteristic distribution diagram module, an acoustic field distribution diagram module and an equipment state coefficient module:

[0105] The electrical operation coefficient module is used to obtain electrical data of target power equipment in the target area and calculate the electrical operation coefficient based on the electrical data; the electrical data includes: actual power and apparent power;

[0106] The temperature characteristic distribution map module is used to obtain the temperature data of the target power equipment in the target area from multiple perspectives, and obtain the temperature characteristic distribution map based on the temperature data fusion;

[0107] The sound field distribution diagram module is used to obtain the sound wave signal set of the target power equipment through the microphone array, and obtain the sound field distribution diagram according to the sound wave signal set;

[0108] The equipment state coefficient module is used to calculate the equipment state coefficient according to the electrical operation coefficient, the sound field distribution diagram and the temperature characteristic distribution diagram, and evaluate the operation state of the power equipment according to the equipment state coefficient.

[0109] A multi-parameter calibration and evaluation system based on cross-modal parameters is provided according to an embodiment of the present invention. By accurately acquiring and analyzing the electrical operation, temperature distribution and acoustic field characteristic data of each power equipment in the target area, the equipment state coefficient is calculated, thereby achieving an accurate evaluation of the operating state of the power equipment, and effectively improving the accuracy and efficiency of fault warning and status monitoring.

[0110] In one embodiment, the temperature feature distribution diagram module includes: a feature extraction module, a data conversion module, a data reconstruction module, a feature fusion module and a temperature feature module:

[0111] A feature extraction module is used to obtain visible light images and infrared images of target power equipment, extract spatial features in the visible light image, and identify pixel features in the infrared image;

[0112] A data conversion module is used to divide pixel features into regional blocks according to the number of channels, convert the regional blocks into frequency spectra through discrete cosine transformation, and encode the frequency spectrum to obtain a two-dimensional matrix;

[0113] A data reconstruction module is used to flatten a two-dimensional matrix to obtain a one-dimensional array, reshape the one-dimensional array to obtain a three-dimensional tensor, and reconstruct each three-dimensional tensor to obtain a target three-dimensional tensor;

[0114] A feature fusion module is used to downsample the target three-dimensional tensor to obtain high-frequency signals and low-frequency signals, and perform feature fusion on the high-frequency signals and the low-frequency signals to obtain target frequency features;

[0115] The temperature feature module is used to input the target frequency feature into the machine learning model to obtain a weight map, and add the spatial feature to the residual path to obtain a temperature feature distribution map.

[0116] In one embodiment, the sound field distribution diagram module includes: an array deployment module, a multi-ring array module, an array rotation module and a sound feature module:

[0117] An array deployment module is used to use the target microphone as a central node and deploy the microphones at preset intervals with the central node as the center of the circle to obtain a target circular array; the circular array includes multiple nodes;

[0118] A multi-ring array module is used to use the central node as an endpoint and the nodes on the circular array as directions to determine the node positioning according to a preset interval to obtain a multi-ring array; the multi-ring array includes multiple circular arrays;

[0119] An array rotation module is used to determine a rotation reference point according to a target circular ring array, and to rotate each circular ring array according to the rotation reference point according to P = (i-1)θ to obtain a calibration multi-ring array; P is a rotation angle, θ is a preset rotation angle, and i is the number of rings;

[0120] The sound feature module is used to collect sound according to the calibrated multi-ring array to obtain a sound data set, and to deploy the sound data set in three-dimensional space according to the spatial features to obtain a sound field distribution map.

[0121] In one embodiment, the device state coefficient module includes: a data extraction module and a device state coefficient acquisition module:

[0122] A data extraction module is used to extract an abnormal sound data set and abnormal temperature data according to the sound field distribution map and the temperature characteristic distribution map; the abnormal sound data set includes the abnormal sound data set of each circular array;

[0123] An equipment status coefficient acquisition module is used to calculate the equipment status coefficient according to the electrical operation coefficient, the abnormal sound data set and the abnormal temperature data;

[0124] Equipment status coefficient calculation formula:

[0125]

[0126] Among them, F is the equipment status coefficient, e is the natural constant, D is the electrical operation coefficient, S i is the abnormal sound data set of the i-th ring, i is the number of rings, T j Abnormal temperature data in the temperature characteristic distribution diagram, m is the number of abnormal temperature areas, α, β i and γ are constant proportional coefficients and are not zero, β i The value decreases as the number of rings increases.

