Intelligent Decision Management Method and Device for Energy-saving Equipment Status Monitoring and Fault Diagnosis

Abstract

The embodiments of the present application provide an intelligent decision-making management method and device for monitoring the state and diagnosing faults of energy-saving equipment. Through an edge computing platform integrating signal conditioning, gain amplification, and digital-to-analog conversion units, high-precision data acquisition and processing are achieved. The system realizes adaptive sampling based on time-series feature extraction and working condition judgment models, and innovatively constructs a collaborative diagnosis system combining rule models, machine learning, and deep networks. By embedding the device feature mapping library into the diagnosis system, combining trend modeling and remaining life prediction, a risk classification and early warning system is established to realize intelligent decision-making and remote management of equipment maintenance. This method provides a comprehensive solution for the predictive maintenance of industrial equipment.

Description

Technical Field

[0001] This application relates to the field of data processing, and particularly to an intelligent decision-making management method and device for energy-saving equipment status monitoring and fault diagnosis. Background Art

[0002] Traditional methods for energy-saving equipment status monitoring and fault diagnosis mainly rely on simple data collection and basic signal processing technologies, and it is difficult to meet the intelligent monitoring requirements of modern industrial equipment under complex working conditions. There are obvious deficiencies in data collection accuracy, real-time performance, and anti-interference ability in the existing technologies, and there is a lack of effective fusion and analysis capabilities for multi-source heterogeneous data.

[0003] At the same time, there are also limitations in the existing systems in terms of fault diagnosis model construction and predictive maintenance. Traditional methods often adopt a single diagnostic model, fail to fully utilize the collaborative advantages of multiple models, and lack in-depth analysis of the correlation between equipment component parameters and fault types. The system is relatively simple in terms of fault feature extraction and pattern recognition, and it is difficult to achieve accurate diagnosis under complex working conditions.

[0004] In addition, the existing technologies also need to be improved in terms of equipment life prediction and maintenance decision-making. There is a lack of effective trend modeling methods and risk warning mechanisms, and early warning of equipment faults and scientific decision-making cannot be achieved. The formulation of maintenance plans often relies on manual experience, lacking systematic and intelligent decision-making support.

[0005] In terms of the edge computing architecture, the existing systems generally have problems such as low utilization rate of computing resources and insufficient real-time processing capabilities. There is a lack of efficient data preprocessing and adaptive sampling mechanisms, and the real-time requirements of large-scale equipment monitoring cannot be met. Solving these problems is of great significance for improving the intelligent level and operation and maintenance efficiency of industrial equipment management. Summary of the Invention

[0006] In view of the problems in the existing technologies, this application provides an intelligent decision-making management method and device for energy-saving equipment status monitoring and fault diagnosis, which can establish a risk grading and early warning system, realize intelligent decision-making and remote management of equipment maintenance, and provide a comprehensive solution for the predictive maintenance of industrial equipment.

[0007] To solve at least one of the above problems, this application provides the following technical solutions:

[0008] In a first aspect, this application provides an intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis, including:

[0009] Integrate the signal conditioning module, gain amplification unit, analog-to-digital conversion unit and operation processor into the edge computing platform, connect the environmental parameter acquisition module and the energy consumption acquisition module to the data acquisition channel of the analog-to-digital conversion unit, filter and amplify the acquired signal based on the signal conditioning module, adjust the gain coefficient of the acquisition channel through the gain amplification unit, use the analog-to-digital conversion unit to convert the conditioned signal into a digital quantity, and transmit the digital quantity to the multi-level data cache through the bus;

[0010] Input the monitoring data in the multi-level data cache into the time series feature extraction model, perform principal component analysis on the extracted time series features to obtain feature vectors, establish a device condition judgment model based on the feature vectors, adjust the sampling frequency according to the condition judgment model, input the adjusted sampling data into the data fusion model, input the fusion data into the collaborative diagnosis system composed of the rule model, machine learning model and deep network model, embed the feature mapping library constructed by the device component parameters, fault types and fault modes into the collaborative diagnosis system, correct the diagnosis result based on the feature mapping library, and input the corrected diagnosis result into the decision-making system;

[0011] Input the output result of the decision-making system and the device historical data into the trend modeling unit, construct a remaining life prediction model based on the trend modeling unit, combine the fault evolution analysis to construct a risk warning function, combine the risk warning function with the remaining life prediction model to establish a hierarchical warning system, generate a maintenance plan according to the hierarchical warning system, input the maintenance plan into the maintenance decision-making unit to generate a device maintenance plan, and allocate the maintenance tasks through the remote communication module based on the device maintenance plan.

[0012] Further, the integration of the signal conditioning module, gain amplification unit, analog-to-digital conversion unit and operation processor into the edge computing platform, and the connection of the environmental parameter acquisition module and the energy consumption acquisition module to the data acquisition channel of the analog-to-digital conversion unit include:

[0013] Construct a heterogeneous computing architecture using a dual-core processor and a programmable gate array, configure the master-slave interface, storage interface and communication interface of the heterogeneous computing architecture as data processing units based on the fieldbus protocol, connect the amplifier circuit, filter circuit and isolation circuit of the signal conditioning module to the input end of the data processing unit, and connect the programmable amplifier and gain adjustment circuit of the gain amplification unit to the output end of the signal conditioning module;

[0014] Connect the sample-and-hold circuit, reference source circuit, and digital quantity output circuit of the digital-to-analog conversion unit to the signal output terminal of the gain amplification unit. Configure the logic function module of the programmable gate array based on the hardware description language. Connect the temperature sensor and humidity sensor of the environmental parameter acquisition module and the current transformer and voltage sensor of the energy consumption acquisition module to the multiple sampling channels of the digital-to-analog conversion unit respectively.

[0015] Further, filter and amplify the acquired signal based on the signal conditioning module, adjust the gain coefficient of the acquisition channel through the gain amplification unit, convert the conditioned signal into a digital quantity using the digital-to-analog conversion unit, and transmit the digital quantity to the multi-level data cache through the bus, including:

[0016] Input the acquired signal into a passive filter circuit in the signal conditioning module for band-pass filtering, perform common-mode rejection on the filtered signal through a differential amplifier, input the rejected signal into an operational amplifier for signal amplification, electrically isolate the amplified signal through an opto-isolator, and perform secondary amplification on the isolated signal based on a programmable gain amplifier. Set the gain coefficient of the programmable gain amplifier as an adjustable parameter, and dynamically adjust the gain coefficient according to the signal amplitude.

[0017] Use the sample-and-hold circuit in the digital-to-analog conversion unit to sample the conditioned signal, compare the sampled signal with the reference voltage generated by the reference source circuit to obtain a digital quantity, transmit the digital quantity to the field programmable gate array buffer through the serial peripheral interface, and transmit the data in the field programmable gate array buffer to the multi-level data cache composed of a dynamic random access memory and a solid-state memory through the system bus.

[0018] Further, input the monitoring data in the multi-level data cache into a time series feature extraction model, perform principal component analysis on the extracted time series features to obtain feature vectors, establish a device condition judgment model based on the feature vectors, adjust the sampling frequency according to the condition judgment model, and input the adjusted sampling data into a data fusion model, including:

[0019] Divide the monitoring data in the multi-level data cache into multiple data windows according to the time series, extract statistical features, frequency domain features, and time-frequency features from the data windows respectively. Combine the mean, variance, skewness, and kurtosis of the statistical features, the frequency center, frequency variance, and power spectral density of the frequency domain features, and the wavelet coefficients and energy entropy of the time-frequency features to form a feature matrix. Standardize the feature matrix, input the standardized feature matrix into the principal component analysis model, calculate the principal component scores based on eigenvalue decomposition, and construct feature vectors from the principal component scores.

[0020] Construct a working condition judgment model based on the support vector machine, input the feature vector into the working condition judgment model for classification, establish a corresponding relationship between the classification result and a preset sampling frequency threshold, select the sampling frequency parameter corresponding to the current working condition according to the corresponding relationship, input the data collected under the sampling frequency parameter into the data fusion model, and calculate the weight coefficients of different data sources based on the attention mechanism in the data fusion model, and perform weighted fusion on the multi-dimensional features according to the weight coefficients.

[0021] Further, input the fusion data into a collaborative diagnosis system composed of a rule model, a machine learning model, and a deep network model, embed a feature mapping library constructed from equipment component parameters, fault types, and fault modes into the collaborative diagnosis system, correct the diagnosis result based on the feature mapping library, and input the corrected diagnosis result into the decision-making system, including:

[0022] Construct a rule model by dividing the fault diagnosis expert knowledge according to equipment components, construct a machine learning model based on the gradient boosting tree algorithm, construct a deep network model using a recurrent neural network, form a collaborative diagnosis system with the rule model, the machine learning model, and the deep network model, input the fusion data into the rule model for rule reasoning, input the machine learning model for classification prediction, and input the deep network model for sequence analysis, construct a feature mapping library from the structural parameters, operating parameters, and fault characteristics of equipment components, and establish a diagnostic confidence scoring system based on the corresponding rules of fault types and fault modes in the feature mapping library;

[0023] Input the diagnosis results of the rule model, the machine learning model, and the deep network model into the diagnostic confidence scoring system for scoring, quantify the reliability of the diagnosis results of each model according to the scoring results, perform weighted combination on the quantified diagnosis results based on the Bayesian inference method to obtain a collaborative diagnosis result, match and correct the collaborative diagnosis result with the fault characteristics in the feature mapping library, and input the corrected diagnosis result into the decision-making system.

[0024] Further, input the output result of the decision-making system and the equipment historical data into the trend modeling unit, construct a remaining life prediction model based on the trend modeling unit, construct a risk warning function in combination with fault evolution analysis, and establish a hierarchical warning system by combining the risk warning function with the remaining life prediction model, including:

[0025] Input the device status indicators output by the decision-making system, the operation records, maintenance records, and fault records in the device historical database into the trend modeling unit. In the trend modeling unit, perform time series decomposition on the status indicators to extract the trend term, cycle term, and random term. Establish a device degradation model based on the Weibull distribution, combine the device degradation model with the long short-term memory network to construct a remaining life prediction model, and use the status indicator decomposition term, degradation parameter, and time scale as input variables in the remaining life prediction model;

[0026] Conduct cluster analysis on the device fault data to identify the fault evolution path, establish a corresponding relationship between the fault evolution path and the status indicator threshold to construct a risk warning function, map the risk level calculated by the risk warning function to the life interval predicted by the remaining life prediction model, and divide the warning level into multiple levels based on the mapping relationship. Set trigger conditions and warning rules for different warning levels respectively to construct a hierarchical warning system.

[0027] Further, generate a maintenance plan according to the hierarchical warning system, input the maintenance plan into the maintenance decision-making unit to generate a device maintenance plan, and allocate maintenance tasks based on the device maintenance plan through the remote communication module, including:

[0028] Associate and match the warning levels in the hierarchical warning system with the maintenance strategy library, construct a maintenance plan generation model based on the maintenance procedures, maintenance cycles, and spare part information in the maintenance strategy library. Use the device status indicators, warning levels, and maintenance resource constraints as input parameters in the maintenance plan generation model to generate a specific maintenance plan, input the maintenance plan into the maintenance decision-making unit, and balance the maintenance cost, maintenance time, and device reliability based on the multi-objective optimization algorithm in the maintenance decision-making unit, and prioritize and schedule the device maintenance items;

[0029] Allocate the maintenance plan output by the maintenance decision-making unit through the remote communication module. Encode the maintenance task information in the remote communication module according to a preset format, establish a communication link for the maintenance task based on the distributed network protocol, and send the encoded maintenance task to the corresponding execution terminal through the communication link. Decode and classify and store the received maintenance task in the execution terminal.

