An industrial air conditioner energy-saving optimization control method based on multi-source data fusion and related equipment

Abstract

The application provides an industrial air conditioner energy-saving optimization control method based on multi-source data fusion and related equipment, and is applied to the technical field of data processing. The application takes industrial air conditioner energy-saving optimization as the core, collects multi-source data such as equipment parameters and environmental variables, and fuses layer processing standards, AI decision operation benchmarks and other system level benchmark data. The multi-source data fusion parameters and the level control strategy parameters are combined, the data processing and instruction execution logic are matched, the hierarchical processing and multi-modal integration are relied on, the collaborative rules are established, and the energy-saving constraint conditions are generated. Based on the above data and constraint collection energy consumption, operation and fault warning data, combined with energy consumption prediction, supply and demand regulation and other functional parameters, energy-saving effect and operation state associated data are generated. Finally, the energy management platform visual analysis and ISO50001 compliance audit output energy-saving optimization and system reliability evaluation results.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an energy-saving optimization control method and related equipment for industrial air conditioning based on multi-source data fusion. Background Technology

[0002] Current industrial air conditioning systems suffer from numerous technical deficiencies in energy-saving control and data management, failing to meet the comprehensive requirements of "high efficiency, stable operation, and compliant management" in industrial scenarios. Specifically: Industrial air conditioning operation involves multi-dimensional data, including equipment parameters (such as compressor frequency and valve opening), environmental variables (such as temperature, humidity, and light intensity), user behavior (such as personnel density), and production plans (such as shift schedules and capacity adjustments). However, existing systems lack a unified data fusion architecture, resulting in isolated data silos. Redundancy at the data layer is not eliminated, feature layer dimensions are not optimized, and decision-making data lacks coordination, leading to low data quality and an inability to effectively support energy-saving decisions. Existing systems often rely on fixed thresholds or simple PID control, failing to incorporate AI algorithms for dynamic optimization and lacking a closed-loop logic of "data fusion - decision generation - execution feedback." For example, cooling supply and demand adjustments do not consider dynamic changes in production plans and environmental loads, easily leading to "overcooling" or "energy waste," making it difficult to achieve system-level energy-saving goals (such as an annual energy efficiency improvement of 3%-5%).

[0003] Existing monitoring systems often focus on single equipment parameters (such as compressor frequency), failing to cover comprehensive data such as energy consumption distribution and optimization effects. Furthermore, data acquisition accuracy (e.g., energy consumption error exceeding 2%) and real-time performance (e.g., latency exceeding 1 second) are substandard. Fault warnings rely solely on single parameter thresholds, lacking multimodal data correlation analysis. The existing system is not aligned with the ISO 50001 energy management system standard, and its energy consumption statistics methods and data acquisition accuracy do not meet compliance requirements. Moreover, it lacks visualization and analysis tools, failing to intuitively present key information such as regional energy consumption differences and equipment fluctuation trends, which is detrimental to energy consumption control and optimization effect evaluation.

[0004] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this disclosure and does not constitute information on prior art known to those skilled in the art. Summary of the Invention

[0005] According to one aspect of this application, an energy-saving optimization control method for industrial air conditioning based on multi-source data fusion is provided, comprising: acquiring multi-source data acquisition parameters and system-level benchmark data; the multi-source data acquisition parameters including equipment parameters, environmental variables, user behavior, and production plan data selected according to the operational needs of industrial air conditioning; and the system-level benchmark data including multi-source data fusion layer processing standards, AI optimization decision layer computation benchmarks, and edge control execution layer control benchmarks; combining the multi-source data fusion parameters and control strategy parameters at each level, synchronously matching the data processing flow and control command execution logic; and establishing collaborative rules for multi-source data fusion and system control based on hierarchical processing technology and multimodal integration design to generate energy-saving optimization control rules for the entire system. It can optimize constraints; based on the combination of multi-source data acquisition parameters, system-level benchmark data, collaborative control logic, and dynamic strategy constraints, it collects energy consumption data, equipment operation data, optimization effect data, and fault early warning data; it processes the collected system operation data and combines it with functional parameters including energy consumption prediction model results and supply and demand balance adjustment parameters to generate correlation data between industrial air conditioning energy-saving effect and system operation status. Among them, functional parameters are grouped into energy consumption precision control parameters and system-level energy-saving optimization parameters; it processes the correlation data and parameter settings based on the visualization analysis of the energy management platform and the ISO50001 compliance audit standard to generate industrial air conditioning energy-saving optimization and system reliability evaluation results.

