Enterprise equipment efficiency dynamic prediction method and device, electronic equipment and storage medium

Abstract

The invention relates to the technical field of computer application, and particularly discloses an enterprise equipment efficiency dynamic prediction method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining original data of different types of equipment, and extracting an original feature vector of each piece of equipment; calculating a weight coefficient of each device based on the historical contribution degree, the current health state and the real-time load of each device, and performing weighted aggregation on the original feature vector to obtain a dynamic weighted feature vector of a device level; aggregating to form a production line layer feature vector; inputting the production line layer feature vector into a pre-constructed equipment efficiency prediction model to generate an equipment efficiency prediction result; the processing process of the method relates to hierarchical aggregation and dynamic weighting, multi-source heterogeneous equipment data can be effectively utilized, and dynamic prediction of the equipment efficiency is realized on the edge side with limited resources, so that dynamic and accurate prediction of the equipment efficiency can be realized, and technical support is provided for intelligent upgrading of factories.

Description

Technical Field

[0001] The present application relates to the field of computer application technologies, and more particularly, to a method, device, electronic device, and storage medium for dynamically predicting the efficiency of enterprise equipment. Background Art

[0002] Traditional methods for predicting the efficiency of enterprise equipment have long dominated the industrial field, and their core relies on centralized servers for large-scale data processing and complex model calculations.

[0003] However, existing equipment efficiency prediction methods largely ignore the dynamic change characteristics of equipment operating states. For example, on complex production lines, the tasks and importance levels borne by different equipment vary, and the real-time load of the equipment itself also fluctuates with changes in production tasks. However, traditional prediction methods often fail to fully consider these key factors, resulting in a deviation between the prediction model and the actual operating state of the equipment, and it is difficult for the prediction results to accurately reflect the true situation of equipment efficiency, and both the prediction accuracy and real-time performance are difficult to meet the requirements of practical applications.

[0004] In response to the above problems, there is currently no effective technical solution. Summary of the Invention

[0005] The purpose of the present application is to provide a method, device, electronic device, and storage medium for dynamically predicting the efficiency of enterprise equipment, so as to achieve dynamic and accurate prediction of equipment efficiency on the premise of paying attention to the dynamic change characteristics of equipment operating states.

[0006] In a first aspect, the present application provides a method for dynamically predicting the efficiency of enterprise equipment for predicting equipment efficiency, and the method includes the following steps: S1. Obtain the original data of different types of equipment, perform data cleaning and format unification processing on the original data, and extract the original feature vectors of each equipment, where the original feature vectors represent the key feature parameters of the corresponding equipment operating state; S2. Calculate the weight coefficients of each equipment based on the historical contribution degree, current health state, and real-time load of each equipment, and perform weighted aggregation on the original feature vectors of the equipment at the same level on the production line according to the weight coefficients to obtain the dynamic weighted feature vectors at the equipment level; S3. Aggregate the dynamic weighted feature vectors of the equipment on the same production line according to the production line process flow to form the feature vectors at the production line level; S4. Input the feature vectors at the production line level into a pre-constructed equipment efficiency prediction model to generate an equipment efficiency prediction result.

[0007] The enterprise equipment efficiency dynamic prediction method of the present application realizes the dynamic configuration of the weight coefficient by comprehensively considering the historical contribution, current health status and real-time load of the equipment, aggregates the dynamic weighted feature vectors of the equipment on the same production line according to the production line process flow to form a production line layer feature vector, thereby abstracting the overall operation state of the production line at a high level, and then inputs the production line layer feature vector into a pre-constructed equipment efficiency prediction model to generate an equipment efficiency prediction result, and finally realizes the prediction output of the equipment efficiency. Its overall processing process involves hierarchical aggregation and dynamic weighting, can effectively utilize multi-source heterogeneous equipment data, and thus can realize the dynamic and accurate prediction of the equipment efficiency, providing technical support for the intelligent upgrade of the factory.

[0008] For the described enterprise equipment efficiency dynamic prediction method, wherein, the historical contribution degree is determined based on the following steps: A1. Obtain the operation data of each device within a preset historical period, and the operation data includes production data, energy consumption data and operation time data; A2. Calculate the unit time workload corresponding to each device according to the production data of each device; A3. Calculate the unit time energy consumption intensity corresponding to each device according to the energy consumption data and operation time data of each device; A4. Calculate the historical contribution degree corresponding to each device according to the unit time workload and the unit time energy consumption intensity.

[0009] The above processing method clarifies the acquisition path of the historical contribution degree of each device. By calculating the unit time workload and the unit time energy consumption intensity, the historical operation data of each device is quantified and standardized. The historical contribution degree obtained by combining the unit time workload and the unit time energy consumption intensity reflects the level of the energy consumption intensity of the device while generating a unit workload. By clarifying the calculation method of the historical contribution degree, the historical performance of the device can be evaluated more objectively, and then a more accurate basis can be provided for the subsequent calculation of the equipment weight coefficient, and finally the accuracy of the equipment efficiency prediction can be improved.

[0010] For the described enterprise equipment efficiency dynamic prediction method, wherein, the step of calculating the weight coefficient of each device based on the historical contribution degree, current health status and real-time load of each device includes: S21. Obtain the depreciation life data and actual working hours data of each device, and calculate the current health status of each device according to the depreciation life data and the actual working hours data; S22. Calculate the real-time load of each device according to the actual operating power and rated power of each device; S23. Calculate the dynamic adjustment coefficient of each device according to the current health status and the real-time load; S24. Calculate the weight coefficients of each device by combining the dynamic adjustment coefficient and the historical contribution degree.

[0011] The above processing process combines the dynamic adjustment coefficient with the historical contribution degree. The finally obtained weight coefficient not only considers the historical efficiency of the device but also takes into account the current operating state of the device, enabling the device weight coefficient to more accurately and dynamically reflect the device efficiency. Therefore, using this weight coefficient for device efficiency prediction can improve the accuracy and reliability of the prediction and solve the problem of low prediction accuracy caused by the unclear quantification method of the weight coefficient.