[0030] In a second aspect, the present application provides an intelligent decision-making management device for energy-saving device status monitoring and fault diagnosis, including:

[0031] The platform computing module is used to integrate the signal conditioning module, the gain amplification unit, the analog-to-digital conversion unit and the operation processor into the edge computing platform, connect the environmental parameter acquisition module and the energy consumption acquisition module to the data acquisition channel of the analog-to-digital conversion unit, filter and amplify the acquired signal based on the signal conditioning module, adjust the gain coefficient of the acquisition channel through the gain amplification unit, convert the conditioned signal into a digital quantity by using the analog-to-digital conversion unit, and transmit the digital quantity to the multi-level data cache through the bus;

[0032] The model construction module is used to input the monitoring data in the multi-level data cache into the time series feature extraction model, perform principal component analysis on the extracted time series features to obtain feature vectors, establish a device condition judgment model based on the feature vectors, adjust the sampling frequency according to the condition judgment model, input the adjusted sampling data into the data fusion model, input the fusion data into the collaborative diagnosis system composed of the rule model, the machine learning model and the deep network model, embed the feature mapping library constructed by the device component parameters, fault types and fault modes into the collaborative diagnosis system, correct the diagnosis result based on the feature mapping library, and input the corrected diagnosis result into the decision-making system;

[0033] The monitoring and diagnosis module is used to input the output result of the decision-making system and the device historical data into the trend modeling unit, construct a remaining life prediction model based on the trend modeling unit, combine the fault evolution analysis to construct a risk warning function, combine the risk warning function with the remaining life prediction model to establish a hierarchical warning system, generate a maintenance plan according to the hierarchical warning system, input the maintenance plan into the maintenance decision-making unit to generate a device maintenance plan, and allocate the maintenance tasks through the remote communication module based on the device maintenance plan.

[0034] In a third aspect, the present application provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the intelligent decision management method for monitoring the state and diagnosing faults of the energy-saving device are implemented.

[0035] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the intelligent decision management method for monitoring the state and diagnosing faults of the energy-saving device are implemented.

[0036] In a fifth aspect, the present application provides a computer program product, including a computer program / instructions. When the computer program / instructions are executed by a processor, the steps of the intelligent decision management method for monitoring the state and diagnosing faults of the energy-saving device are implemented.

[0037] As can be seen from the above technical solutions, the present application provides an intelligent decision-making management method and device for energy-saving equipment status monitoring and fault diagnosis. Through an edge computing platform integrating signal conditioning, gain amplification, and analog-to-digital conversion units, high-precision data acquisition and processing are achieved. The system realizes adaptive sampling based on time-series feature extraction and operating condition judgment models, and innovatively constructs a collaborative diagnosis system combining rule models, machine learning, and deep networks. By embedding the equipment feature mapping library into the diagnosis system and combining trend modeling and remaining life prediction, a risk grading early warning system is established to realize intelligent decision-making and remote management of equipment maintenance. This method provides a comprehensive solution for the predictive maintenance of industrial equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0039] Figure 1 It is one of the flow diagrams of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis in the embodiments of the present application;

[0040] Figure 2 It is another flow diagram of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis in the embodiments of the present application;

[0041] Figure 3 It is the third flow diagram of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis in the embodiments of the present application;

[0042] Figure 4 It is the fourth flow diagram of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis in the embodiments of the present application;

[0043] Figure 5 It is the fifth flow diagram of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis in the embodiments of the present application;

[0044] Figure 6 It is the sixth flow diagram of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis in the embodiments of the present application;

[0045] Figure 7 It is the seventh flow diagram of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis in the embodiments of the present application;

[0046] Figure 8Structural diagram of the intelligent decision-making management device for energy-saving equipment status monitoring and fault diagnosis in the embodiments of the present application;

[0047] Figure 9 Schematic structural diagram of the electronic device in the embodiments of the present application.

[0048] Reference numerals:

[0049] Electronic device 9600, central processing unit 9100, memory 9140, communication module 9110, input unit 9120, audio processor 9130, display 9160, power supply 9170, buffer memory 9141, application / function storage unit 9142, data storage unit 9143, driver program storage unit 9144, antenna 9111, speaker 9131, microphone 9132. Detailed implementation manners

[0050] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Apparently, the described embodiments are some but not all of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without making creative efforts shall fall within the protection scope of the present application.

[0051] In the technical solutions of the present application, the acquisition, storage, use, processing, etc. of data all comply with the relevant provisions of national laws and regulations.

[0052] Considering the problems existing in the prior art, the present application provides an intelligent decision-making management method and device for energy-saving equipment status monitoring and fault diagnosis. Through an edge computing platform integrating signal conditioning, gain amplification and analog-to-digital conversion units, high-precision data acquisition and processing are realized. The system realizes adaptive sampling based on time-series feature extraction and working condition judgment models, and innovatively constructs a collaborative diagnosis system combining rule models, machine learning and deep networks. By embedding the device feature mapping library into the diagnosis system and combining trend modeling and remaining life prediction, a risk grading early warning system is established to realize intelligent decision-making and remote management of equipment maintenance. This method provides a comprehensive solution for the predictive maintenance of industrial equipment.

[0053] In order to be able to establish a risk grading early warning system, realize intelligent decision-making and remote management of equipment maintenance, and provide a comprehensive solution for the predictive maintenance of industrial equipment, the present application provides an embodiment of an intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis. Refer to Figure 1 , the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis specifically includes the following contents:

[0054] Step S101: Integrate the signal conditioning module, gain amplification unit, digital-to-analog conversion unit, and arithmetic processor into the edge computing platform. Connect the environmental parameter acquisition module and energy consumption acquisition module to the data acquisition channels of the digital-to-analog conversion unit. Filter and amplify the acquired signals based on the signal conditioning module, adjust the gain coefficient of the acquisition channels through the gain amplification unit, convert the conditioned signals into digital quantities using the digital-to-analog conversion unit, and transmit the digital quantities to the multi-level data cache through the bus;

[0055] Optionally, in the hardware integration process of the edge computing platform in this embodiment, a modular hierarchical design architecture is adopted. The bottom layer of the signal conditioning module uses a multi-stage operational amplifier circuit to implement signal preprocessing. The first-stage operational amplifier is configured in a differential amplifier mode, and a high common-mode rejection ratio circuit is used to suppress the common-mode interference signals in the industrial field; the second-stage operational amplifier uses an inverting proportional amplification circuit to achieve preliminary signal amplification, and the amplification factor is accurately set through a feedback resistor network; the third-stage operational amplifier constructs a twin-T active filter circuit, and combined with the high input impedance characteristics of the operational amplifier, it effectively filters out power frequency interference and high-frequency noise.

[0056] In the signal conditioning circuit design of this embodiment, a multi-channel adaptive gain control mechanism is implemented. The gain amplification unit uses a high-precision programmable gain amplifier, and its gain coefficient can be adjusted in real time through the SPI bus. The gain range covers 1 - 1000 times, and the resolution reaches 0.1 dB. The signal amplitude is monitored in real time through a 16-bit precision ADC. Combining the proportional-integral control algorithm, when the signal amplitude is lower than 30% of the range, the gain coefficient is gradually increased; when the signal amplitude exceeds 70% of the range, the gain coefficient is dynamically decreased to ensure that the signal always operates within the best dynamic range.

[0057] In the implementation process of the digital-to-analog conversion unit in this embodiment, a high-performance pipelined ADC architecture is adopted. The sample-and-hold circuit uses a bootstrap switch technology to achieve signal sampling through capacitor charging and discharging, effectively suppressing the non-linearity of the switch-on resistance, and the sampling accuracy is better than 14 bits; the comparator uses a dynamic latch comparator structure, combined with a clock control circuit to achieve fast comparison, while reducing the static power consumption; the digital output circuit uses a low-voltage differential signal transmission method to improve signal integrity and enhance the anti-interference ability through differential pairs. The reference source circuit uses a temperature-compensated bandgap structure to provide a 2.5V high-precision reference voltage, and the temperature drift is better than 3 ppm / ℃.

[0058] In the design of the environmental parameter acquisition module of this embodiment, a multi-parameter collaborative acquisition mechanism is implemented. The temperature sensor uses a four-wire PT100 platinum resistor, which is combined with a high-precision constant current source excitation circuit to measure temperature through a Wheatstone bridge. The measurement range covers -50°C to 150°C; the humidity sensor uses a highly reliable polymer capacitive sensor, which converts the capacitance change into a frequency signal through a multivibrator to achieve accurate measurement of relative humidity; the air quality sensor uses an electrochemical principle, and the pollutant concentration is characterized by a microampere-level current signal, which is combined with a transimpedance amplifier to achieve signal conversion.

[0059] In the construction of the energy consumption acquisition module of this embodiment, a time-division multiplexing data acquisition strategy is adopted. The voltage sampling channel uses a high-precision resistor voltage division network, which is combined with an isolation amplifier to achieve electrical isolation; the current sampling channel uses a highly linear Hall sensor, and the measurement accuracy is improved through closed-loop compensation technology. The sampling clock is generated by a phase-locked loop to ensure the sampling synchronization of the voltage and current channels, providing an accurate basis for power calculation. The power calculation uses a digital multiplier to perform instantaneous power operation, and the average power is obtained through a sliding window average.

[0060] In the data transmission link of this embodiment, a multi-level cache architecture is adopted. The first-level cache uses a dual-port RAM to achieve real-time caching of sampling data; the second-level cache uses a large-capacity SDRAM to support long-term data storage; the third-level cache uses an industrial-grade SSD to achieve persistent data storage. The data transmission adopts the DMA method, and the real-time transfer of data is realized through an interrupt management mechanism. The bus protocol adopts the standard ModBus-RTU format, supporting multi-machine communication.

[0061] Through the above technical solutions, this embodiment realizes the highly reliable acquisition and processing of industrial field signals. This solution ensures the accurate acquisition of weak signals through multi-level signal conditioning and adaptive gain control; guarantees the accuracy of data conversion through high-performance ADC and isolation technology; adopts a multi-level cache mechanism to realize the reliable storage and transmission of data. In practical applications, this solution can adapt to complex industrial environments and provide an accurate data basis for equipment status monitoring.

[0062] Step S102: Input the monitoring data in the multi-level data cache into the time-series feature extraction model, perform principal component analysis on the extracted time-series features to obtain feature vectors, establish an equipment condition judgment model based on the feature vectors, adjust the sampling frequency according to the condition judgment model, input the adjusted sampling data into the data fusion model, input the fusion data into the collaborative diagnosis system composed of a rule model, a machine learning model, and a deep network model, embed the feature mapping library constructed by equipment component parameters, fault types, and fault modes into the collaborative diagnosis system, correct the diagnosis result based on the feature mapping library, and input the corrected diagnosis result into the decision-making system;

[0063] Optionally, in the process of extracting temporal features in this embodiment, a multi-scale analysis method is adopted. First, the monitoring data is segmented by time windows, and the window length is adaptively adjusted according to the device characteristics. A shorter window is used for parameters that change rapidly, and a longer window is used for parameters that change slowly. In each time window, statistical features (mean, standard deviation, skewness, kurtosis), time-domain features (maximum value, minimum value, peak factor, waveform factor), and frequency-domain features (center frequency, power spectral density, frequency band energy ratio) are extracted respectively.