[0006] Another aspect of this application discloses an industrial air conditioning energy-saving optimization control device based on multi-source data fusion, comprising: an acquisition module for acquiring multi-source data acquisition parameters and system-level benchmark data; the multi-source data acquisition parameters including equipment parameters, environmental variables, user behavior, and production plan data selected according to the operational needs of the industrial air conditioning; and the system-level benchmark data including multi-source data fusion layer processing standards, AI optimization decision-making layer computational benchmarks, and edge control execution layer regulation benchmarks; and a processing module for combining the multi-source data fusion parameters and control strategy parameters at each level, synchronously matching the data processing flow with the control command execution logic, and establishing collaborative rules for multi-source data fusion and system control based on hierarchical processing technology and multimodal integration design. Generate energy-saving optimization constraints for the entire system; process constraints based on multi-source data acquisition parameters, system-level baseline data, collaborative control logic, and dynamic strategy combinations, collecting energy consumption data, equipment operation data, optimization effect data, and fault early warning data; process the collected system operation data, combining functional parameters including energy consumption prediction model results and supply-demand balance adjustment parameters, to generate correlation data between industrial air conditioning energy-saving effect and system operation status. Among them, functional parameters are grouped into energy consumption precision control parameters and system-level energy-saving optimization parameters; process the correlation data and parameter settings based on the energy management platform's visualization analysis and ISO50001 compliance audit standards to generate industrial air conditioning energy-saving optimization and system reliability evaluation results.

[0007] According to another aspect of this application, an electronic device includes: a first processor; and a memory for storing executable instructions of the first processor; wherein the first processor is configured to execute the above-described industrial air conditioning energy-saving optimization control method based on multi-source data fusion by executing the executable instructions.

[0008] According to another aspect of this application, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a second processor, implements the above-described industrial air conditioning energy-saving optimization control method based on multi-source data fusion.

[0009] This application provides an industrial air conditioning energy-saving optimization control method and related equipment based on multi-source data fusion. It constructs a closed-loop process of "data input - rule establishment - data acquisition - correlation analysis - evaluation output" around industrial air conditioning energy-saving optimization. First, it acquires multi-source data collection parameters such as equipment, environment, user behavior, and production plans, as well as system-level benchmark data for the fusion layer, AI decision layer, and edge control layer. Then, through hierarchical processing and multimodal integration, it establishes collaborative rules to generate energy-saving constraints. Subsequently, it collects energy consumption, equipment operation, optimization effect, and fault early warning data, which, after preprocessing and fusion, are combined with energy consumption prediction models and supply-demand balance adjustment parameters to generate correlation data between energy-saving effects and operating status. Finally, relying on visualization analysis and the ISO50001 standard, it outputs energy-saving optimization and system reliability evaluation results. The core technology lies in breaking down "information silos" through multi-source data fusion, achieving dynamic optimization by combining AI algorithms, and balancing energy consumption control, operational stability, and compliance, thus solving problems such as low energy efficiency and insufficient data management in traditional systems.

[0010] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0011] Figure 1 The flowchart illustrates an embodiment of an energy-saving optimization control method for industrial air conditioning based on multi-source data fusion provided in this application.