[0012] The described dynamic prediction method for enterprise device efficiency, wherein step S2 further includes the following steps: S25. Obtain the operating parameters of each device, where the operating parameters include actual working voltage information, actual working current information, and actual ambient temperature information; S26. Calculate the parameter adjustment coefficient of each device according to the actual working voltage information, actual working current information, and actual ambient temperature information; S27. Compensate and adjust the weight coefficient of each device based on the parameter adjustment coefficient.

[0013] The described dynamic prediction method for enterprise device efficiency, wherein step S1 includes: S11. Obtain the original data of different types of devices, evaluate the quality of the original data of each device based on the data missing rate. If the data missing rate of the original data is lower than the preset threshold, it is determined as high-quality data; otherwise, it is determined as low-quality data; S12. For the high-quality data, use an outlier detection method based on statistical principles for data cleaning and perform format standardization processing to obtain the cleaned high-quality data; S13. For the low-quality data, use a missing value filling method based on data mining for data cleaning and perform format standardization processing to obtain the cleaned low-quality data; S14. Extract the time-domain statistical features, frequency-domain statistical features, and wavelet transform features of the cleaned high-quality data and the cleaned low-quality data respectively to form the original feature vector of each device.

[0014] The described dynamic prediction method for enterprise device efficiency, wherein step S3 includes: S31. Obtain the device relationship graph, where the device relationship graph uses nodes to represent devices, edges to represent the technological process relationships between devices, and the weight of the edges represents the material transfer amount between devices; S32. Search and obtain the production line chain based on the device relationship graph and the magnitude relationship of the edge weights; S33. Extract the dynamic weighted feature vectors of the corresponding devices according to the production line chain, and integrate them using the weights of the edges to generate the production line layer feature vectors.

[0015] In the described dynamic prediction method for enterprise equipment efficiency, the equipment efficiency prediction model includes a teacher model and a student model compressed and simplified based on the teacher model. Step S4 includes: S41. Monitor the resource utilization rate of edge devices, where the resource utilization rate includes CPU utilization rate, memory utilization rate, and storage utilization rate; S42. If any resource utilization rate exceeds the preset threshold, input the production line layer feature vectors into the student model to generate the equipment efficiency prediction result; if all resource utilization rates do not exceed the preset threshold, input the production line layer feature vectors into the teacher model to generate the equipment efficiency prediction result.

[0016] In a second aspect, the present application also provides a dynamic prediction device for enterprise equipment efficiency for predicting equipment efficiency, including: An acquisition module for acquiring the original data of different types of devices, performing data cleaning and format standardization processing on the original data, and extracting the original feature vectors of each device, where the original feature vectors represent the key feature parameters of the operating state of the corresponding device; A first integration module for calculating the weight coefficients of each device based on the historical contribution degree, current health status, and real-time load of each device, and performing weighted aggregation on the original feature vectors of the devices at the same level on the production line according to the weight coefficients to obtain the dynamic weighted feature vectors at the device level; A second integration module for aggregating the dynamic weighted feature vectors of the devices on the same production line according to the production line process flow to form the production line layer feature vectors; A prediction module for inputting the production line layer feature vectors into a pre-constructed equipment efficiency prediction model to generate the equipment efficiency prediction result.

[0017] The enterprise equipment efficiency dynamic prediction device of the present application realizes the dynamic configuration of the weight coefficient by comprehensively considering the historical contribution of the equipment, the current health status and the real-time load, so as to highlight the influence of important equipment and equipment with abnormal status on the production line efficiency, reduce the influence of unimportant equipment and equipment with good status, and aggregate the dynamic weighted feature vectors of the equipment on the same production line according to the production line process flow to form a production line layer feature vector, thereby abstracting the overall operation status of the production line at a high level. Then, the production line layer feature vector is input into a pre-constructed equipment efficiency prediction model to generate an equipment efficiency prediction result, and finally the prediction output of the equipment efficiency is realized. Its overall processing process involves hierarchical aggregation and dynamic weighting, which can effectively utilize multi-source heterogeneous equipment data, and thus can realize the dynamic and accurate prediction of equipment efficiency, providing technical support for the intelligent upgrade of the factory.

[0018] In a third aspect, the present application further provides an electronic device, including a processor and a memory. The memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps in the method provided in the first aspect above are run.

[0019] In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps in the method provided in the first aspect above are run.

[0020] As can be seen from the above, the present application provides an enterprise equipment efficiency dynamic prediction method, device, electronic device and storage medium. Among them, compared with the traditional method that relies on a centralized server for data processing and model calculation, the enterprise equipment efficiency dynamic prediction method realizes the dynamic configuration of the weight coefficient by comprehensively considering the historical contribution of the equipment, the current health status and the real-time load, so as to highlight the influence of important equipment and equipment with abnormal status on the production line efficiency, reduce the influence of unimportant equipment and equipment with good status, and aggregate the dynamic weighted feature vectors of the equipment on the same production line according to the production line process flow to form a production line layer feature vector, thereby abstracting the overall operation status of the production line at a high level. Then, the production line layer feature vector is input into a pre-constructed equipment efficiency prediction model to generate an equipment efficiency prediction result, and finally the prediction output of the equipment efficiency is realized. Its overall processing process involves hierarchical aggregation and dynamic weighting, which can effectively utilize multi-source heterogeneous equipment data, and realize the dynamic prediction of equipment efficiency on the resource-constrained edge side, so as to be able to realize the dynamic and accurate prediction of equipment efficiency, providing technical support for the intelligent upgrade of the factory. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a flowchart of the enterprise equipment efficiency dynamic prediction method provided by an embodiment of the present application.

[0022] Figure 2 This is a schematic structural diagram of the enterprise equipment efficiency dynamic prediction device provided by the embodiment of the present application.