[0064] In the feature dimensionality reduction link of this embodiment, an adaptive principal component analysis mechanism is implemented. The influence of dimensions is eliminated by feature standardization, the feature correlation coefficient matrix is calculated, and the principal components are obtained by the eigenvalue decomposition method. A cumulative contribution rate threshold is introduced to automatically determine the number of principal components, and the selected principal components are linearly combined to form a feature vector. The feature vector not only retains the main information of the original features but also significantly reduces the data dimension.

[0065] In the construction of the operating condition judgment model in this embodiment, an improved support vector machine algorithm is adopted. The radial basis function is selected as the kernel function, and its parameters are optimized and determined through grid search. The performance of the model is evaluated by cross-validation, and the penalty factor is dynamically adjusted. The input of the model is the dimensionality-reduced feature vector, and the output is the device operating condition category, including multiple states such as startup, stable operation, load fluctuation, and fault.

[0066] In the sampling frequency adjustment strategy of this embodiment, an operating condition adaptive mechanism is implemented. The mapping relationship between the operating condition category and the sampling frequency is established. The sampling frequency is increased in the startup and fault states of the device to capture transient features; the sampling frequency is appropriately reduced in the stable operation state to reduce data redundancy. The sampling frequency adjustment adopts a smooth transition strategy to avoid sampling mutations.

[0067] In the design of the data fusion model in this embodiment, a hierarchical fusion architecture is adopted. In the first layer, data-level fusion is performed, and the state estimation of multi-source data is realized through a Kalman filter; in the second layer, feature-level fusion is performed, and multi-dimensional features are combined using the Dempster-Shafer evidence theory; in the third layer, decision-level fusion is performed, and multiple judgment results are synthesized based on the fuzzy integral method.

[0068] In the construction of the collaborative diagnosis system in this embodiment, a multi-model collaboration mechanism is implemented. The rule model constructs a diagnostic rule base based on expert knowledge and uses a forward reasoning engine for fault reasoning; the machine learning model adopts an ensemble learning method to combine base classifiers such as random forests and gradient boosting trees; the deep network model adopts a long short-term memory network structure to capture the fault evolution process through temporal feature learning.

[0069] In the construction process of the feature mapping library in this embodiment, a hierarchical organizational structure is adopted. The device component parameters include geometric parameters, material parameters, operating parameters, etc., and are classified according to component functions; the fault types include mechanical faults, electrical faults, control faults, etc., and each type of fault is further divided into multiple specific fault modes; the fault mode description includes multiple dimensions such as fault characteristics, evolution laws, and influencing factors.

[0070] In the diagnosis result correction link of this embodiment, a confidence evaluation mechanism is realized. Based on the feature mapping library, a fault diagnosis scoring criterion is established, and confidence scores are calculated for the rule diagnosis result, the machine learning diagnosis result, and the deep learning diagnosis result respectively. Multiple diagnosis results are fused through a weighted voting method, and the weight coefficient is positively correlated with the confidence score.

[0071] In the design of the decision system interface in this embodiment, a standardized data format is adopted. The diagnosis results include key information such as fault types, fault degrees, and confidence levels, and are encapsulated in JSON format; the diagnosis process data includes intermediate data such as feature extraction results and model diagnosis results, and is stored in binary format; the system configuration information includes model parameters, threshold settings, etc., and is managed in XML format.

[0072] Through the above technical solutions, this embodiment realizes the intelligent diagnosis of the device state. This solution extracts the typical features of faults through multi-scale feature extraction and adaptive dimensionality reduction; improves the accuracy and reliability of diagnosis through multi-model collaborative diagnosis and feature mapping correction; adopts a working condition adaptive data acquisition strategy to optimize the system resource utilization efficiency. In practical applications, this solution can timely detect device anomalies and provide decision-making support for predictive maintenance.

[0073] Step S103: Input the output result of the decision system and the device historical data into the trend modeling unit, build a remaining life prediction model based on the trend modeling unit, construct a risk warning function in combination with fault evolution analysis, combine the risk warning function with the remaining life prediction model to establish a hierarchical warning system, generate a maintenance plan according to the hierarchical warning system, input the maintenance plan into the maintenance decision unit to generate a device maintenance plan, and allocate maintenance tasks through the remote communication module based on the device maintenance plan.

[0074] Optionally, in the construction process of the trend modeling unit in this embodiment, the state indicators output by the decision system are first subjected to time series decomposition. The wavelet decomposition method is used to decompose the state indicators into a trend term, a periodic term, and a random term. The trend term reflects the long-term change trend of device performance degradation, the periodic term reflects the periodic fluctuation characteristics of device operation, and the random term contains the influence of environmental interference and measurement noise.

[0075] In the implementation of the remaining useful life prediction model, a two-layer prediction architecture is adopted. The bottom layer uses an improved Weibull distribution model to describe the degradation process of the device. The model parameters are determined by the maximum likelihood estimation method, taking into account influencing factors such as the device's usage time, load level, and environmental conditions. The upper layer uses a long short-term memory network to achieve dynamic prediction. The network input includes the decomposition terms of the state indicators, the degradation model parameters, and the time scale information, and adaptively selects key temporal features through the attention mechanism.

[0076] In the process of fault evolution analysis in this embodiment, a multi-dimensional clustering mechanism is implemented. Spatiotemporal clustering analysis is performed on historical fault data to identify typical fault evolution patterns. The clustering features include the time features, spatial features, and correlation features of the fault occurrence. The density peak clustering algorithm is used to automatically determine the clustering centers, and the clustering effect is evaluated through the silhouette coefficient. Based on the clustering results, a fault evolution path diagram is constructed to describe the development laws of different fault types.

[0077] In the construction of the risk warning function in this embodiment, a multi-index fusion method is adopted. The mapping relationship between the state indicators and the fault risk levels is established, and the risk score is calculated through the fuzzy comprehensive evaluation method. The evaluation factors include the deviation degree, change rate, and fluctuation amplitude of the state indicators, and the weight coefficients of each factor are determined based on expert knowledge. The risk warning function adopts a piecewise linear structure, and different warning thresholds are set in different risk intervals.

[0078] In the process of establishing the hierarchical warning system in this embodiment, an adaptive threshold adjustment mechanism is implemented. The output result of the risk warning function is correlated with the remaining useful life prediction result to establish a two-dimensional warning matrix. The horizontal axis represents the risk level, the vertical axis represents the remaining useful life interval, and the matrix elements define the warning levels. The warning threshold is dynamically adjusted according to the device operation status to avoid false alarms and missed alarms.

[0079] In the link of generating the maintenance plan in this embodiment, an intelligent planning strategy is adopted. The corresponding maintenance strategy is selected based on the warning level. The strategy library includes various plans such as preventive maintenance, condition-based maintenance, and emergency maintenance. The generation of the maintenance plan takes into account multiple factors such as the device importance, maintenance resource constraints, and production plan, and determines the optimal maintenance time window through a multi-objective optimization algorithm.

[0080] In the design of the maintenance decision-making unit in this embodiment, a dynamic scheduling mechanism is implemented. The formulation of the maintenance plan adopts the rolling horizon optimization method to minimize the maintenance cost and downtime on the premise of ensuring the device reliability. The decision variables include the execution order, start and end times, and resource allocation of the maintenance items, and the genetic algorithm is used to solve the optimization problem.

[0081] In the implementation of the remote communication module in this embodiment, a distributed architecture is adopted. The distribution of maintenance tasks is based on the publish-subscribe mode, and the asynchronous transmission of task information is realized through a message queue. The task information is encapsulated in a standardized format, including key information such as maintenance objects, execution time, and required resources. The security of task distribution is ensured through identity authentication and permission management.

[0082] In the task execution monitoring link of this embodiment, a feedback closed-loop mechanism is established. The execution status of maintenance tasks is collected in real time, including information such as progress, quality, and resource consumption. The maintenance effect is evaluated through data analysis, the maintenance strategy library is dynamically updated, and the continuous optimization of the maintenance plan is realized. The maintenance records are automatically archived in the historical database to provide data support for subsequent analysis.

[0083] Through the above technical solutions, this embodiment realizes the intelligent management of equipment maintenance. This solution accurately evaluates the health status of equipment through multi-dimensional trend analysis and life prediction; realizes the optimal allocation of maintenance resources through a multi-level early warning mechanism and intelligent maintenance planning; adopts a distributed task management and closed-loop feedback mechanism to improve the execution efficiency of maintenance work. In practical applications, this solution can effectively reduce the equipment failure rate, extend the service life of the equipment, and improve the intelligent level of equipment management.

[0084] As can be seen from the above description, the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis provided by the embodiment of this application can realize high-precision data acquisition and processing through an edge computing platform integrating a signal conditioning, gain amplification, and analog-to-digital conversion unit. The system realizes adaptive sampling based on time-series feature extraction and working condition judgment models, and innovatively constructs a collaborative diagnosis system combining a rule model, machine learning, and a deep network. By embedding the equipment feature mapping library into the diagnosis system and combining trend modeling and remaining life prediction, a risk classification early warning system is established to realize intelligent decision-making and remote management of equipment maintenance. This method provides a comprehensive solution for the predictive maintenance of industrial equipment.

[0085] In an embodiment of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis of this application, refer to Figure 2 , it may specifically include the following content:

[0086] Step S201: Construct a heterogeneous computing architecture using a dual-core processor and a programmable gate array, configure the master-slave interface, storage interface, and communication interface of the heterogeneous computing architecture as data processing units based on the fieldbus protocol, connect the amplification circuit, filtering circuit, and isolation circuit of the signal conditioning module to the input end of the data processing unit, and connect the programmable amplifier and gain adjustment circuit of the gain amplification unit to the output end of the signal conditioning module;

[0087] Step S202: Connect the sample-and-hold circuit, reference source circuit, and digital quantity output circuit of the digital-to-analog conversion unit to the signal output end of the gain amplification unit. Configure the logic function module of the programmable gate array based on the hardware description language, and connect the temperature sensor and humidity sensor of the environmental parameter acquisition module and the current transformer and voltage sensor of the energy consumption acquisition module to the multi-channel sampling channels of the digital-to-analog conversion unit respectively.

[0088] Optionally, in the design of the heterogeneous computing architecture of this embodiment, a dual-core ARM Cortex-A processor is selected as the main controller, and an FPGA of the Xilinx series is used to build a heterogeneous computing platform. The dual-core processor adopts a master-slave architecture, where one core is responsible for system management and communication tasks, and the other core focuses on data processing and algorithm operations. The FPGA realizes high-speed data interaction with the processor through the AXI bus, and realizes data acquisition, signal preprocessing, and real-time control functions inside the FPGA.