[0012] Figure 2 The diagram shows a structural schematic of an industrial air conditioning energy-saving optimization control device based on multi-source data fusion, according to an embodiment of this application. Detailed Implementation

[0013] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0014] The following is combined Figure 1This application describes an energy-saving optimization control method for industrial air conditioning based on multi-source data fusion, according to exemplary embodiments thereof. It should be noted that the following application scenarios are shown only to facilitate understanding of the spirit and principles of this application, and the embodiments of this application are not limited in any way. Rather, the embodiments of this application are applicable to any suitable scenario.

[0015] In one implementation, Figure 1 A schematic diagram of a process flow for an industrial air conditioning energy-saving optimization control method based on multi-source data fusion according to an embodiment of this application is shown.

[0016] S101, acquire multi-source data acquisition parameters and system-level baseline data.

[0017] In one implementation, the multi-source data acquisition parameters need to be determined based on the actual operating scenario and control requirements of the industrial air conditioning system. The core data covers four dimensions: equipment, environment, user behavior, and production plan. Parameters and examples for each dimension are as follows. Regarding equipment parameters, the focus is on key indicators of the core operating components of the industrial air conditioning system. Real-time data collection of equipment operating status reflects the unit's efficiency. Examples include compressor frequency (e.g., current operating frequency 50Hz, which can be dynamically adjusted within a range of 30-70Hz according to cooling demand) and valve opening (e.g., chilled water valve opening 60%, adjustment accuracy ±5%, used to control the cooling capacity delivery rate).

[0018] For environmental variables, data on environmental conditions in the industrial air conditioning service area are collected to provide a basis for load forecasting and operation strategy adjustment. Examples include temperature and humidity (e.g., the current workshop temperature is 26℃ and humidity is 65%, and the temperature and humidity control thresholds are set to 24-28℃ and 55%-70%, respectively) and light intensity (e.g., the natural light intensity during the day is 800 lux, which is used to determine whether the air conditioning load needs to be adjusted in conjunction with lighting heat dissipation).

[0019] Based on user behavior parameters, data is collected on the activity patterns of personnel and equipment usage in industrial scenarios to match actual energy demand. Examples include personnel density (e.g., the real-time personnel density in Workshop A is 2 people / ㎡, and the density rises to 3 people / ㎡ during peak hours (8:00-12:00, 14:00-18:00), requiring a corresponding increase in cooling intensity) and equipment start-up and shutdown (e.g., production equipment in the workshop starts up at 9:00, and the heat dissipation increases after startup, so the air conditioning needs to adjust its operating parameters 15 minutes in advance).

[0020] Based on production planning data and combined with data collected from factory production schedules, we can adapt in advance to load changes brought about by production activities. Examples include production shifts (e.g., if a production line operates on a two-shift system, with 8:00-20:00 as the production period, the air conditioning needs to maintain stable operation; 20:00-8:00 the next day is the shutdown period, during which the operating power can be reduced) and capacity adjustment plans (e.g., if the capacity increases by 20% next week, the number of production equipment in operation will increase accordingly, and the air conditioning needs to anticipate a 15% increase in cooling demand in advance).

[0021] The system-level benchmark data needs to be determined based on the functional positioning of the three-level technical architecture of industrial air conditioning: "multi-source data fusion layer - AI optimization decision layer - edge control execution layer". The benchmark data and examples for each level are as follows. For the multi-source data fusion layer processing standard, a unified processing rule is provided for heterogeneous data integration to ensure data quality and fusion effectiveness. Examples include: data layer redundancy removal standard (using the Kalman filter algorithm, setting the filter coefficient to 0.8, to perform redundancy removal processing on the collected compressor frequency raw data, eliminating abnormal data with fluctuations exceeding ±3Hz); feature layer dimensionality reduction standard (using PCA dimensionality reduction technology to reduce the dimensionality of 5 dimensions of environmental variables such as temperature, humidity, and light to 2 principal components, retaining more than 90% of the information in the original data); and decision layer data integration standard (using a multi-model voting mechanism, setting 3 prediction models (LSTM, ARIMA, BP neural network), when the prediction error of at least 2 models is <3%, the prediction result is used as the decision basis).