[0127] In one embodiment, the device state coefficient module further includes:

[0128] The first judgment module is used to determine that the state of the target power equipment is abnormal if the equipment state coefficient is greater than the state threshold, increase the acquisition frequency of multi-parameter data and issue a fault alarm;

[0129] The second judgment module is used to determine that the state of the target power equipment is normal if the equipment state coefficient is not greater than the state threshold, maintain a normal data acquisition frequency and perform continuous monitoring.

[0130] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A multi-parameter calibration evaluation method based on cross-modal parameters, characterized in that: The method comprises: Acquire electrical data of target power equipment in the target area, and calculate the electrical operation coefficient according to the electrical data; the electrical data includes: actual power and apparent power; Acquire temperature data of target power equipment in a target area from multiple perspectives, and obtain a temperature characteristic distribution map based on the fusion of the temperature data; Acquire a sound wave signal set of the target power equipment through a microphone array, and obtain a sound field distribution diagram according to the sound wave signal set; The equipment state coefficient is calculated according to the electrical operation coefficient, the sound field distribution diagram and the temperature characteristic distribution diagram, and the operation state of the electric equipment is evaluated according to the equipment state coefficient.

2. The multi-parameter calibration evaluation method based on cross-modal parameters according to claim 1, characterized in that: The temperature characteristic distribution diagram obtained by fusing the temperature data includes: Acquire a visible light image and an infrared image of the target power equipment, extract spatial features in the visible light image, and identify pixel features in the infrared image; Dividing the pixel features into regional blocks according to the number of channels, converting the regional blocks into frequency spectra by discrete cosine transform, and encoding the frequency spectra to obtain a two-dimensional matrix; Flatten the two-dimensional matrix to obtain a one-dimensional array, reshape the one-dimensional array to obtain a three-dimensional tensor, and reconstruct each three-dimensional tensor to obtain the target three-dimensional tensor; Downsampling the target three-dimensional tensor to obtain a high-frequency signal and a low-frequency signal, and fusing the high-frequency signal and the low-frequency signal to obtain a target frequency feature; The target frequency feature is input into the machine learning model to obtain a weight map, and the spatial feature is added to the residual path to obtain a temperature feature distribution map.

3. The multi-parameter calibration evaluation method based on cross-modal parameters according to claim 2, characterized in that: The steps of deploying the microphone array include: The target microphone is taken as the central node, and the microphones are deployed at preset intervals with the central node as the center to obtain a target circular array; the circular array includes multiple nodes; The central node is used as an endpoint, and the nodes on the circular array are used as directions to determine the node positioning according to a preset interval to obtain a multi-ring array; the multi-ring array includes a plurality of circular arrays; Determine a rotation reference point according to the target circular ring array, and rotate each circular ring array according to P=(i-1)θ according to the rotation reference point to obtain a calibration multi-ring array; P is the rotation angle, θ is the preset rotation angle, and i is the number of rings; Sound is collected according to the calibrated multi-ring array to obtain a sound data set, and the sound data set is deployed in three-dimensional space according to the spatial characteristics to obtain a sound field distribution map.

4. The multi-parameter calibration evaluation method based on cross-modal parameters according to claim 3 is characterized in that: The equipment state coefficient calculated according to the electrical operation coefficient, the sound field distribution diagram and the temperature characteristic distribution diagram includes: Extracting an abnormal sound data set and abnormal temperature data according to the sound field distribution map and the temperature characteristic distribution map; the abnormal sound data set includes the abnormal sound data sets of each circular array; Calculate the equipment status coefficient according to the electrical operation coefficient, the abnormal sound data set and the abnormal temperature data; Equipment status coefficient calculation formula: Among them, F is the equipment status coefficient, e is the natural constant, D is the electrical operation coefficient, S i is the abnormal sound data set of the ith ring, n is the total number of rings, T j Abnormal temperature data in the temperature characteristic distribution diagram, m is the number of abnormal temperature areas, α, β i and γ are constant proportional coefficients and are not zero, β i The value decreases as the number of rings increases.

5. The multi-parameter calibration evaluation method based on cross-modal parameters according to claim 1, characterized in that: Evaluating the operating status of the power equipment according to the equipment status coefficient includes: If the equipment status coefficient is greater than the status threshold, the target power equipment is judged to be in an abnormal state, the acquisition frequency of multi-parameter data is increased, and a fault alarm is issued; If the device status coefficient is not greater than the status threshold, it is determined that the target power device is in normal status, and a normal data acquisition frequency is maintained and continuous monitoring is performed.