[0089] In the implementation process of the data processing unit of this embodiment, a modular interface design is adopted. The master-slave interface adopts the RS485 bus protocol, supporting multi-point communication and long-distance data transmission; the storage interface includes an SDRAM controller and a Flash controller, which are used for real-time data caching and program storage respectively; the communication interface integrates an Ethernet MAC controller and a CAN bus controller to realize multi-protocol data transmission. Each interface module is flexibly connected through the crossbar switch matrix inside the FPGA.

[0090] In the design of the signal conditioning module of this embodiment, a multi-stage signal processing link is realized. The amplification circuit uses a low-noise operational amplifier to construct a differential input stage, and sets the initial gain through a precision resistor network; the filter circuit includes a cascaded structure of a high-pass filter and a low-pass filter, and the cut-off frequency can be adjusted through an RC network; the isolation circuit uses a digital isolator and a DC-DC isolated power supply to provide 3000V electrical isolation ability.

[0091] In the construction process of the gain amplification unit of this embodiment, a digitally controlled programmable gain amplifier is adopted. The gain adjustment circuit realizes real-time gain control through the SPI interface, the gain range covers 0dB to 60dB, and the step accuracy reaches 0.5dB. An overvoltage protection circuit and an offset calibration circuit are integrated to ensure the reliability of signal processing. The gain value is dynamically updated through a look-up table in the FPGA.

[0092] In the design of the digital-to-analog conversion unit of this embodiment, high-precision data acquisition is realized. The sample-and-hold circuit uses a rail-to-rail input operational amplifier to build a follower circuit to ensure sampling accuracy; the reference source circuit provides a 2.5V reference voltage based on the bandgap principle, and the temperature drift is controlled within 5ppm / °C; the digital quantity output circuit uses LVDS differential signal transmission to improve anti-interference ability.

[0093] In this embodiment, in the configuration of the FPGA logic function module, a modular HDL design method is adopted. The timing control module, data cache module, and interface protocol module are implemented through VHDL language. The timing control module generates precise sampling clocks and control signals; the data cache module implements multi-level FIFO caching and supports data pipelining; the interface protocol module processes data interaction with external devices.

[0094] In the implementation of the environmental parameter acquisition module of this embodiment, a multi-channel parallel acquisition mechanism is established. The temperature sensor uses a PT100 platinum resistance, and a four-wire measurement circuit is used to eliminate the influence of lead impedance; the humidity sensor selects a high-precision capacitive sensor, which is converted into a voltage signal through an operational amplifier integration circuit. The sensor signals are connected to the analog input channels of the ADC after signal conditioning.

[0095] In the design of the energy consumption acquisition module of this embodiment, synchronous sampling of voltage and current is realized. The current transformer uses an open-loop Hall sensor with a measurement range covering 0 - 100A, and a precision operational amplifier is used to realize the conversion of current signals; the voltage sensor uses a high-precision resistor voltage division network, and electrical isolation is realized through an isolation amplifier. The gain and offset of the sampling channels can be dynamically adjusted through the FPGA.

[0096] In this embodiment, the time-division multiplexing of sensor signals is realized through a multiplexer, and the sampling timing is strictly synchronized by the control signals generated by the FPGA. A digital filtering algorithm is implemented inside the FPGA to perform real-time processing on the sampled data and filter out power frequency interference and high-frequency noise. The processed data is transmitted to the processor through the DMA method to achieve efficient data acquisition.

[0097] Through the above technical solutions, this embodiment realizes the highly reliable acquisition and processing of industrial field signals. This solution improves the system processing ability through a heterogeneous computing architecture, ensures the acquisition accuracy through multi-level signal conditioning, and improves the system scalability through modular design. In practical applications, this solution can adapt to complex industrial environments and provide a reliable data basis for equipment status monitoring.

[0098] In an embodiment of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis in this application, refer to Figure 3 , and it may specifically include the following content:

[0099] Step S301: Input the acquired signal into a passive filter circuit in the signal conditioning module for band-pass filtering, suppress the common mode of the filtered signal through a differential amplifier, amplify the suppressed signal through an operational amplifier, electrically isolate the amplified signal through an opto-isolator, and perform secondary amplification on the isolated signal based on a programmable gain amplifier. Set the gain coefficient of the programmable gain amplifier as an adjustable parameter, and dynamically adjust the gain coefficient according to the signal amplitude.

[0100] Step S302: Sample the conditioned signal using the sample-and-hold circuit in the analog-to-digital conversion unit, compare the sampled signal with the reference voltage generated by the reference source circuit to obtain a digital quantity, transmit the digital quantity to the field-programmable gate array buffer through a serial peripheral interface, and transmit the data in the field-programmable gate array buffer to a multi-level data cache composed of a dynamic random access memory and a solid-state memory through a system bus.

[0101] Optionally, in the design of the passive filter circuit of the signal conditioning module in this embodiment, a multi-stage RC network is used to implement the band-pass filtering function. The low-frequency cut-off part adopts a high-pass filter structure, which is composed of a precision thin-film capacitor and a metal-film resistor, and the cut-off frequency is designed at 0.1 Hz to effectively suppress the DC offset and low-frequency drift; the high-frequency cut-off part adopts a low-pass filter structure, and the cut-off frequency is set at 1 kHz to suppress high-frequency interference and noise. The filter circuit adopts a symmetric design to ensure the consistency of the phase characteristics of the signal channel.

[0102] In the design of the differential amplifier in this embodiment, a high-precision instrumentation amplifier is used to construct the differential input stage. The common-mode rejection ratio is set through a precision resistor network, and the common-mode rejection ratio is better than 100 dB in the power frequency range. The differential amplifier is powered by a dual power supply, and an over-voltage protection circuit is designed at the input stage to prevent damage to the chip caused by large-signal impact. The common-mode voltage range covers ±10 V, meeting the requirements for common-mode interference suppression in industrial sites.

[0103] In the design of the operational amplifier stage in this embodiment, the function of low-noise amplification is realized. A rail-to-rail input / output operational amplifier is used, and a fixed gain is set through a negative feedback network. The bandwidth of the operational amplifier meets the signal bandwidth requirements. An offset calibration circuit is integrated, and the DC offset is eliminated by adjusting an external potentiometer to ensure the accuracy of the amplified signal.

[0104] In the application of the opto-isolator in this embodiment, a high-speed digital isolator is used to achieve electrical isolation. A linear optocoupler is used at the input end of the isolator to ensure the linear transmission of analog signals; a differential output structure is used at the output end to improve the anti-interference ability. The bandwidth of the isolator covers the signal bandwidth to ensure the minimum signal distortion. A DC-DC isolation module is used at the power supply end to provide an isolated power supply.

[0105] In this embodiment, adaptive gain control is achieved in the design of the programmable gain amplifier. A digitally controlled programmable gain amplifier is used, and the gain range is divided into multiple gears. The gain control is realized through the SPI interface. The gain adjustment adopts a look-up table method, and the optimal gain coefficient is selected according to the signal amplitude. A soft switching strategy is adopted during gain switching to avoid signal jumps caused by sudden gain changes.

[0106] In this embodiment, in the implementation of the sample and hold circuit, a sampling circuit is constructed using a high-speed sampling switch and a hold capacitor. The sampling switch adopts a CMOS transmission gate structure to reduce the on-resistance of the switch; the hold capacitor adopts a high-quality polypropylene capacitor to reduce the dielectric absorption effect. The sampling clock is generated by the FPGA, and the sampling period can be dynamically adjusted according to the signal bandwidth.

[0107] In this embodiment, in the design of the reference source circuit, a high-precision bandgap reference source is adopted. The temperature coefficient of the reference voltage is optimized through curvature compensation technology, and the power supply rejection ratio is improved through multi-stage filtering. The reference source circuit integrates a buffer amplifier to provide a reference voltage with a low output impedance and ensure the load capacity.

[0108] In this embodiment, in the digital quantity transmission link, a differential signal transmission mechanism is adopted. The serial peripheral interface adopts the SPI protocol, the clock frequency is adjustable, and multi-device cascading is supported. The data frame format includes a synchronization header, data bits, and a check bit, and the data integrity is ensured through CRC check. The transmission line adopts differential pair wiring to improve the anti-interference ability.

[0109] In this embodiment, in the design of the FPGA buffer, a multi-level buffer architecture is realized. The first-level buffer adopts distributed RAM to achieve real-time data caching; the second-level buffer adopts block RAM to support batch data transmission. The buffer control is realized by a state machine, which supports data flow control and exception handling.

[0110] In this embodiment, in the implementation of the multi-level data cache, a hierarchical storage structure is adopted. Dynamic random access memory is used for temporary data storage, supporting high-speed read and write access; solid-state memory is used for data persistent storage, and the wear leveling algorithm is adopted to extend the service life. The data transmission adopts the DMA method, and the data is transported in real time through interrupt management.

[0111] Through the above technical solutions, this embodiment realizes the high-precision conditioning and acquisition of industrial field signals. This solution improves the signal-to-noise ratio of the signal through multi-stage filtering and amplification; enhances the anti-interference ability through electrical isolation and differential transmission; and adopts a multi-level cache mechanism to ensure the reliable storage of data. In practical applications, this solution can effectively cope with complex industrial environments and provide high-quality data support for equipment status monitoring.

[0112] In an embodiment of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis in this application, seeFigure 4 , it may specifically include the following content:

[0113] Step S401: Divide the monitoring data in the multi-level data cache into multiple data windows according to the time series, extract statistical features, frequency-domain features and time-frequency features for the data windows respectively, form a feature matrix by combining the mean, variance, skewness, kurtosis of the statistical features, the frequency center, frequency variance, power spectral density of the frequency-domain features, and the wavelet coefficients, energy entropy of the time-frequency features, perform standardization processing on the feature matrix, input the standardized feature matrix into the principal component analysis model, calculate the principal component scores based on eigenvalue decomposition, and construct a feature vector with the principal component scores;

[0114] Step S402: Construct a working condition judgment model based on the support vector machine, input the feature vector into the working condition judgment model for classification, establish a corresponding relationship between the classification result and a preset sampling frequency threshold, select the sampling frequency parameter corresponding to the current working condition according to the corresponding relationship, input the data collected at the sampling frequency parameter into the data fusion model, calculate the weight coefficients of different data sources based on the attention mechanism in the data fusion model, and perform weighted fusion on the multi-dimensional features according to the weight coefficients.

[0115] Optionally, in the process of data window division in this embodiment, an adaptive window mechanism is adopted. Set the basic window length according to the operating characteristics of the device, use a shorter window for rapidly changing parameters, and use a longer window for slowly changing parameters. Set a 50% overlap rate between windows, which not only ensures the continuity of feature extraction but also avoids omission of important features. The window length is automatically calculated by the characteristic frequency of the device to ensure capturing the complete feature cycle.

[0116] In the statistical feature extraction link of this embodiment, multi-dimensional feature calculation is realized. The mean reflects the overall level of the parameter, the variance characterizes the degree of dispersion of the data, the skewness describes the asymmetry of the data distribution, and the kurtosis characterizes the steepness of the data distribution. Through the combination of these statistics, the statistical characteristics of the signal are comprehensively characterized. The feature calculation adopts a recursive algorithm, which improves the calculation efficiency.

[0117] In the process of frequency-domain feature extraction in this embodiment, an improved fast Fourier transform algorithm is adopted. The frequency center reflects the main frequency components of the signal, the frequency variance describes the degree of concentration of the frequency distribution, and the power spectral density characterizes the energy distribution of each frequency component. Windowing is performed through a Hanning window to reduce spectral leakage. The spectral analysis adopts a segmented averaging method to improve the reliability of the spectral estimation.