[0022] To optimize the computational benchmark of AI decision-making, the accuracy requirements and parameter range of AI algorithm computation are clarified to ensure the scientific nature of optimization decisions. Examples include energy efficiency ratio prediction accuracy (building an energy consumption prediction model based on neural networks, setting the energy efficiency ratio prediction error to be <3%, such as predicting an energy efficiency ratio of 3.2 under a certain operating condition, while the actual operating energy efficiency ratio should be within the range of 3.1-3.3), and genetic algorithm optimization benchmark (using the energy efficiency ratio prediction result as the fitness function of the genetic algorithm, setting the algorithm iteration count to 50 times and the population size to 100, ensuring a balance between computational efficiency and optimization effect, and quickly locating the optimal operating parameter range of the equipment).

[0023] To standardize the control accuracy and operating mode of the equipment execution layer, and ensure the accurate implementation of control commands, examples include temperature control accuracy (using PID algorithm to control air conditioning operation, setting temperature control accuracy to ±0.5℃, such as a target temperature of 25℃, the actual operating temperature needs to be stable at 24.5-25.5℃) and multi-mode switching benchmark (supporting two modes: strong cooling and silent energy saving, setting to automatically switch to strong cooling mode when the workshop temperature is higher than 28℃, and switch to silent energy saving mode when it is lower than 24℃, with a mode switching response time of <10 seconds).

[0024] S102 combines the multi-source data fusion parameters and control strategy parameters at each level, synchronously matches the data processing flow and control command execution logic, establishes collaborative rules for multi-source data fusion and system control based on hierarchical processing technology and multimodal integration design, and generates energy-saving optimization constraints for the entire system.

[0025] In one implementation, a correlation analysis is performed on the multi-source data fusion parameters and control strategy parameters at each level to generate a correlation analysis result of the data fusion and control strategy adaptation simulation. The multi-source data fusion parameters include Kalman filter redundancy removal parameters at the data layer, PCA dimensionality reduction / CCA subspace learning parameters at the feature layer, multi-model voting parameters at the decision layer, and LSTM model parameters for dynamic load forecasting. The Kalman filter redundancy removal parameters at the data layer are used to filter noise and redundant information in the original data. Key parameters include the filter gain (e.g., set to 0.8) and the state estimation error covariance matrix (initialized as a diagonal matrix with diagonal elements of 0.1). For example, for the collected compressor frequency raw data (including ±5Hz fluctuation noise), after processing with this parameter, abnormal data with fluctuation amplitudes exceeding ±3Hz can be removed, retaining valid operating data.

[0026] The PCA dimensionality reduction parameters for the feature layer include the principal component contribution rate (set to ≥90%) and the number of dimensions after dimensionality reduction (e.g., reducing 5-dimensional environmental data such as temperature, humidity, and light intensity to 2-dimensional). The CCA subspace learning parameters include the cross-modal data correlation threshold (set to ≥0.7). For example, CCA processing is performed on equipment parameters (compressor frequency, valve opening) and environmental variables (temperature and humidity) to select feature combinations with a correlation of ≥0.7, thereby improving data fusion efficiency.

[0027] The multi-model voting parameters for the decision-making level include the number of models participating in the vote (e.g., 3 models: LSTM, ARIMA, and BP neural network) and the voting pass threshold (e.g., at least 2 models have a prediction error of <3%). For example, in dynamic load forecasting, when the prediction errors of LSTM and BP neural network for cooling demand are 2.5% and 2.8% respectively, and the error of ARIMA is 4%, the voting pass threshold is met, and the average prediction value of the former two is used as the basis for decision-making.