6. A multi-parameter calibration evaluation system based on cross-modal parameters, characterized in that: The system includes: an electrical operation coefficient module, a temperature characteristic distribution diagram module, an acoustic field distribution diagram module and an equipment status coefficient module: The electrical operation coefficient module is used to obtain electrical data of target power equipment in the target area and calculate the electrical operation coefficient according to the electrical data; the electrical data includes: actual power and apparent power; The temperature characteristic distribution diagram module is used to obtain temperature data of target power equipment in a target area from multiple perspectives, and obtain a temperature characteristic distribution diagram based on the fusion of the temperature data; The sound field distribution diagram module is used to obtain a sound wave signal set of the target power equipment through a microphone array, and obtain a sound field distribution diagram by fitting the sound wave signal set; The equipment state coefficient module is used to calculate the equipment state coefficient according to the electrical operation coefficient, the sound field distribution diagram and the temperature characteristic distribution diagram, and evaluate the operation state of the power equipment according to the equipment state coefficient.

7. A multi-parameter calibration and evaluation system based on cross-modal parameters according to claim 6, characterized in that: The temperature characteristic distribution diagram module includes: a feature extraction module, a data conversion module, a data reconstruction module, a feature fusion module and a temperature characteristic module: The feature extraction module is used to obtain a visible light image and an infrared image of the target power equipment, extract spatial features in the visible light image, and identify pixel features of the infrared image; The data conversion module is used to divide the pixel features according to the number of channels to obtain regional blocks, convert the regional blocks into frequency spectra through discrete cosine transformation, and encode the frequency spectrum to obtain a two-dimensional matrix; The data reconstruction module is used to flatten the two-dimensional matrix to obtain a one-dimensional array, reshape the one-dimensional array to obtain a three-dimensional tensor, and reconstruct each three-dimensional tensor to obtain a target three-dimensional tensor; The feature fusion module is used to downsample the target three-dimensional tensor to obtain a high-frequency signal and a low-frequency signal, and perform feature fusion on the high-frequency signal and the low-frequency signal to obtain a target frequency feature; The temperature feature module is used to input the target frequency feature into the machine learning model to obtain a weight map, and add the spatial feature to the residual path to obtain a temperature feature distribution map.

8. The multi-parameter calibration and evaluation system based on cross-modal parameters according to claim 6, characterized in that: The sound field distribution diagram module includes: an array deployment module, a multi-ring array module, an array rotation module and a sound feature module: The array deployment module is used to use the target microphone as a central node, and deploy the microphones at preset intervals with the central node as the center of the circle to obtain a target circular array; the circular array includes multiple nodes; The multi-ring array module is used to use the central node as an endpoint and the nodes on the circular array as directions to determine the node positioning according to a preset interval to obtain a multi-ring array; the multi-ring array includes a plurality of circular arrays; The array rotation module is used to determine a rotation reference point according to the target circular ring array, and rotate each circular ring array according to P=(i-1)θ according to the rotation reference point to obtain a calibration multi-ring array; P is a rotation angle, θ is a preset rotation angle, and i is the number of rings; The sound feature module is used to collect sound according to the calibrated multi-ring array to obtain a sound data set, and to perform three-dimensional spatial deployment of the sound data set according to the spatial feature to obtain a sound field distribution map.

9. The multi-parameter calibration and evaluation system based on cross-modal parameters according to claim 6, characterized in that: The device state coefficient module includes: a data extraction module and a device state coefficient acquisition module: The data extraction module is used to extract abnormal sound data sets and abnormal temperature data according to the sound field distribution map and the temperature characteristic distribution map; the abnormal sound data sets include abnormal sound data sets of each circular array; The equipment state coefficient acquisition module is used to calculate the equipment state coefficient according to the electrical operation coefficient, the abnormal sound data set and the abnormal temperature data; Equipment status coefficient calculation formula: Among them, F is the equipment status coefficient, e is the natural constant, D is the electrical operation coefficient, S i is the abnormal sound data set of the ith ring, n is the total number of rings, T j Abnormal temperature data in the temperature characteristic distribution diagram, m is the number of abnormal temperature areas, α, β i and γ are constant proportional coefficients and are not zero, β i The value decreases as the number of rings increases.

10. The multi-parameter calibration and evaluation system based on cross-modal parameters according to claim 6, characterized in that: The equipment status coefficient module also includes: The first judgment module is used to determine that the state of the target power equipment is abnormal, increase the acquisition frequency of multi-parameter data and issue a fault alarm if the equipment state coefficient is greater than the state threshold; The second judgment module is used to determine that the state of the target power equipment is normal if the equipment state coefficient is not greater than the state threshold, maintain a normal data acquisition frequency and perform continuous monitoring.

Images