[0118] In this embodiment, a multi-scale analysis mechanism is implemented in time-frequency feature extraction. Wavelet transform is used for time-frequency decomposition, and db4 wavelet is selected as the basis function for 5-layer decomposition. Wavelet coefficients reflect the time-varying characteristics of the signal in different frequency bands, and energy entropy describes the complexity of the signal. The high-frequency components are analyzed in detail through wavelet packet decomposition to improve the resolution of time-frequency features.

[0119] In this embodiment, an adaptive normalization method is adopted in the feature matrix normalization process. Features with different dimensions are normalized and transformed to eliminate the influence of dimensions. The normalization parameters are dynamically updated through a sliding window to adapt to the non-stationary characteristics of the signal. Outliers are identified and processed by the 3σ criterion to ensure the reliability of the normalization results.

[0120] In this embodiment, an adaptive dimensionality reduction mechanism is implemented in the construction of the principal component analysis model. Eigenvectors and eigenvalues are calculated through eigenvalue decomposition, and the number of principal components is automatically determined according to the cumulative contribution rate. The principal component scores are calculated by projecting onto the eigenvectors, retaining the main information of the original features. The dimensionality reduction process takes into account the correlation between features and eliminates data redundancy.

[0121] In this embodiment, an improved support vector machine algorithm is adopted in the design of the working condition judgment model. The radial basis function is selected as the kernel function, and the kernel parameters are optimized through cross-validation. The penalty factor and kernel function parameters are determined by the grid search method to improve the classification accuracy. The model is trained using the online learning method, supporting incremental updates.

[0122] In this embodiment, a working condition adaptive mechanism is implemented in the sampling frequency adjustment strategy. A mapping relationship between working condition categories and sampling frequencies is established. The sampling frequency is increased during equipment startup and fault states to capture transient features; the sampling frequency is appropriately reduced during stable operation to reduce data redundancy. The sampling frequency adjustment adopts a smooth transition strategy to avoid sampling mutations.

[0123] In this embodiment, a multi-level fusion architecture is adopted in the construction of the data fusion model. The attention mechanism determines the weight coefficients by calculating the feature correlation degree, and the weight calculation takes into account the importance and temporal dependence of the features. The fusion process is divided into feature-level fusion and decision-level fusion. The feature-level fusion adopts the weighted average method, and the decision-level fusion adopts the voting mechanism.

[0124] In this embodiment, a dynamic weight adjustment mechanism is implemented in the feature weighted fusion link. The weight coefficients are optimized through the backpropagation algorithm, and the loss function includes a reconstruction error term and a sparsity constraint term. The fusion results are evaluated through residual analysis, and the weight parameters are dynamically adjusted. The fusion process takes into account the data reliability and suppresses abnormal data sources.

[0125] Through the above technical solutions, this embodiment realizes the intelligent recognition and feature extraction of the device state. This solution obtains the typical features of the state through multi-dimensional feature extraction and adaptive dimensionality reduction; improves the accuracy of state recognition through working condition adaptive sampling and multi-source data fusion; and enhances the adaptability of the system by adopting a dynamic weight adjustment mechanism. In practical applications, this solution can accurately identify the device working conditions and provide a reliable data basis for state monitoring and fault diagnosis.

[0126] In an embodiment of the intelligent decision-making management method for energy-saving device state monitoring and fault diagnosis in this application, refer to Figure 5 , and it may specifically include the following content:

[0127] Step S501: Construct a rule model by dividing the fault diagnosis expert knowledge according to the device components, construct a machine learning model based on the gradient boosting tree algorithm, construct a deep network model using a recurrent neural network, form a collaborative diagnosis system with the rule model, the machine learning model, and the deep network model, input the fusion data into the rule model for rule reasoning, input the machine learning model for classification prediction, input the deep network model for sequence analysis, construct a feature mapping library with the structural parameters, operating parameters, and fault characteristics of the device components, and establish a diagnostic confidence scoring system based on the corresponding rules of fault types and fault modes in the feature mapping library;

[0128] Step S502: Input the diagnostic results of the rule model, the machine learning model, and the deep network model into the diagnostic confidence scoring system for scoring, quantify the reliability of the diagnostic results of each model according to the scoring results, perform weighted combination on the quantified diagnostic results based on the Bayesian inference method to obtain a collaborative diagnosis result, match and correct the collaborative diagnosis result with the fault characteristics in the feature mapping library, and input the corrected diagnostic result into the decision-making system.

[0129] Optionally, in the process of constructing the rule model in this embodiment, a hierarchical knowledge representation method is adopted. An expert knowledge base is established for different device components, including the typical fault characteristics and diagnostic rules of key components such as bearings, gears, and motors. The rules are described in an "if-then" structure, where the antecedent contains the logical combination of fault characteristics, and the consequent gives the fault type and severity. The rule base is organized by a decision tree to improve the rule retrieval efficiency.

[0130] In the design of the machine learning model in this embodiment, an ensemble learning mechanism is realized. The gradient boosting tree algorithm uses a serial generation method of multiple decision trees, and each tree is trained for the residuals of the previous tree. The feature selection process considers information gain and feature importance, and determines the number and depth of the trees through cross-validation. The model training adopts a batch learning strategy and supports incremental updates.

[0131] In the construction of the deep network model in this embodiment, a long short-term memory network structure is adopted. The input layer receives the time series feature sequence, and controls the information flow through the input gate, forget gate, and output gate. The hidden layer adopts a multi-layer stacked structure to enhance the feature extraction ability of the model. The network training uses the backpropagation algorithm, and the loss function includes a classification error term and a regularization term.

[0132] In the implementation of the collaborative diagnosis system in this embodiment, a multi-model collaboration mechanism is established. The rule model is responsible for the rapid diagnosis of known faults, the machine learning model processes the pattern recognition of complex faults, and the deep network model captures the time series evolution characteristics of faults. The three models run in parallel, and the result fusion is achieved through subsequent confidence evaluation.

[0133] In the process of constructing the feature mapping library in this embodiment, multi-dimensional feature association is realized. The structural parameters include static features such as geometric dimensions and material properties, the operating parameters include dynamic features such as rotational speed and load, and the fault features include monitoring indicators such as vibration features and temperature features. The mapping relationship between features is established through association rule mining to form a fault diagnosis knowledge graph.

[0134] In the design of the diagnostic confidence scoring system in this embodiment, a multi-level evaluation method is adopted. The scoring criteria are established based on the integrity, consistency, and similarity of the fault features. The integrity evaluates the feature coverage, the consistency evaluates the stability of the diagnostic results, and the similarity evaluates the matching degree with historical cases. The scoring results are calculated through the fuzzy comprehensive evaluation method.

[0135] In the process of quantifying the reliability of the diagnostic results in this embodiment, an uncertainty evaluation mechanism is realized. The reliability of the rule model is calculated through the rule matching degree, the reliability of the machine learning model is evaluated through the probability output, and the reliability of the deep network model is estimated through the prediction variance. The reliability indicators are uniformly quantified through normalization processing.

[0136] In the Bayesian inference process in this embodiment, a dynamic weight update mechanism is established. The prior probability is established based on the historical diagnostic accuracy of each model, and the posterior probability is calculated through the Bayesian formula. The weight coefficient is positively correlated with the posterior probability, realizing the adaptive fusion of the diagnostic results.

[0137] In the diagnostic result correction link in this embodiment, a knowledge-driven optimization mechanism is realized. The collaborative diagnostic results are matched with the fault features in the feature mapping library for similarity, and the most similar fault mode is identified. The diagnostic results are corrected through expert rules to improve the diagnostic accuracy. The correction process takes into account the influence of the equipment operating conditions and environmental factors.

[0138] In this embodiment, a standardized data format is adopted in the design of the decision system interface. The diagnostic results include information such as fault type, fault location, fault severity, and reliability, and are encapsulated in a structured format. The diagnostic process data includes intermediate data such as feature extraction results and model diagnostic results, supporting the traceability of the diagnostic process.

[0139] Through the above technical solution, this embodiment realizes the intelligent diagnosis of equipment faults. This solution improves the accuracy of fault diagnosis through multi-model collaboration and knowledge-driven; enhances the credibility of diagnostic results through reliability assessment and result correction; and improves the scalability of the system by adopting a standardized interface design. In practical applications, this solution can accurately identify various faults and provide reliable decision-making support for equipment maintenance.

[0140] In an embodiment of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis in this application, see Figure 6 , it may specifically include the following content:

[0141] Step S601: Input the equipment status indicators output by the decision system, the operation records, maintenance records, and fault records in the equipment historical database into the trend modeling unit. In the trend modeling unit, perform time series decomposition on the status indicators to extract the trend term, periodic term, and random term, establish an equipment degradation model based on the Weibull distribution, combine the equipment degradation model with a long short-term memory network to construct a remaining life prediction model, and use the status indicator decomposition terms, degradation parameters, and time scale as input variables in the remaining life prediction model;

[0142] Step S602: Perform clustering analysis on the equipment fault data to identify the fault evolution path, establish a corresponding relationship between the fault evolution path and the status indicator threshold to construct a risk warning function, map the risk level calculated by the risk warning function to the life interval predicted by the remaining life prediction model, divide the warning level into multiple levels based on the mapping relationship, and set trigger conditions and warning rules for different warning levels respectively to construct a hierarchical warning system.

[0143] Optionally, in the time series decomposition process of the trend modeling unit in this embodiment, a multi-scale analysis method is adopted. The status indicators are decomposed into different frequency components through wavelet decomposition. The trend term reflects the long-term change trend, the periodic term reflects the periodic characteristics of equipment operation, and the random term includes short-term fluctuations and noise. The decomposition process uses an adaptive threshold for denoising to improve the accuracy of trend extraction.

[0144] In the construction of the equipment degradation model in this embodiment, a parameter adaptive estimation mechanism is realized. The shape parameter and scale parameter of the Weibull distribution model are determined by the maximum likelihood estimation method, taking into account the influence of equipment operation time, load level and environmental conditions. The model parameters are dynamically updated in a sliding window manner to adapt to the time-varying characteristics of equipment degradation.

[0145] In the design of the remaining useful life prediction model in this embodiment, a two-layer prediction architecture is adopted. The underlying Weibull model describes the basic degradation law of the equipment, and the parameter estimation takes into account the statistical characteristics of historical data; the upper-layer long short-term memory network captures complex non-linear features, and the network structure contains multiple memory units, and the information flow is controlled through a gating mechanism.

[0146] In the process of input variable processing in this embodiment, a feature fusion mechanism is realized. The decomposed items of the state index extract time domain features and frequency domain features through feature engineering. The degradation parameters include the key parameters of the Weibull model and the degradation rate, and the time scale information includes the historical operation time and the prediction time window. The correlation between features is modeled through an attention mechanism.

[0147] In the identification of the fault evolution path in this embodiment, a multi-dimensional clustering strategy is established. The density peak clustering algorithm is used to perform clustering analysis on the fault data, and the clustering features include fault symptoms, evolution speed and propagation path. The clustering effect is evaluated by the silhouette coefficient, and the optimal number of clusters is automatically determined. The clustering results form a fault evolution map.