[0028] The parameters of the dynamic load forecasting LSTM model include the time step (set to 1 hour, i.e., based on historical data per hour), the number of hidden layer neurons (e.g., 128), and the prediction error threshold (<3%). For example, by combining the environmental temperature and humidity and production plan data of the past 24 hours, the cooling demand can be predicted 1-4 hours in advance. When the predicted cooling demand for a certain period is 100kW, the actual demand should be within the range of 97-103kW to meet the error requirements.

[0029] The control strategy parameters for the multi-source data fusion layer include data acquisition frequency (e.g., equipment parameters are collected once every 10 seconds, environmental variables are collected once every minute), and data preprocessing trigger conditions (e.g., when the fluctuation range of the raw data exceeds ±10%, Kalman filtering is automatically initiated). The control strategy parameters for the AI ​​optimization decision layer include the number of genetic algorithm iterations (e.g., 50 times), population size (e.g., 100 individuals), and the computational scale of cloud-based operating conditions (e.g., calculating 100,000 equipment combination operating conditions each time). For example, when generating the optimal operating scheme for the chiller / heater unit, the scheme with the highest energy efficiency ratio is selected from 100,000 operating conditions through 50 iterations. The control strategy parameters for the edge control execution layer include the PID algorithm proportional gain (e.g., 2.0), integral time (e.g., 5 seconds), derivative time (e.g., 1 second), and multi-mode switching trigger thresholds (e.g., switching to strong cooling mode when the temperature is above 28℃, and switching to silent energy-saving mode when the temperature is below 24℃).

[0030] The correlation between multi-source data fusion parameters and control strategy parameters was calculated using the Pearson correlation coefficient. For example, the correlation between the Kalman filter gain (0.8) of the data layer and the data acquisition frequency (10 seconds / time) of the multi-source data fusion layer was 0.85, indicating high compatibility and that the filter gain effectively matched the redundancy removal requirements of high-frequency acquired data. The correlation between the dimension (2-dimensional) after PCA dimensionality reduction of the feature layer and the cloud-based operating condition calculation volume (100,000 types) of the AI ​​optimization decision layer was 0.78, indicating that the dimensionality-reduced feature data could efficiently support large-scale operating condition calculations. The correlation between the time step of the dynamic load prediction LSTM model (1 hour) and the integration time (5 seconds) of the PID algorithm in the edge control execution layer was 0.65, requiring further optimization of compatibility in subsequent steps. Finally, a correlation analysis result table was generated, labeling the correlation values ​​and compatibility levels of each parameter combination (high: ≥0.8, medium: 0.6-0.79, low: <0.6).

[0031] The correlation analysis results are filtered to generate a list of target simulation variables. This list includes multi-source data fusion layer processing parameters, AI optimization decision parameters, edge control execution parameters, data-control response timeliness parameters, and multimodal data integration error parameters. Core parameters of the fusion layer with a correlation coefficient ≥ 0.6 are selected. Examples include: Kalman filter gain (0.8, correlation coefficient 0.85): used to simulate the impact of different filter intensities on data quality; PCA dimensionality reduction dimension (2-dimensional, correlation coefficient 0.78): a key variable simulating the correlation between feature data dimension and decision efficiency; and LSTM model time step (1 hour, correlation coefficient 0.65): used to simulate the impact of time-series data acquisition intervals on load forecasting accuracy.

[0032] Filter decision-level parameters with a correlation coefficient ≥ 0.6. Examples include: number of iterations in the genetic algorithm (50 iterations, correlation coefficient 0.72): simulating the impact of the number of iterations on the accuracy of the optimization scheme; cloud-based operating condition computational scale (100,000 scenarios, correlation coefficient 0.78): serving as a variable to balance the computational scale and decision-making efficiency in the simulation. Filter execution-level parameters with a correlation coefficient ≥ 0.6. Examples include: PID algorithm proportional coefficient (2.0, correlation coefficient 0.68): simulating the impact of the proportional coefficient on temperature control accuracy (±0.5℃); multi-mode switching temperature threshold (24℃ / 28℃, correlation coefficient 0.75): used to simulate the impact of mode switching timing on energy-saving effects. Filter time-related parameters with a correlation coefficient ≥ 0.6. Examples include: data fusion processing time (e.g., ≤ 1 second, correlation coefficient 0.82): simulating the matching relationship between fusion processing speed and control command generation timeliness; cloud-based decision command issuance time (e.g., ≤ 2 seconds, correlation coefficient 0.76): ensuring that decision results can be transmitted to the edge execution layer in a timely manner.