[0023] Figure 3 This is a schematic structural diagram of the electronic device provided by the embodiment of the present application.

[0024] Reference numerals: 201, acquisition module; 202, first integration module; 203, second integration module; 204, prediction module; 301, processor; 302, memory; 303, communication bus. Detailed implementation manners

[0025] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Usually, the components of the embodiments of the present application described and shown in the accompanying drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application to be protected, but only represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present application.

[0026] It should be noted that similar reference numerals and letters indicate similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of the present application, the terms "first", "second", etc. are only used for distinguishing descriptions, and cannot be understood as indicating or implying relative importance.

[0027] In a first aspect, please refer to Figure 1 , some embodiments of the present application provide an enterprise equipment efficiency dynamic prediction method for predicting equipment efficiency. The method includes the following steps: S1. Obtain the original data of different types of equipment, perform data cleaning and format standardization processing on the original data, and extract the original feature vectors of each equipment. The original feature vectors represent the key feature parameters of the corresponding equipment operation state; S2. Calculate the weight coefficients of each equipment based on the historical contribution degree, current health status and real-time load of each equipment, and perform weighted aggregation on the original feature vectors of the equipment at the same level on the production line according to the weight coefficients to obtain the dynamic weighted feature vectors at the equipment level; S3. According to the production line process flow, aggregate the dynamic weighted feature vectors of the equipment on the same production line to form the production line layer feature vectors; S4. Input the production line layer feature vectors into a pre-constructed equipment efficiency prediction model to generate equipment efficiency prediction results.

[0028] Specifically, in step S1, the original data comes from different types of equipment in a large factory workshop. The data cleaning process may include operations such as removing duplicate values, handling missing values and outliers, etc., to ensure data quality. The format unification process can convert data in different formats into a unified format, such as timestamp format, numerical format, etc., for convenient subsequent processing. The extraction of the original feature vectors can be based on time domain, frequency domain or time-frequency domain analysis methods. For example, feature parameters such as mean, variance, and spectral energy are extracted. These original feature vectors represent the key feature parameters of the corresponding equipment operation state and can lay a data foundation for subsequent efficiency prediction.

[0029] More specifically, step S2 is used to calculate the weight coefficients of each equipment. It comprehensively considers the historical contribution degree, current health status and real-time load of the equipment, so as to dynamically reflect the importance and operation status of the equipment in the production process. Among them, the historical contribution degree can be calculated based on data such as the output and energy consumption of the equipment in the historical period, reflecting the historical performance of the equipment. The current health status can be evaluated based on data such as the depreciation years and actual working hours of the equipment, reflecting the current aging degree of the equipment. The real-time load can be calculated based on data such as the actual operating power and rated power of the equipment, reflecting the current operating pressure of the equipment. The calculation of the weight coefficients can be achieved by methods such as weighted average method and exponential weighted method. The calculated weight coefficients are then used to perform weighted aggregation on the original feature vectors of the equipment at the same level on the production line (such as multiple sets of parallel equipment in the same processing stage), thereby obtaining the dynamically weighted feature vectors at the equipment level. The weighted aggregation process can adopt methods such as linear weighting and non-linear weighting to fuse the original feature vectors of the equipment at the same level. This weighted aggregation method can highlight the influence of important equipment and equipment with abnormal states on the production line efficiency.

[0030] More specifically, step S3 aggregates the dynamically weighted feature vectors of the equipment on the same production line according to the production line process flow to form the production line layer feature vectors. Among them, the aggregation of the production line layer feature vectors can be achieved by methods such as summation, averaging, and splicing. The production line layer feature vectors can represent the operation state of the production line as a whole and provide input for the efficiency prediction at the production line level.

[0031] More specifically, in step S4, the equipment efficiency prediction model can be a machine learning model, a deep learning model, etc., such as a support vector machine, a recurrent neural network, etc. This model can be trained using the production line layer feature vectors and equipment efficiency data determined from historical equipment operation data. The equipment efficiency prediction result output by the model can provide an enterprise with a quantitative evaluation of the equipment operation status, assisting managers in making production decisions and equipment maintenance.

[0032] Compared with the traditional method that relies on a centralized server for data processing and model calculation, in the enterprise equipment efficiency dynamic prediction method of the embodiment of the present application, by comprehensively considering the historical contribution, current health status, and real-time load of the equipment to calculate the equipment weight coefficient, the dynamic configuration of the weight coefficient is realized to highlight the impact of important equipment and equipment with abnormal status on the production line efficiency, reduce the impact of unimportant equipment and equipment in good status, and according to the production line process flow, aggregate the dynamic weighted feature vectors of the equipment on the same production line to form a production line layer feature vector, so as to abstract the overall operation status of the production line at a high level, and then input the production line layer feature vector into a pre-constructed equipment efficiency prediction model to generate an equipment efficiency prediction result, and finally realize the prediction output of the equipment efficiency. Its overall processing process involves hierarchical aggregation and dynamic weighting, which can effectively utilize multi-source heterogeneous equipment data and realize the dynamic prediction of equipment efficiency on the resource-constrained edge side, so as to realize the dynamic and accurate prediction of equipment efficiency and provide technical support for the intelligent upgrade of the factory.

[0033] In some preferred embodiments, the historical contribution degree is determined based on the following steps: A1. Obtain the operation data of each device within a preset historical period, where the operation data includes production data, energy consumption data, and operation time data; A2. Calculate the unit time workload corresponding to each device according to the production data of each device; A3. Calculate the unit time energy consumption intensity corresponding to each device according to the energy consumption data and operation time data of each device; A4. Calculate the historical contribution degree corresponding to each device according to the unit time workload and the unit time energy consumption intensity.

[0034] Specifically, the preset historical period can be set according to actual needs; in step A2, the unit time workload is calculated by dividing the total production volume of the device within the preset historical period by the total operation time, so as to quantify the production capacity of the device per unit time. The workload unit can be quantifiable indicators such as the number of products or the number of processed parts.