[0148] In the construction process of the risk warning function in this embodiment, a multi-index fusion method is adopted. The state index threshold is set based on the fault evolution path, and the threshold setting takes into account the equipment characteristics and operation requirements. The risk level is calculated through fuzzy comprehensive evaluation, and the evaluation factors include the deviation degree, change rate and fluctuation amplitude of the state index.

[0149] In the warning level mapping link of this embodiment, a two-dimensional evaluation mechanism is realized. A two-dimensional evaluation matrix of risk level and remaining life interval is established, with the horizontal axis representing the risk level and the vertical axis representing the remaining life interval. The matrix elements define the warning levels, and the mapping relationship is determined through expert knowledge rules. The warning levels are dynamically adjusted according to the equipment status.

[0150] In the design of the hierarchical warning system in this embodiment, a hierarchical warning architecture is adopted. The warnings are divided into four levels: attention, warning, alarm and emergency, and different trigger conditions and warning rules are set for each level. The trigger conditions include single-index thresholds and multi-index combination conditions, and the warning rules define the generation and transmission methods of warning information.

[0151] In this embodiment, an adaptive threshold mechanism is implemented in the trigger condition setting. The basic threshold is determined based on the statistical characteristics of the device operation data, and the threshold level is corrected by a dynamic adjustment coefficient. The threshold adjustment takes into account the influence of the device operation conditions and environmental factors to avoid false alarms and missed alarms.

[0152] In the process of formulating the warning rules in this embodiment, a multi-level response mechanism is established. Different warning levels correspond to different processing procedures and response measures, including adjustment of the status monitoring frequency, formulation of maintenance plans, and emergency disposal, etc. The warning information is transmitted through a standardized interface to support the collaborative response of multiple systems.

[0153] Through the above technical solutions, this embodiment realizes the intelligent warning of the device status. This solution accurately evaluates the device health status through multi-dimensional trend analysis and life prediction; realizes the early identification of risks through fault evolution analysis and hierarchical warning; and adopts an adaptive warning mechanism to improve the accuracy of warning. In practical applications, this solution can timely detect potential faults and provide predictive decision-making support for device maintenance.

[0154] In an embodiment of the intelligent decision-making management method for energy-saving device status monitoring and fault diagnosis in this application, refer to Figure 7 , and it may specifically include the following content:

[0155] Step S701: Associate and match the warning levels in the hierarchical warning system with the maintenance strategy library, construct a maintenance plan generation model based on the maintenance procedures, maintenance cycles, and spare part information in the maintenance strategy library, use the device status indicators, warning levels, and maintenance resource constraints as input parameters in the maintenance plan generation model to generate a specific maintenance plan, input the maintenance plan into the maintenance decision-making unit, and in the maintenance decision-making unit, weigh the maintenance cost, maintenance time, and device reliability based on a multi-objective optimization algorithm, and prioritize and schedule the device maintenance items;

[0156] Step S702: Assign tasks to the maintenance plan output by the maintenance decision-making unit through the remote communication module. In the remote communication module, encode the maintenance task information according to a preset format, establish a communication link for the maintenance task based on the distributed network protocol, and send the encoded maintenance task to the corresponding execution terminal through the communication link. In the execution terminal, decode and classify and store the received maintenance task.

[0157] Optionally, in the process of constructing the maintenance strategy library in this embodiment, a hierarchical knowledge organization method is adopted. The maintenance procedures include standard operation processes, technical requirements, and quality standards. The maintenance cycles are formulated based on the device characteristics and operation conditions. The spare part information includes inventory levels, procurement cycles, and replacement strategies. The strategy library is stored and retrieved through a relational database to support the dynamic update of the strategies.

[0158] In the design of the maintenance plan generation model in this embodiment, a knowledge-driven reasoning mechanism is implemented. The input parameters are preprocessed through feature engineering. The device status indicators adopt a multi-dimensional evaluation method. A mapping relationship is established between the warning levels and the maintenance rules. The maintenance resource constraints include elements such as personnel, tools, and spare parts. The model generates maintenance plans based on a method that combines rule-based reasoning and case-based reasoning.

[0159] In the construction of the maintenance decision-making unit in this embodiment, a multi-level decision-making architecture is adopted. The multi-objective optimization algorithm uses an improved NSGA-II algorithm. The objective functions include minimizing the maintenance cost, optimizing the maintenance time, and maximizing the device reliability. The trade-off relationship between different objectives is represented by the Pareto optimal solution set, and the decision maker can select the most suitable solution according to the actual needs.

[0160] In the process of prioritizing maintenance projects in this embodiment, a comprehensive evaluation system is established. The evaluation indicators include the device importance, the fault risk level, the maintenance urgency, and the resource availability. The weights of the indicators are determined by the analytic hierarchy process, and the weighted scoring method is used to calculate the project priorities. The sorting results consider the dependencies between projects.

[0161] In the optimization of the time arrangement in this embodiment, a dynamic scheduling mechanism is implemented. A scheduling model is constructed based on the project priorities and the resource constraints, and a genetic algorithm is used to solve the optimal scheduling plan. The scheduling process considers minimizing the device downtime and maximizing the utilization rate of maintenance resources, and supports real-time adjustment and dynamic optimization.

[0162] In the design of the remote communication module in this embodiment, a hierarchical communication architecture is adopted. The maintenance task information is encoded in XML format. The data structure includes task identification, execution requirements, resource requirements, and time nodes. The data encryption technology is used in the encoding process to ensure the security of information transmission.

[0163] In the implementation of the distributed network protocol in this embodiment, a reliable transmission mechanism is established. A connection-oriented communication mode is adopted, and a communication link is established through a handshake protocol. The protocol supports data fragmentation and retransmission mechanisms to ensure the reliable transmission of large-scale maintenance task information. The network topology adopts a star structure to improve the communication efficiency.

[0164] In the communication link management in this embodiment, a dynamic routing strategy is adopted. The route is selected based on the link quality and the network load, and the link backup and switching mechanisms are supported. The bandwidth adaptive technology is used in the communication process to optimize the data transmission performance. The link status is monitored in real time through a heartbeat mechanism.

[0165] In the design of the execution terminal of this embodiment, an intelligent parsing mechanism is implemented. The received maintenance tasks are decoded by an XML parser, and the parsing results are subjected to format verification and integrity check. The task information is classified according to type and priority, and the storage uses a distributed database, supporting local caching and synchronous updates.

[0166] In the process of task classification and storage of this embodiment, a multi-dimensional index structure is established. The storage model includes task attribute indexes and time indexes, supporting multi-condition queries and statistical analysis. The storage system adopts a distributed architecture, providing data backup and disaster recovery mechanisms. The task execution status is managed through a state machine.

[0167] Through the above technical solutions, this embodiment realizes the intelligent management of equipment maintenance. This solution generates a scientific and reasonable maintenance plan through knowledge-driven and multi-objective optimization; ensures the efficient execution of maintenance tasks through distributed communication and intelligent terminals; and improves the collaborative efficiency of maintenance work by adopting a dynamic management mechanism. In practical applications, this solution can achieve the optimal allocation of maintenance resources and improve the efficiency and quality of equipment maintenance.

[0168] In order to be able to establish a risk grading and early warning system, realize the intelligent decision-making and remote management of equipment maintenance, and provide a comprehensive solution for the predictive maintenance of industrial equipment, this application provides an embodiment of an intelligent decision-making management device for realizing all or part of the content of the intelligent decision-making management method for the status monitoring and fault diagnosis of the energy-saving equipment, see Figure 8 , the intelligent decision-making management device for the status monitoring and fault diagnosis of the energy-saving equipment specifically includes the following:

[0169] The platform computing module 10 is used to integrate the signal conditioning module, the gain amplification unit, the analog-to-digital conversion unit, and the arithmetic processor into the edge computing platform, connect the environmental parameter acquisition module and the energy consumption acquisition module to the data acquisition channel of the analog-to-digital conversion unit, filter and amplify the acquired signal based on the signal conditioning module, adjust the gain coefficient of the acquisition channel through the gain amplification unit, convert the conditioned signal into a digital quantity by using the analog-to-digital conversion unit, and transmit the digital quantity to the multi-level data cache through the bus;

[0170] The model construction module 20 is configured to input the monitoring data in the multi-level data cache into a time series feature extraction model, perform principal component analysis on the extracted time series features to obtain feature vectors, establish an equipment condition judgment model based on the feature vectors, adjust the sampling frequency according to the condition judgment model, input the adjusted sampling data into a data fusion model, input the fusion data into a collaborative diagnosis system composed of a rule model, a machine learning model, and a deep network model, embed a feature mapping library constructed by equipment component parameters, fault types, and fault modes into the collaborative diagnosis system, correct the diagnosis result based on the feature mapping library, and input the corrected diagnosis result into a decision-making system;

[0171] The monitoring and diagnosis module 30 is configured to input the output result of the decision-making system and the equipment historical data into a trend modeling unit, construct a remaining life prediction model based on the trend modeling unit, construct a risk warning function in combination with fault evolution analysis, combine the risk warning function with the remaining life prediction model to establish a hierarchical warning system, generate a maintenance plan according to the hierarchical warning system, input the maintenance plan into a maintenance decision-making unit to generate an equipment maintenance plan, and allocate maintenance tasks based on the equipment maintenance plan through a remote communication module.

[0172] As can be seen from the above description, the intelligent decision-making management device for energy-saving equipment status monitoring and fault diagnosis provided by the embodiments of the present application can achieve high-precision data acquisition and processing through an edge computing platform integrating a signal conditioning, gain amplification, and analog-to-digital conversion unit. The system realizes adaptive sampling based on time series feature extraction and a condition judgment model, and innovatively constructs a collaborative diagnosis system combining a rule model, machine learning, and a deep network. By embedding an equipment feature mapping library into the diagnosis system and combining trend modeling and remaining life prediction, a risk grading warning system is established to realize intelligent decision-making and remote management of equipment maintenance. This method provides a comprehensive solution for the predictive maintenance of industrial equipment.

[0173] From a hardware perspective, in order to be able to establish a risk grading warning system, realize intelligent decision-making and remote management of equipment maintenance, and provide a comprehensive solution for the predictive maintenance of industrial equipment, the embodiments of the present application provide an electronic device for implementing all or part of the content in the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis. The electronic device specifically includes the following contents:

[0174] A processor, a memory, a communications interface, and a bus; wherein, the processor, the memory, and the communications interface complete communication with each other through the bus; the communications interface is used to implement information transmission between the intelligent decision-making management device for energy-saving device status monitoring and fault diagnosis and related devices such as the core business system, the user terminal, and the relevant database, etc.; the logic controller may be a desktop computer, a tablet computer, a mobile terminal, etc., and this embodiment is not limited thereto. In this embodiment, the logic controller may be implemented with reference to the embodiments of the intelligent decision-making management method for energy-saving device status monitoring and fault diagnosis in the embodiments, as well as the embodiments of the intelligent decision-making management device for energy-saving device status monitoring and fault diagnosis. The content is incorporated herein, and the repeated parts will not be elaborated again.

[0175] It can be understood that the user terminal may include a smart phone, a tablet electronic device, a network set-top box, a portable computer, a desktop computer, a personal digital assistant (PDA), a vehicle-mounted device, a smart wearable device, etc. Among them, the smart wearable device may include smart glasses, a smart watch, a smart bracelet, etc.