[0033] Error parameters with a correlation coefficient ≥ 0.6 were selected. Examples include the cross-modal data (equipment parameters and environmental variables) integration error (≤ 2%, correlation coefficient 0.69): the impact of simulation error on the quality of fused information. The error between load forecast and actual demand (< 3%, correlation coefficient 0.81): a core error indicator for verifying the adaptability of fused parameters to control strategies.

[0034] The correlation analysis results and the target simulation variable list are processed to generate parameter combination constraints, including hierarchical collaborative constraints conforming to hierarchical processing technology, time-series synchronization constraints for multi-source data fusion and system control, and functional adaptation constraints for multimodal integrated design. Combining the correlation analysis results (medium-to-high adaptation parameter combinations) with the target simulation variables, constraints are generated from three dimensions: hierarchical collaboration, time-series synchronization, and functional adaptation, ensuring efficient collaboration between multi-source data fusion and system control. Based on the hierarchical fusion logic of "data layer - feature layer - decision layer," the parameter transmission and processing rules at each level are constrained. For data layer processing constraints: the Kalman filter redundancy removal parameter (gain 0.8) must prioritize processing the raw data of equipment parameters (compressor frequency, valve opening) and environmental variables (temperature and humidity). Only processed data can enter the feature layer, and the data layer processing time must be ≤1 second, matching the data-control response timeliness parameter. Regarding the feature layer-decision layer collaboration constraint: the dimensionality (2D) data after PCA dimensionality reduction must meet the cloud-based working condition calculation requirements of the AI-optimized decision layer, that is, the dimensionality-reduced data must be able to support 100,000 kinds of working condition calculations, and the latency of feature data being transmitted to the decision layer must be ≤0.5 seconds to ensure efficient data flow between layers.

[0035] To address the synchronization constraints between load forecasting and control execution, the dynamic load forecasting LSTM model (time step 1 hour) needs to output cooling demand forecast results 1-4 hours in advance. The edge control execution layer needs to adjust the PID algorithm parameters (proportional coefficient 2.0) based on these results 15 minutes in advance to ensure seamless timing between forecasting and execution. For example, if the predicted cooling demand increases at 9:00, the PID parameter adjustment should begin at 8:45. Regarding the synchronization constraints between data acquisition and fusion, equipment parameters are collected every 10 seconds, and Kalman filtering must be completed within 1 second of acquisition to ensure that the fused data can promptly support the AI ​​decision-making layer's operating condition calculations every 5 minutes, avoiding time-series discrepancies.

[0036] To ensure the compatibility between multimodal data (equipment, environment, user behavior, production plan) and system control functions, the LSTM model must simultaneously integrate equipment parameters (compressor frequency), environmental variables (temperature and humidity), and production plan data (production shifts) for multimodal data compatibility with load forecasting. The integration error of each modality should be ≤2% to ensure the prediction results match actual cooling demands. For example, if the production shift is from 8:00 to 20:00, the model should prioritize incorporating equipment heat dissipation and personnel activity data during that period. Regarding the compatibility between fused data and optimization decision-making, the decision-making basis output by the multi-model voting parameters (two models with an error <3% pass) must support the genetic algorithm in generating a coordinated solution for chiller / heater units, water pumps, and terminal fans. Furthermore, the solution must meet the waste heat conversion requirements of condensing heat recovery technology (producing 75℃ process hot water) to ensure compatibility between fused data and energy-saving functions.