[0035] More specifically, in step A3, the energy consumption intensity per unit time is calculated by dividing the total energy consumption of the device during the historical period by the total operating time, so as to evaluate the energy consumption level of the device per unit time. The energy consumption unit can be kilowatt-hour, joule, etc.

[0036] More specifically, in step A4, the historical contribution degree is obtained by dividing the workload per unit time by the energy consumption intensity per unit time, that is, historical contribution degree = workload per unit time / energy consumption intensity per unit time. It reflects the workload generated by the device under unit energy consumption. The higher the ratio, the higher the historical contribution degree of the device, which can quantify the effectiveness level of the device during the historical period.

[0037] More specifically, the above processing method clarifies the acquisition path of the historical contribution degree of each device. By calculating the workload per unit time and the energy consumption intensity per unit time, the historical operation data of each device is quantified and standardized. The historical contribution degree obtained by combining the workload per unit time and the energy consumption intensity per unit time reflects the level of energy consumption intensity while the device generates unit workload. By clarifying the calculation method of the historical contribution degree, the historical performance of the device can be evaluated more objectively, and thus a more accurate basis can be provided for the calculation of the subsequent device weight coefficient, ultimately improving the accuracy of device effectiveness prediction.

[0038] In some preferred embodiments, the steps of calculating the weight coefficient of each device based on the historical contribution degree, current health status and real-time load of each device include: S21. Obtain the depreciation life data and actual working hours data of each device, and calculate the current health status of each device according to the depreciation life data and the actual working hours data; S22. Calculate the real-time load of each device according to the actual operating power and rated power of each device; S23. Calculate the dynamic adjustment coefficient of each device according to the current health status and real-time load; S24. Calculate the weight coefficient of each device by combining the dynamic adjustment coefficient and the historical contribution degree.

[0039] Specifically, in step S21, the current health status quantifies the health level of the device by the ratio of the used duration of the device to the theoretical life duration of the device. The depreciation life data represents the theoretical service life of the device, and the preset annual working hours is the planned working time of the device per year. The product of the two gives the total theoretical working hours of the device. Among them, the calculation formula of the current health status is preferably: current health status = 1 - (actual working hours data / (depreciation life data * preset annual working hours)). Through this formula calculation, the health status of the device can be quantified. The longer the working time, the lower the current health status value.

[0040] More specifically, in step S22, the real-time load is calculated from the actual operating power and the rated power of the device, which reflects the current operating intensity of the device. Among them, the calculation formula of the real-time load is preferably: real-time load = actual power / rated power.

[0041] More specifically, in step S23, the dynamic adjustment coefficient is preferably obtained by weighted summation of the current health state and the real-time load, and is preferably calculated by the following formula: dynamic adjustment coefficient = α * current health state + β * real-time load, where the weight coefficients α and β are used to adjust the relative importance of the health state and the real-time load in the dynamic adjustment coefficient, and can be adjusted according to actual application requirements.

[0042] More specifically, in step S24, the weight coefficient is preferably obtained by multiplying the historical contribution degree by the device dynamic adjustment coefficient. Among them, the historical contribution degree reflects the efficiency output ability of the device in the historical period, and the dynamic adjustment coefficient adjusts the contribution degree according to the current state of the device (current health state and real-time load) to obtain the weight coefficient, so that the final weight coefficient comprehensively considers the historical performance and the current state of the device.

[0043] More specifically, the above processing process combines the dynamic adjustment coefficient with the historical contribution degree. The finally obtained weight coefficient takes into account both the historical efficiency of the device and the current operating state of the device, so that the device weight coefficient can more accurately and dynamically reflect the device efficiency. Thus, using this weight coefficient for device efficiency prediction can improve the accuracy and reliability of the prediction, and solve the problem of low prediction accuracy caused by the unclear quantification method of the weight coefficient.

[0044] In some preferred embodiments, step S2 further includes the following steps: S25. Obtain the operating parameters of each device, where the operating parameters include actual working voltage information, actual working current information, and actual ambient temperature information; S26. Calculate the parameter adjustment coefficient of each device according to the actual working voltage information, actual working current information, and actual ambient temperature information; S27. Compensate and adjust the weight coefficient of each device based on the parameter adjustment coefficient.

[0045] Specifically, in step S25, the operating parameters can be obtained by real-time collection via sensors. The sensors are configured to monitor the actual working voltage, actual working current, and actual ambient temperature of the device, and these parameters can directly reflect the real-time state of the device operation.

[0046] More specifically, in step S26, the calculation process of the parameter adjustment coefficient may include parameter standardization and weighted summation processing, so that different devices have standardized parameter adjustment coefficients, and the parameter adjustment coefficient can quantify the influence degree of operating parameters on the device efficiency.

[0047] More specifically, step S27 is used to compensate and adjust the weight coefficient of each device according to the parameter adjustment coefficient, realizing the dynamic adjustment of the weight coefficient based on the operating parameters, so that the adjusted weight coefficient can reflect the real-time change of the device operating parameters.

[0048] More specifically, the above processing process further considers the influence of device operating parameters on the weight coefficient. Among them, the parameter adjustment coefficient obtained by combining step S25 and step S26 aims to quantify the comprehensive influence of voltage, current and temperature on the device efficiency. These parameters can more deeply reflect the real-time state of device operation and the change of external environment. The weight coefficient after compensation and adjustment in step S27 can more comprehensively and accurately reflect the actual efficiency state of the device.

[0049] More specifically, in step S27, the weight coefficient is preferably adjusted based on the following formula: the weight coefficient after compensation and adjustment = the weight coefficient before compensation and adjustment * (1 + parameter adjustment coefficient).