[0176] In practical applications, part of the intelligent decision-making management method for energy-saving device status monitoring and fault diagnosis may be executed on the electronic device side as described above, or all operations may be completed in the client device. Specifically, it can be selected according to the processing capacity of the client device and the limitations of the user usage scenario, etc. This application does not make any limitations in this regard. If all operations are completed in the client device, the client device may further include a processor.

[0177] The above-mentioned client device may have a communication module (i.e., a communication unit), and may be communicatively connected to a remote server to implement data transmission with the server. The server may include a server on the task scheduling center side, and may also include a server on an intermediate platform in other implementation scenarios, such as a server on a third-party server platform communicatively linked to the task scheduling center server. The server may include a single computer device, or may include a server cluster composed of multiple servers, or a server structure of a distributed device.

[0178] Figure 9 It is a schematic block diagram of the system composition of the electronic device 9600 according to an embodiment of the present application. As Figure 9 shown, the electronic device 9600 may include a central processing unit 9100 and a memory 9140; the memory 9140 is coupled to the central processing unit 9100. It should be noted that this Figure 9is exemplary; other types of structures can also be used to supplement or replace this structure to achieve telecommunication functions or other functions.

[0179] In one embodiment, the function of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis can be integrated into the central processing unit 9100. Among them, the central processing unit 9100 can be configured to perform the following controls:

[0180] Step S101: Integrate the signal conditioning module, the gain amplification unit, the analog-to-digital conversion unit and the arithmetic processor into the edge computing platform, connect the environmental parameter acquisition module and the energy consumption acquisition module to the data acquisition channel of the analog-to-digital conversion unit, filter and amplify the acquired signal based on the signal conditioning module, adjust the gain coefficient of the acquisition channel through the gain amplification unit, use the analog-to-digital conversion unit to convert the conditioned signal into a digital quantity, and transmit the digital quantity to the multi-level data cache through the bus;

[0181] Step S102: Input the monitoring data in the multi-level data cache into the time series feature extraction model, perform principal component analysis on the extracted time series features to obtain feature vectors, establish a device condition judgment model based on the feature vectors, adjust the sampling frequency according to the condition judgment model, input the adjusted sampling data into the data fusion model, input the fusion data into the collaborative diagnosis system composed of the rule model, the machine learning model and the deep network model, embed the feature mapping library constructed by the device component parameters, fault types and fault modes into the collaborative diagnosis system, correct the diagnosis result based on the feature mapping library, and input the corrected diagnosis result into the decision-making system;

[0182] Step S103: Input the output result of the decision-making system and the device historical data into the trend modeling unit, construct a remaining life prediction model based on the trend modeling unit, combine the fault evolution analysis to construct a risk warning function, combine the risk warning function with the remaining life prediction model to establish a hierarchical warning system, generate a maintenance plan according to the hierarchical warning system, input the maintenance plan into the maintenance decision-making unit to generate a device maintenance plan, and allocate the maintenance tasks through the remote communication module based on the device maintenance plan.

[0183] As can be seen from the above description, the electronic device provided by the embodiments of the present application realizes high-precision data acquisition and processing through an edge computing platform that integrates signal conditioning, gain amplification, and analog-to-digital conversion units. The system realizes adaptive sampling based on timing feature extraction and working condition judgment models, and innovatively constructs a collaborative diagnosis system that combines a rule model, machine learning, and a deep network. By embedding the device feature mapping library into the diagnosis system and combining trend modeling and remaining life prediction, a risk classification and early warning system is established to realize intelligent decision-making and remote management of equipment maintenance. This method provides a comprehensive solution for the predictive maintenance of industrial equipment.

[0184] In another embodiment, the intelligent decision-making management device for energy-saving equipment status monitoring and fault diagnosis can be separately configured from the central processing unit 9100. For example, the intelligent decision-making management device for energy-saving equipment status monitoring and fault diagnosis can be configured as a chip connected to the central processing unit 9100, and the functions of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis are realized through the control of the central processing unit.

[0185] As Figure 9 shown, the electronic device 9600 may further include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It should be noted that the electronic device 9600 does not necessarily have to include Figure 9 all the components shown in Figure 9 ; in addition, the electronic device 9600 may further include

[0186] components not shown in Figure 9 ; reference may be made to the prior art.

[0187] Among them, the memory 9140 can be, for example, one or more of a buffer, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory, or other suitable devices. The above information related to failures can be stored, and in addition, programs for executing relevant information can be stored. And the central processing unit 9100 can execute the programs stored in the memory 9140 to realize information storage or processing, etc.

[0188] The input unit 9120 provides input to the central processing unit 9100. The input unit 9120 is, for example, a key or a touch input device. The power supply 9170 is used to supply power to the electronic device 9600. The display 9160 is used to display display objects such as images and texts. The display can be, for example, an LCD display, but is not limited thereto.

[0189] The memory 9140 may be a solid-state memory, such as a read-only memory (ROM), a random access memory (RAM), a SIM card, etc. It may also be a memory that stores information even when power is off, can be selectively erased and has more data. Examples of such a memory are sometimes referred to as EPROMs, etc. The memory 9140 may also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142, which is used to store application programs and function programs or the processes for operating the electronic device 9600 by the central processing unit 9100.

[0190] The memory 9140 may also include a data storage unit 9143, which is used to store data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various drivers of the electronic device for communication functions and / or for performing other functions of the electronic device (such as a messaging application, an address book application, etc.).

[0191] The communication module 9110 is a transmitter / receiver that transmits and receives signals via the antenna 9111. The communication module 9110 (transmitter / receiver) is coupled to the central processing unit 9100 to provide input signals and receive output signals, which may be the same as in the case of a conventional mobile communication terminal.

[0192] Based on different communication technologies, multiple communication modules 9110 may be provided in the same electronic device, such as a cellular network module, a Bluetooth module, and / or a wireless local area network module, etc. The communication module 9110 (transmitter / receiver) is also coupled to the speaker 9131 and the microphone 9132 via the audio processor 9130 to provide an audio output via the speaker 9131 and receive an audio input from the microphone 9132, thereby implementing normal telecommunication functions. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. In addition, the audio processor 9130 is also coupled to the central processing unit 9100, so that recording can be performed on the local machine through the microphone 9132 and the sound stored on the local machine can be played through the speaker 9131.

[0193] An embodiment of the present application further provides a computer-readable storage medium capable of implementing all steps of the intelligent decision-making management method for energy-saving device status monitoring and fault diagnosis in the above embodiments where the execution entity is a server or a client. A computer program is stored on the computer-readable storage medium. When the computer program is executed by a processor, all steps of the intelligent decision-making management method for energy-saving device status monitoring and fault diagnosis in the above embodiments where the execution entity is a server or a client are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0194] Step S101: Integrate the signal conditioning module, gain amplification unit, analog-to-digital conversion unit, and arithmetic processor into the edge computing platform. Connect the environmental parameter acquisition module and the energy consumption acquisition module to the data acquisition channels of the analog-to-digital conversion unit. Filter and amplify the acquired signals based on the signal conditioning module. Adjust the gain coefficient of the acquisition channels through the gain amplification unit. Use the analog-to-digital conversion unit to convert the conditioned signals into digital quantities, and transmit the digital quantities to the multi-level data cache through the bus;

[0195] Step S102: Input the monitoring data in the multi-level data cache into the time series feature extraction model. Perform principal component analysis on the extracted time series features to obtain feature vectors. Establish a device working condition judgment model based on the feature vectors. Adjust the sampling frequency according to the working condition judgment model. Input the adjusted sampling data into the data fusion model. Input the fusion data into the collaborative diagnosis system composed of the rule model, machine learning model, and deep network model. Embed the feature mapping library constructed by the device component parameters, fault types, and fault modes into the collaborative diagnosis system. Correct the diagnosis results based on the feature mapping library, and input the corrected diagnosis results into the decision-making system;

[0196] Step S103: Input the output result of the decision-making system and the device historical data into the trend modeling unit. Construct a remaining life prediction model based on the trend modeling unit. Combine the fault evolution analysis to construct a risk warning function. Combine the risk warning function with the remaining life prediction model to establish a hierarchical warning system. Generate a maintenance plan according to the hierarchical warning system. Input the maintenance plan into the maintenance decision-making unit to generate a device maintenance plan. Allocate the maintenance tasks through the remote communication module based on the device maintenance plan.

[0197] As can be seen from the above description, the computer-readable storage medium provided by the embodiments of the present application realizes high-precision data acquisition and processing through an edge computing platform integrating signal conditioning, gain amplification, and analog-to-digital conversion units. The system realizes adaptive sampling based on timing feature extraction and working condition judgment models, and innovatively constructs a collaborative diagnosis system combining rule models, machine learning, and deep networks. By embedding the device feature mapping library into the diagnosis system and combining trend modeling and remaining life prediction, a risk grading and early warning system is established to realize intelligent decision-making and remote management of equipment maintenance. This method provides a comprehensive solution for the predictive maintenance of industrial equipment.

[0198] An embodiment of the present application further provides a computer program product capable of implementing all steps of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis in which the execution subject in the above embodiments is a server or a client. When the computer program / instructions are executed by a processor, the steps of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis are implemented. For example, the computer program / instructions implement the following steps:

[0199] Step S101: Integrate a signal conditioning module, a gain amplification unit, an analog-to-digital conversion unit, and an arithmetic processor into an edge computing platform, connect an environmental parameter acquisition module and an energy consumption acquisition module to the data acquisition channel of the analog-to-digital conversion unit, filter and amplify the acquired signal based on the signal conditioning module, adjust the gain coefficient of the acquisition channel through the gain amplification unit, use the analog-to-digital conversion unit to convert the conditioned signal into a digital quantity, and transmit the digital quantity to a multi-level data cache through a bus;

[0200] Step S102: Input the monitoring data in the multi-level data cache into a timing feature extraction model, perform principal component analysis on the extracted timing features to obtain feature vectors, establish a device working condition judgment model based on the feature vectors, adjust the sampling frequency according to the working condition judgment model, input the adjusted sampling data into a data fusion model, input the fusion data into a collaborative diagnosis system composed of a rule model, a machine learning model, and a deep network model, embed a feature mapping library constructed by device component parameters, fault types, and fault modes into the collaborative diagnosis system, correct the diagnosis result based on the feature mapping library, and input the corrected diagnosis result into a decision-making system;

[0201] Step S103: Input the output result of the decision-making system and the device historical data into the trend modeling unit. Based on the trend modeling unit, construct a remaining life prediction model, construct a risk warning function in combination with fault evolution analysis, combine the risk warning function with the remaining life prediction model to establish a hierarchical warning system, generate a maintenance plan according to the hierarchical warning system, input the maintenance plan into the maintenance decision-making unit to generate a device maintenance plan, and allocate maintenance tasks based on the device maintenance plan through the remote communication module.

[0202] As can be seen from the above description, the computer program product provided by the embodiment of the present application realizes high-precision data acquisition and processing through an edge computing platform integrating a signal conditioning, gain amplification, and analog-to-digital conversion unit. The system realizes adaptive sampling based on timing feature extraction and a working condition judgment model, and innovatively constructs a collaborative diagnosis system combining a rule model, machine learning, and a deep network. By embedding a device feature mapping library into the diagnosis system and combining trend modeling and remaining life prediction, a risk grading warning system is established to realize intelligent decision-making and remote management of device maintenance. This method provides a comprehensive solution for the predictive maintenance of industrial equipment.