[0037] S103 processes multi-source data acquisition parameters, system-level baseline data, collaborative control logic, and dynamic strategy combination constraints to collect energy consumption data, equipment operation data, optimization effect data, and fault early warning data.

[0038] In one implementation, multi-source data acquisition parameters, system-level benchmark data, collaborative control logic, and dynamic strategy combination constraints are extracted and classified to generate energy consumption data acquisition dimension information, equipment operation data monitoring item information, optimization effect data evaluation dimension information, and fault early warning data triggering basis information. Multi-source data acquisition parameters include equipment parameters (compressor frequency, valve opening degree), environmental variables (temperature, humidity, light intensity), etc.; system-level benchmark data includes the Kalman filter standard of the multi-source data fusion layer and the PID temperature control accuracy (±0.5℃) of the edge control execution layer, etc.; collaborative control logic includes data flow rules from the data layer to the feature layer to the decision layer; and dynamic strategy combination constraints include the range of collaborative parameters for the cold and heat source units and the fault prediction response time (<5 minutes).

[0039] For energy consumption data collection dimensions, extract energy consumption-related parameters and constraints, and clarify the collection dimensions. Examples include collection objects (energy consumption of cold and heat source units, water pump energy consumption, terminal fan energy consumption, and total energy consumption of each workshop area) and collection cycles (equipment energy consumption is collected once every 15 minutes, and total regional energy consumption is summarized once every hour). This information is generated based on the "equipment parameters" in the multi-source data collection parameters and the "energy-saving optimization threshold" in the dynamic strategy combination constraints, ensuring coverage of all objects in energy consumption statistics.

[0040] For equipment operation data monitoring items, core operating parameters and monitoring requirements of the equipment are extracted. Examples include compressor (operating frequency, monitoring range 30-70Hz, out of range is marked as abnormal), valve (cold water valve opening, monitoring accuracy ±5%, matching PID control accuracy), water pump (speed, monitoring range 1000-3000 rpm), and terminal fan (operating mode, monitoring whether it switches normally between strong cooling / silent energy-saving mode). This information is generated based on the "edge control execution layer control benchmark" in the system-level benchmark data and the "equipment control logic" in the collaborative control logic, which fits the equipment operation status monitoring requirements.

[0041] For the data evaluation dimensions of optimization effect, relevant evaluation indicators for energy-saving optimization effect are extracted. Examples include energy saving rate (monthly energy saving rate, which needs to match the annual energy efficiency improvement benchmark of 3%-5% of the adaptive learning algorithm), power saving of abnormal production lines (monthly power saving of a single production line, with an evaluation range of "12,000 kWh per month for abnormal production lines"), and energy efficiency improvement rate (annual energy efficiency improvement value, which needs to meet ≥3%). This information is generated based on the "AI optimization decision layer operation benchmark" in the system-level benchmark data and the "energy efficiency closed-loop evolution" content in the key function innovation, ensuring that the evaluation indicators are consistent with the energy-saving target.

[0042] Based on the fault warning data trigger information, relevant parameters and triggering conditions for fault prediction are extracted. Examples include equipment parameter deviation (warning is triggered when compressor frequency fluctuation exceeds ±5Hz or valve opening abnormal fluctuation exceeds ±10%), fault prediction response time (it takes less than 5 minutes from detecting an anomaly to generating a warning), and dual cold source redundancy switching trigger (when the main cold source fails, the backup cold source must start within 10 minutes to ensure 7×24-hour temperature control). This information is generated based on the "safety and reliability assurance" content in key functional innovation and the "fault risk threshold" in the dynamic strategy combination constraints, thus clarifying the fault warning triggering standard.