[0050] In some preferred embodiments, step S26 includes: S261. Based on a preset standardization conversion formula, convert the actual working voltage information, actual working current information and actual ambient temperature information into standardized voltage, standardized current and standardized temperature respectively; S262. According to the preset voltage influence factor, current influence factor and temperature influence factor, weighted sum the standardized voltage, standardized current and standardized temperature respectively to obtain the parameter adjustment coefficient corresponding to each device.

[0051] Specifically, in step S261, the preset standardization conversion formula is used to convert the actual working voltage information, actual working current information and actual ambient temperature information into dimensionless standardized values. The standardization conversion formula is preferably: parameter standardization value = (actual parameter value - parameter mean) / parameter standard deviation. This standardization conversion formula can eliminate the influence caused by different physical dimensions and value ranges. By subtracting the parameter mean and dividing by the parameter standard deviation, the original data is converted into the standard normal distribution range with a mean of 0 and a standard deviation of 1. Thus, voltage, current and temperature data with different dimensions and value ranges can be compared and calculated under the same standard.

[0052] More specifically, in step S262, the preset voltage influence factor, current influence factor, and temperature influence factor represent the relative importance of the impacts of voltage, current, and temperature on the device performance, and they can be fixed values preset according to the device characteristics and operation experience. After obtaining the standardized voltage, standardized current, and standardized temperature in this step, multiply them by the corresponding influence factors respectively and sum the products to obtain the parameter adjustment coefficient, so that the parameter adjustment coefficient comprehensively considers the combined impact of voltage, current, and temperature on the device performance.

[0053] More specifically, the above processing process standardizes the actual working voltage information, actual working current information, and actual ambient temperature information, eliminates the impacts of dimension and numerical range differences on subsequent calculations, and then realizes the weighted summation of different operating parameters by introducing the preset influence factors, thereby obtaining a comprehensive parameter adjustment coefficient, enabling the parameter adjustment coefficient to more scientifically and reasonably reflect the combined impact of operating parameters on the device performance, and further being used for the compensation adjustment of the device weight coefficient, making the adjustment of the device weight coefficient more refined and accurate.

[0054] In some preferred embodiments, step S1 includes: S11. Obtain the original data of different types of devices, evaluate the quality of the original data of each device based on the data missing rate. If the data missing rate of the original data is lower than the preset threshold, it is determined as high-quality data; otherwise, it is determined as low-quality data. S12. For high-quality data, adopt an outlier detection method based on statistical principles for data cleaning and perform format standardization processing to obtain the cleaned high-quality data. S13. For low-quality data, adopt a missing value filling method based on data mining for data cleaning and perform format standardization processing to obtain the cleaned low-quality data. S14. Extract the time-domain statistical features, frequency-domain statistical features, and wavelet transform features of the cleaned high-quality data and the cleaned low-quality data respectively to form the original feature vector of each device.

[0055] Specifically, in step S11, the data missing rate evaluation is used for the preliminary judgment of the quality of the original data of each device, providing a basis for subsequent data processing, and it can be realized by collecting communication data packets within a preset time window for data collection frequency analysis.

[0056] More specifically, steps S12 and S13 adopt different data cleaning strategies for different quality data to ensure the effectiveness and pertinence of data cleaning.

[0057] More specifically, for data cleaning of high-quality data in step S12, a statistical outlier detection method can be adopted, such as the 3σ principle or the box plot method, to identify and process outliers. For the format standardization process, operations such as timestamp alignment, data type conversion, and dimension unification can be used to ensure data format consistency.

[0058] More specifically, for data cleaning of low-quality data in step S13, a data mining missing value filling method can be adopted, such as K-nearest neighbor filling, regression filling, or multiple imputation, to recover the information of missing data. The format standardization process is the same as that of high-quality data.

[0059] More specifically, in step S14, the time-domain statistical features can include mean, standard deviation, peak value, kurtosis, etc. The frequency-domain statistical features can be obtained through Fourier transform, such as energy spectrum, power spectrum, etc. The wavelet transform features can be extracted by discrete wavelet transform, such as the wavelet coefficients of the db4 wavelet basis; these features are combined into an original feature vector to comprehensively characterize the operating state of the device.

[0060] More specifically, the above processing process distinguishes data through quality assessment and adopts different data cleaning methods, ensuring the effectiveness and pertinence of data cleaning in the case of differences in the quality of the original data, providing high-quality feature data support for subsequent device efficiency prediction, and thus improving the accuracy of device efficiency prediction.

[0061] In some preferred embodiments, the process of obtaining the data missing rate includes: B1. Collect communication data packets of each device within a preset time window, and parse the communication data packets to extract the data collection frequency; B2. If the data collection frequency is lower than the preset frequency threshold, it is determined that data is missing, and the data missing rate is calculated according to the data collection frequency and the preset frequency threshold.

[0062] Specifically, in step S1, parsing the communication data packets to extract the data collection frequency can adopt conventional existing data parsing and extraction means, so it will not be elaborated here.

[0063] More specifically, in step B2, the preset frequency threshold is the data transmission speed that the device should have under normal working conditions, which can be adjusted according to the device type, data application scenario, and the requirement for data integrity; when the actual data collection frequency is lower than this threshold, the system determines that the data is missing; in the embodiments of the present application, the data missing rate is preferably calculated based on the following formula: data missing rate = (1 - actual collection frequency / preset collection frequency) * 100%.

[0064] More specifically, the above formula quantifies the deviation degree between the actual acquisition frequency and the preset frequency threshold, so as to obtain the degree of data loss, providing data support for the selection of subsequent differential data cleaning and feature extraction strategies.

[0065] In some preferred embodiments, step S3 includes: S31. Obtain an equipment relationship graph, where the equipment relationship graph represents equipment with nodes, represents the technological process relationship between equipment with edges, and the weight of the edge represents the material transfer amount between equipment; S32. Based on the equipment relationship graph and the magnitude relationship of the weights of the edges, search for and obtain a production line chain; S33. Extract the dynamic weighted feature vectors of the corresponding equipment according to the production line chain, and integrate them using the weights of the edges to generate a production line layer feature vector.