[0203] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, apparatus, or computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0204] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (apparatuses), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for realizing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0205] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device that realizes the functions in the processFigure 1 one process or multiple processes and / or blocks Figure 1 the functions specified in one block or multiple blocks.

[0206] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0207] In the present invention, specific embodiments are used to elaborate on the principles and implementation manners of the present invention. The description of the above embodiments is only for helping to understand the method of the present invention and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. An intelligent decision-making management method for monitoring the status and diagnosing faults of energy-saving equipment, characterized in that The method includes: Integrating a signal conditioning module, a gain amplification unit, a digital-to-analog conversion unit, and an arithmetic processor into an edge computing platform, connecting an environmental parameter acquisition module and an energy consumption acquisition module to the data acquisition channels of the digital-to-analog conversion unit, filtering and amplifying the acquired signals based on the signal conditioning module, adjusting the gain coefficient of the acquisition channels through the gain amplification unit, converting the conditioned signals into digital quantities using the digital-to-analog conversion unit, and transmitting the digital quantities to a multi-level data cache through a bus; Inputting the monitoring data in the multi-level data cache into a time series feature extraction model, performing principal component analysis on the extracted time series features to obtain feature vectors, constructing a working condition judgment model based on a support vector machine, inputting the feature vectors into the working condition judgment model for classification, establishing a corresponding relationship between the classification results and a preset sampling frequency threshold, selecting the sampling frequency parameters corresponding to the current working condition according to the corresponding relationship, inputting the adjusted sampling data into a data fusion model, inputting the fusion data into a collaborative diagnosis system composed of a rule model, a machine learning model, and a deep network model, embedding a feature mapping library constructed by device component parameters, fault types, and fault modes into the collaborative diagnosis system, correcting the diagnosis results based on the feature mapping library, and inputting the corrected diagnosis results into a decision-making system; Inputting the device status indicators output by the decision-making system, the operation records, maintenance records, and fault records in the device historical database into a trend modeling unit, decomposing the status indicators into trend terms, periodic terms, and random terms through time series decomposition in the trend modeling unit, establishing a device degradation model based on the Weibull distribution, combining the device degradation model with a long short-term memory network to construct a remaining life prediction model, using the decomposed terms of the status indicators, degradation parameters, and time scales as input variables in the remaining life prediction model, constructing a risk warning function by combining fault evolution analysis, combining the risk warning function with the remaining life prediction model to establish a hierarchical warning system, generating a maintenance plan according to the hierarchical warning system, inputting the maintenance plan into a maintenance decision-making unit to generate a device maintenance plan, and allocating maintenance tasks through a remote communication module based on the device maintenance plan.

2. The intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis according to claim 1, wherein The integration of the signal conditioning module, the gain amplification unit, the digital-to-analog conversion unit, and the arithmetic processor into the edge computing platform, and the connection of the environmental parameter acquisition module and the energy consumption acquisition module to the data acquisition channels of the digital-to-analog conversion unit include: Constructing a heterogeneous computing architecture using a dual-core processor and a programmable gate array, configuring the master-slave interface, storage interface, and communication interface of the heterogeneous computing architecture as data processing units based on a field bus protocol, connecting the amplification circuit, filtering circuit, and isolation circuit of the signal conditioning module to the input end of the data processing unit, and connecting the programmable amplifier and gain adjustment circuit of the gain amplification unit to the output end of the signal conditioning module; Connect the sample and hold circuit, reference source circuit, and digital quantity output circuit of the digital-to-analog conversion unit to the signal output terminal of the gain amplification unit. Configure the logic function module of the programmable gate array based on the hardware description language. Connect the temperature sensor and humidity sensor of the environmental parameter acquisition module, and the current transformer and voltage sensor of the energy consumption acquisition module to the multiple sampling channels of the digital-to-analog conversion unit respectively.

3. The intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis according to claim 1, characterized in that, Filter and amplify the acquired signal based on the signal conditioning module, adjust the gain coefficient of the acquisition channel through the gain amplification unit, convert the conditioned signal into a digital quantity using the digital-to-analog conversion unit, and transmit the digital quantity to the multi-level data cache through the bus, including: Input the acquired signal into a passive filter circuit in the signal conditioning module for band-pass filtering, perform common-mode rejection on the filtered signal through a differential amplifier, input the suppressed signal into an operational amplifier for signal amplification, perform electrical isolation on the amplified signal through an opto-isolator, perform secondary amplification on the isolated signal based on a programmable gain amplifier, set the gain coefficient of the programmable gain amplifier as an adjustable parameter, and dynamically adjust the gain coefficient according to the signal amplitude. Use the sample and hold circuit in the digital-to-analog conversion unit to sample the conditioned signal, compare the sampled signal with the reference voltage generated by the reference source circuit to obtain a digital quantity, transmit the digital quantity to the field programmable gate array buffer through the serial peripheral interface, and transmit the data in the field programmable gate array buffer to the multi-level data cache composed of a dynamic random access memory and a solid-state memory through the system bus.

4. The intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis according to claim 1, characterized in that Input the monitoring data in the multi-level data cache into the time series feature extraction model, and perform principal component analysis on the extracted time series features to obtain feature vectors, including: Divide the monitoring data in the multi-level data cache into multiple data windows according to the time series, extract statistical features, frequency domain features, and time-frequency features from the data windows respectively. Combine the mean, variance, skewness, and kurtosis of the statistical features, the frequency center, frequency variance, and power spectral density of the frequency domain features, and the wavelet coefficients and energy entropy of the time-frequency features to form a feature matrix. Standardize the feature matrix, input the standardized feature matrix into the principal component analysis model, calculate the principal component scores based on eigenvalue decomposition, and construct feature vectors from the principal component scores.

5. The intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis according to claim 1, characterized in that Input the fusion data into the collaborative diagnosis system composed of a rule model, a machine learning model, and a deep network model. Embed the feature mapping library constructed by device component parameters, fault types, and fault modes into the collaborative diagnosis system. Correct the diagnosis results based on the feature mapping library, and input the corrected diagnosis results into the decision-making system, including: Construct a rule model by dividing the fault diagnosis expert knowledge according to the equipment components, construct a machine learning model based on the gradient boosting tree algorithm, construct a deep network model using a recurrent neural network, and form a collaborative diagnosis system with the rule model, the machine learning model, and the deep network model. Input the fusion data into the rule model for rule reasoning, input it into the machine learning model for classification prediction, and input it into the deep network model for sequence analysis. Construct a feature mapping library with the structural parameters, operating parameters, and fault characteristics of the equipment components, and establish a diagnostic confidence scoring system based on the corresponding rules between the fault types and fault modes in the feature mapping library; Input the diagnostic results of the rule model, the machine learning model, and the deep network model into the diagnostic confidence scoring system for scoring, quantify the reliability of the diagnostic results of each model according to the scoring results, and perform weighted combination on the quantified diagnostic results based on the Bayesian inference method to obtain a collaborative diagnosis result. Match and correct the collaborative diagnosis result with the fault characteristics in the feature mapping library, and input the corrected diagnostic result into the decision-making system.

6. The intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis according to claim 1, characterized in that, Construct a risk warning function by combining fault evolution analysis, and combine the risk warning function with the remaining life prediction model to establish a hierarchical warning system, including: Perform clustering analysis on the equipment fault data to identify the fault evolution path, establish a corresponding relationship between the fault evolution path and the state index threshold to construct a risk warning function, map the risk level calculated by the risk warning function to the life interval predicted by the remaining life prediction model, divide the warning level into multiple levels based on the mapping relationship, and set trigger conditions and warning rules for different warning levels to construct a hierarchical warning system.

7. The intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis according to claim 1, characterized in that, Generate a maintenance plan according to the hierarchical warning system, input the maintenance plan into the maintenance decision-making unit to generate an equipment maintenance plan, and allocate maintenance tasks based on the equipment maintenance plan through the remote communication module, including: Associate and match the warning levels in the hierarchical warning system with the maintenance strategy library, construct a maintenance plan generation model based on the repair procedures, maintenance cycles, and spare part information in the maintenance strategy library. In the maintenance plan generation model, use the equipment state index, warning level, and maintenance resource constraints as input parameters to generate a specific maintenance plan, input the maintenance plan into the maintenance decision-making unit, and balance the maintenance cost, maintenance time, and equipment reliability based on the multi-objective optimization algorithm in the maintenance decision-making unit, and prioritize and schedule the equipment maintenance projects; Allocate the maintenance plan output by the maintenance decision-making unit through the remote communication module. In the remote communication module, encode the maintenance task information according to a preset format, establish a communication link for the maintenance task based on the distributed network protocol, and send the encoded maintenance task to the corresponding execution terminal through the communication link. In the execution terminal, decode and classify and store the received maintenance task.

8. An intelligent decision-making management device for monitoring the status and diagnosing faults of an energy-saving device, characterized in that, The device includes: The platform computing module is used to integrate the signal conditioning module, the gain amplification unit, the analog-to-digital conversion unit and the operation processor into the edge computing platform, connect the environmental parameter acquisition module and the energy consumption acquisition module to the data acquisition channels of the analog-to-digital conversion unit, filter and amplify the acquired signals based on the signal conditioning module, adjust the gain coefficient of the acquisition channels through the gain amplification unit, convert the conditioned signals into digital quantities by using the analog-to-digital conversion unit, and transmit the digital quantities to the multi-level data cache through the bus; The model construction module is used to input the monitoring data in the multi-level data cache into the time series feature extraction model, perform principal component analysis on the extracted time series features to obtain feature vectors, construct a working condition judgment model based on the support vector machine, input the feature vectors into the working condition judgment model for classification, establish a corresponding relationship between the classification results and the preset sampling frequency threshold, select the sampling frequency parameters corresponding to the current working condition according to the corresponding relationship, input the adjusted sampling data into the data fusion model, input the fusion data into the collaborative diagnosis system composed of the rule model, the machine learning model and the deep network model, embed the feature mapping library constructed by the equipment component parameters, fault types and fault modes into the collaborative diagnosis system, correct the diagnosis results based on the feature mapping library, and input the corrected diagnosis results into the decision-making system; The monitoring and diagnosis module is used to input the equipment status indicators output by the decision-making system, the operation records, maintenance records and fault records in the equipment historical database into the trend modeling unit, decompose the status indicators into trend items, periodic items and random items by time series decomposition in the trend modeling unit, establish an equipment degradation model based on the Weibull distribution, combine the equipment degradation model with the long short-term memory network to construct a remaining life prediction model, use the status indicator decomposition items, degradation parameters and time scales as input variables in the remaining life prediction model, construct a risk warning function by combining fault evolution analysis, combine the risk warning function with the remaining life prediction model to establish a hierarchical warning system, generate a maintenance plan according to the hierarchical warning system, input the maintenance plan into the maintenance decision-making unit to generate an equipment maintenance plan, and allocate maintenance tasks based on the equipment maintenance plan through the remote communication module.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it realizes the steps of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it realizes the steps of the intelligent decision-making management method for energy-saving equipment status monitoring and fault diagnosis according to any one of claims 1 to 7.

Images