[0066] Specifically, the equipment relationship graph in step S1 is a pre-constructed and determined relationship graph, which can be constructed in the following way: Determine all the equipment in the factory workshop as the nodes of the graph, and then determine the connection relationships between the equipment according to the actual production process. These connection relationships form the edges of the graph, and the weights of the edges can be configured and set according to the average material transfer amount between the equipment or the tightness of production collaboration between the equipment. Thus, the interdependent relationship between the equipment on the production line and the material flow situation can be represented by this equipment relationship graph.

[0067] More specifically, in step S32, the process of searching for and obtaining a production line chain can be: Starting from the starting equipment of the equipment relationship graph, use a preset search algorithm to select the equipment with the greatest relationship connected to the current equipment based on the magnitude relationship of the weights of the edges, reach the next equipment along this edge, and add this equipment to the production line chain. Repeat this process until reaching the end equipment of the production line or the weight of the edge is lower than the set threshold.

[0068] More specifically, in step S33, the process of generating a production line layer feature vector can be: Obtain the dynamic weighted feature vectors of each equipment on the production line chain, and then perform weighted averaging or weighted integration on these feature vectors according to the position of the equipment in the production line and the weights of the edges to generate the final production line layer feature vector. Among them, the weights of the edges can reflect the importance of the equipment in the production line and the material relevance. Thus, the core features of the production line can be more effectively extracted, reducing the interference of non-critical equipment or low-material-transfer equipment on the production line layer feature vector, thereby improving the representativeness of the production line layer feature vector and the accuracy of the prediction model.

[0069] In some preferred embodiments, step S32 includes: S321. Initialize the production line chain as an empty linked list according to the equipment relationship graph, and use the weight of the edges in the equipment relationship graph as the material transfer efficiency between equipment. S322. Select the starting equipment from the equipment relationship graph. Based on the magnitude relationship of the edge weights, use the greedy algorithm to preferentially select the downstream equipment that is connected to the current equipment and has the highest material transfer efficiency. Add this downstream equipment to the production line chain, and update the current equipment to this downstream equipment. Repeat this step until reaching the end equipment or the material transfer efficiency is lower than the preset threshold. S323. If the number of equipment in the production line chain is less than the preset quantity threshold, then lower the material transfer efficiency threshold and re - execute S322. Otherwise, use the current production line chain as the production line chain obtained by the search.

[0070] Specifically, in step S321, the equipment relationship graph can be constructed as a data structure, such as an adjacency list or an adjacency matrix, for representing the connection relationship between equipment and the material transfer efficiency. The weight of the edge can be quantified as a specific value, such as the amount of material transferred per unit time. The purpose of initializing the empty linked list is to store the equipment on the production line chain searched, laying a foundation for subsequent searches.

[0071] More specifically, step S322 introduces the greedy algorithm. After selecting the starting equipment from the equipment relationship graph, it preferentially selects the downstream equipment that is connected to the current equipment and has the highest material transfer efficiency, and adds this downstream equipment to the production line chain. Repeat this step until reaching the end equipment or the material transfer efficiency is lower than the preset threshold. Thus, it ensures the effectiveness and efficiency of equipment connection during the construction of the production line chain, and ensures the integrity and rationality of the production line chain.

[0072] More specifically, step S323 considers the number of equipment in the production line chain. If the number of equipment in the production line chain is less than the preset quantity threshold, then lower the material transfer efficiency threshold and re - execute S322 to ensure that the production line chains that may be missed when the material transfer efficiency threshold is relatively high are searched again, improving the integrity of the production line chain search. If the number of equipment in the production line chain is not less than the preset quantity threshold, then use the current production line chain as the production line chain obtained by the search to ensure the effectiveness of the search result. Among them, the preset quantity threshold can be set according to the number of equipment in the actual production line.

[0073] More specifically, the above - mentioned processing method uses the greedy algorithm to search for the production line chain. It can efficiently find the production line chain that meets the material transfer efficiency requirements in the case of complex equipment relationships. By preferentially selecting the paths with high material transfer efficiency, it ensures the efficiency and quality of the production line chain. By dynamically adjusting the material transfer efficiency threshold when the number of equipment in the production line chain is insufficient, it improves the integrity of the search result and ensures that potential production line chains will not be missed.

[0074] In some preferred embodiments, step S33 includes: S331. Obtain the dynamic weighted feature vector of the devices corresponding to the production line chain, and the production line chain is denoted as: C k =(e k,1 ,e k,2 ,…e k,i ,…,e k,mk ), where C k is the k-th production line chain, which is composed of a sequence of device nodes, e k,i represents the i-th device node on the k-th production line chain, and mk is the total number of devices; S332. Obtain the static connection weight between device node e k,i and e k,i+1 according to the weight of the edge ; S333. Generate a dynamic normalization factor S k that satisfies: (1) S334. Calculate the production line layer feature vector L k that satisfies: (2) where is the dynamic weighted feature vector of device node e k,i , is the dynamic weighted feature vector of the device node at the very end.

[0075] Specifically, step S331 is used to extract the dynamic weighted feature vector of the device corresponding to the device node, laying a data foundation for the subsequent feature integration process.

[0076] More specifically, in step S332, the static connection weight is obtained according to the weight of the edge in the device relationship graph, and this weight represents the material transfer amount between devices.

[0077] More specifically, in step S333, the dynamic normalization factor is calculated by summing the connection weights of all adjacent devices in the production line chain and then adding 1, and is used for subsequent vector normalization processing.

[0078] More specifically, in step S334, for each pair of adjacent devices in the production line chain, the dynamic weighted feature vector of the previous device is multiplied by their static connection weight, all the product results are accumulated, and the dynamic weighted feature vector of the last device in the production line chain is added. The final result is divided by the normalization factor to obtain the production line layer feature vector.

[0079] More specifically, in the above processing method, for each pair of adjacent devices, the dynamic weighted feature vector of the previous device is multiplied by the static connection weight between them, and these products are accumulated. The accumulated result is divided by the dynamic normalization factor to obtain the production line layer feature vector. Through this weighted summation strategy in step S33, the connection strength information between devices can be effectively incorporated into the production line layer feature vector. Among them, the greater the connection weight, the stronger the correlation between devices, so that the operating state of the entire production line can be more accurately reflected, enabling the subsequent device efficiency prediction model to make more accurate predictions based on this production line layer feature vector, overcoming the problem in traditional methods that the connection relationship between devices is not fully considered, and improving the accuracy and reliability of the prediction.

[0080] In some preferred embodiments, the device efficiency prediction model includes a teacher model and a student model simplified and compressed based on the teacher model. Step S4 includes: S41. Monitor the resource utilization rate of the edge device. The resource utilization rate includes CPU utilization rate, memory utilization rate, and storage utilization rate; S42. If any resource utilization rate exceeds the preset threshold, input the production line layer feature vector into the student model to generate a device efficiency prediction result; if all resource utilization rates do not exceed the preset threshold, input the production line layer feature vector into the teacher model to generate a device efficiency prediction result.

[0081] Specifically, the monitoring of the resource utilization rate can be achieved by deploying a monitoring agent on the edge device to periodically collect system performance index data. The preset threshold can be configured as a critical value reflecting the resource load status of the edge device and can be set according to usage requirements.

[0082] More specifically, step S42 dynamically selects the device efficiency prediction model based on the resource utilization rate monitored in step S41. Among them, the device efficiency prediction model includes a teacher model and a student model. The teacher model can be a complex model structure, such as a deep neural network model, with high prediction accuracy but large computational resource consumption. The student model is a lightweight model obtained by simplifying and compressing the teacher model, such as a shallow neural network model or a linear regression model. The prediction accuracy may be slightly reduced, but the resource consumption is significantly reduced. When step S41 monitors that any resource utilization rate exceeds the preset threshold, it indicates that the edge device resources are tense, and the system switches to the student model for prediction. When all resource utilization rates do not exceed the preset threshold, it indicates that the edge device resources are sufficient, and the system selects the teacher model for prediction. Thus, through the dynamic switching and use of the teacher model and the student model, the technical solution can adaptively select an appropriate prediction model under different resource load states, effectively reducing the resource consumption of model deployment and operation on the edge side while ensuring the prediction accuracy of device efficiency, and is more suitable for application scenarios with limited edge computing resources.

[0083] More specifically, the student model learns knowledge from the teacher model through model compression and simplification techniques such as knowledge distillation, model pruning, etc., and significantly reduces the model complexity. On the premise of ensuring a certain prediction accuracy, the student model significantly reduces the computational amount and resource consumption, so that it can run quickly on resource-constrained edge devices and reduce the resource occupancy rate.

[0084] For a second aspect, please refer to Figure 2 , some embodiments of the present application further provide an enterprise equipment efficiency dynamic prediction device for predicting equipment efficiency, including: An acquisition module 201, configured to acquire the original data of different types of equipment, perform data cleaning and format unification processing on the original data, and extract the original feature vectors of each equipment, where the original feature vectors represent the key feature parameters of the corresponding equipment operating state; A first integration module 202, configured to calculate the weight coefficient of each equipment based on the historical contribution degree, current health state, and real-time load of each equipment, and perform weighted aggregation on the original feature vectors of the equipment at the same level on the production line according to the weight coefficient to obtain a dynamic weighted feature vector at the equipment level; A second integration module 203, configured to aggregate the dynamic weighted feature vectors of the equipment on the same production line according to the production line process flow to form a production line layer feature vector; A prediction module 204, configured to input the production line layer feature vector into a pre-constructed equipment efficiency prediction model to generate an equipment efficiency prediction result.

[0085] The enterprise equipment efficiency dynamic prediction device of the embodiments of the present application realizes the dynamic configuration of the weight coefficient by comprehensively considering the historical contribution, current health state, and real-time load of the equipment, so as to highlight the influence of important equipment and equipment with abnormal states on the production line efficiency, reduce the influence of unimportant equipment and equipment in good states, and aggregate the dynamic weighted feature vectors of the equipment on the same production line according to the production line process flow to form a production line layer feature vector, so as to abstract the overall operation state of the production line at a high level, and then input the production line layer feature vector into a pre-constructed equipment efficiency prediction model to generate an equipment efficiency prediction result, and finally realize the prediction output of the equipment efficiency. Its overall processing process involves hierarchical aggregation and dynamic weighting, which can effectively utilize multi-source heterogeneous equipment data and realize the dynamic prediction of equipment efficiency on the resource-constrained edge side, so as to realize the dynamic and accurate prediction of equipment efficiency and provide technical support for the intelligent upgrade of the factory.

[0086] In some preferred embodiments, the enterprise equipment efficiency dynamic prediction device of the embodiments of the present application is used to execute the enterprise equipment efficiency dynamic prediction method provided in the above first aspect.

[0087] In a third aspect, please refer to Figure 3 , some embodiments of the present application further provide a schematic structural diagram of an electronic device. The present application provides an electronic device, including: a processor 301 and a memory 302. The processor 301 and the memory 302 are interconnected and communicate with each other through a communication bus 303 and / or other forms of connection mechanisms (not shown), and the memory 302 stores computer-readable instructions executable by the processor 301. When the electronic device runs, the processor 301 executes the computer-readable instructions to execute the method in any optional implementation manner of the above embodiments.

[0088] In a fourth aspect, embodiments of the present application provide a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method in any optional implementation manner of the above embodiments is executed. Among them, the computer-readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM for short), electrically erasable programmable read-only memory (EEPROM for short), erasable programmable read-only memory (EPROM for short), programmable read-only memory (PROM for short), read-only memory (ROM for short), magnetic memory, flash memory, magnetic disk or optical disk.

[0089] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For another example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some communication interfaces, and the indirect coupling or communication connection of the devices or units can be in an electrical, mechanical or other form.

[0090] In addition, the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed over multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0091] Furthermore, in each embodiment of the present application, the various functional modules may be integrated together to form an independent part, or each module may exist alone, or two or more modules may be integrated to form an independent part.

[0092] In this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations.

[0093] The above description is only for the embodiments of the present application and is not used to limit the protection scope of the present application. For those skilled in the art, the present application can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for dynamically predicting the efficiency of enterprise equipment, which is used for predicting the equipment efficiency, and is characterized in that, The method includes the following steps: S1. Obtain the original data of different types of devices, perform data cleaning and format standardization on the original data, and extract the original feature vectors of each device, where the original feature vectors represent the key feature parameters of the operating state of the corresponding device; S2. Calculate the weight coefficients of each device based on the historical contribution degree, current health status, and real-time load of each device, and perform weighted aggregation on the original feature vectors of the devices at the same level on the production line according to the weight coefficients to obtain the dynamic weighted feature vectors at the device level; S3. Aggregate the dynamic weighted feature vectors of the devices on the same production line according to the production line process flow to form the feature vectors at the production line level; S4. Input the feature vectors at the production line level into a pre-constructed device efficiency prediction model to generate the device efficiency prediction results.

2. The dynamic prediction method for enterprise equipment effectiveness according to claim 1, characterized in that The historical contribution degree is determined based on the following steps: A1. Obtain the operation data of each device within a preset historical period, where the operation data includes production data, energy consumption data, and operation time data; A2. Calculate the unit time workload corresponding to each device according to the production data of each device; A3. Calculate the unit time energy consumption intensity corresponding to each device according to the energy consumption data and operation time data of each device; A4. Calculate the historical contribution degree corresponding to each device according to the unit time workload and the unit time energy consumption intensity.

3. A dynamic prediction method for enterprise equipment efficiency according to claim 1, characterized in that The step of calculating the weight coefficients of each device based on the historical contribution degree, current health status, and real-time load of each device includes: S21. Obtain the depreciation life data and actual working hours data of each device, and calculate the current health status of each device according to the depreciation life data and the actual working hours data; S22. Calculate the real-time load of each device according to the actual operating power and rated power of each device; S23. Calculate the dynamic adjustment coefficient of each device according to the current health status and the real-time load; S24. Calculate the weight coefficients of each device by combining the dynamic adjustment coefficient and the historical contribution degree.

4. A dynamic prediction method for enterprise equipment efficiency according to claim 3, characterized in that, Step S2 further includes the following steps: S25. Obtain the operation parameters of each device, where the operation parameters include actual working voltage information, actual working current information, and actual ambient temperature information; S26. Calculate the parameter adjustment coefficient of each device according to the actual working voltage information, actual working current information, and actual ambient temperature information; S27. Compensate and adjust the weight coefficients of each device based on the parameter adjustment coefficient.

5. A dynamic prediction method for enterprise equipment efficiency according to claim 1, characterized in that Step S1 includes: S11. Obtain the original data of different types of devices, evaluate the quality of the original data of each device based on the data missing rate. If the data missing rate of the original data is lower than the preset threshold, it is determined as high-quality data, otherwise it is determined as low-quality data; S12. For the high-quality data, use an outlier detection method based on statistical principles for data cleaning and perform format standardization to obtain the cleaned high-quality data; S13. For the low-quality data, use a missing value filling method based on data mining for data cleaning and perform format standardization to obtain the cleaned low-quality data; S14. Respectively extract the time-domain statistical features, frequency-domain statistical features, and wavelet transform features of the cleaned high-quality data and the cleaned low-quality data to form the original feature vectors of each device.

6. A dynamic prediction method for enterprise equipment efficiency according to claim 1, characterized in that Step S3 includes: S31. Obtain a device relationship graph, where the nodes in the device relationship graph represent devices, the edges represent the technological process relationships between devices, and the weights of the edges represent the material transfer amounts between devices; S32. Based on the device relationship graph and the magnitude relationship of the edge weights, search and obtain a production line chain; S33. Extract the dynamic weighted feature vectors of the corresponding devices according to the production line chain, and use the edge weights for integration to generate the production line layer feature vector.

7. A dynamic prediction method for enterprise equipment efficiency according to claim 1, characterized in that, The device efficiency prediction model includes a teacher model and a student model compressed and simplified based on the teacher model. Step S4 includes: S41. Monitor the resource utilization rate of edge devices, where the resource utilization rate includes CPU utilization rate, memory utilization rate, and storage utilization rate; S42. If any resource utilization rate exceeds the preset threshold, input the production line layer feature vector into the student model to generate a device efficiency prediction result; if all resource utilization rates do not exceed the preset threshold, input the production line layer feature vector into the teacher model to generate a device efficiency prediction result.

8. An enterprise equipment efficiency dynamic prediction device for predicting equipment efficiency, characterized in that, It includes: An acquisition module, configured to acquire the original data of different types of devices, perform data cleaning and format standardization processing on the original data, and extract the original feature vectors of each device, where the original feature vectors represent the key feature parameters of the operating states of the corresponding devices; A first integration module, configured to calculate the weight coefficients of each device based on the historical contribution degree, current health status, and real-time load of each device, and perform weighted aggregation on the original feature vectors of the devices at the same level on the production line according to the weight coefficients to obtain the dynamic weighted feature vectors at the device level; A second integration module, configured to aggregate the dynamic weighted feature vectors of the devices on the same production line according to the production line technological process to form a production line layer feature vector; A prediction module, configured to input the production line layer feature vector into a pre-constructed device efficiency prediction model to generate a device efficiency prediction result.

9. An electronic device, characterized in that, It includes a processor and a memory, where the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps in the method according to any one of claims 1-7 are run.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps in the method according to any one of claims 1-7 are run.

Images