[0043] Based on the technical specifications for industrial air conditioning system operation monitoring, the system processes information on energy consumption data collection dimensions, equipment operation data monitoring items, optimization effect data evaluation dimensions, and fault warning data triggering criteria to generate energy consumption data collection accuracy matching information and equipment operation data real-time evaluation information. According to the technical specifications for industrial air conditioning system operation monitoring (such as data collection accuracy and real-time requirements), the above four types of information are processed to generate energy consumption data collection accuracy matching information and equipment operation data real-time evaluation information, ensuring that the collected information complies with industry standards.

[0044] Referring to the requirement in the standard that "energy consumption data acquisition accuracy must be ≤2%", accuracy matching is performed on the dimensions of energy consumption data acquisition. For energy consumption acquisition of cold and heat source units, electricity meters with an accuracy of Class 1 are selected to ensure that the acquisition error is ≤1%, matching the accuracy requirements of the standard. For the aggregation of total regional energy consumption, a data fusion algorithm is used to eliminate acquisition errors from different devices, ensuring that the aggregated error is ≤2%. Finally, accuracy matching results are generated, labeling the acquisition device model (e.g., electricity meter model DT862), accuracy class (Class 1), and error range (≤1% / ≤2%) for each energy consumption acquisition object, forming energy consumption data acquisition accuracy matching information to ensure the accuracy and reliability of energy consumption data.

[0045] Based on the requirement in the standard that "the data acquisition delay for equipment operation must be ≤1 second," a real-time assessment of the equipment operation data monitoring items is conducted. For compressor frequency acquisition, a high-frequency acquisition module is used, with an acquisition delay of 0.5 seconds, meeting the ≤1-second requirement. For valve opening acquisition, data is transmitted via a real-time communication protocol (such as Modbus-TCP), with a delay of 0.8 seconds, meeting the real-time standard. Finally, a real-time assessment result is generated, indicating the acquisition delay (0.5 seconds / 0.8 seconds), communication protocol (Modbus-TCP), and whether the standard requirement is met (yes) for each monitoring item, forming real-time assessment information for equipment operation data and ensuring real-time feedback of equipment data.

[0046] Based on energy consumption data acquisition accuracy matching information and equipment operation data real-time assessment information, target data in multi-source data acquisition parameters and system-level benchmark data are marked and filtered to generate operation data correlation filtering results. The marking rule for target data is as follows: if the data acquisition accuracy meets the accuracy matching information requirements and the real-time performance meets the real-time assessment information requirements, it is marked as "valid data"; otherwise, it is marked as "invalid data" and discarded.

[0047] For the screening of multi-source data acquisition parameters, the equipment parameters "compressor frequency acquisition" (accuracy ±1Hz, delay 0.5 seconds, meets the standard) and "valve opening acquisition" (accuracy ±5%, delay 0.8 seconds, meets the standard) are marked as valid data; the environmental variable "illuminance acquisition" (original acquisition delay 2 seconds, exceeds the 1-second requirement) is marked as invalid data, removed and re-included after replacing with high-frequency acquisition equipment.

[0048] For the system-level benchmark data screening, the edge control execution layer's "PID temperature control accuracy ±0.5℃" (matches the "valve opening accuracy" in the equipment operation data monitoring item, meeting the standard) and the AI ​​optimization decision layer's "energy consumption prediction error <3%" (matches the "energy saving rate calculation accuracy" in the optimization effect data evaluation dimension, meeting the standard) were marked as valid data; the multi-source data fusion layer's "Kalman filter redundancy removal parameter (original filter coefficient 0.6, resulting in a data accuracy error of 3%, exceeding the 2% requirement)" was marked as invalid data, and after adjusting the filter coefficient to 0.8 (the error decreased to 1.5%), it was remarked as valid data.

[0049] Finally, the system generates a correlation screening result of the running data, summarizes the list of valid data, and marks whether the accuracy and real-time performance of each data point meet the standards and the adjustment measures (such as changing the model of the light acquisition equipment or adjusting the Kalman filter coefficient) to ensure that all subsequent data collected meet the standards.

[0050] The results of the correlation screening of operational data are integrated and quantified to generate industrial air conditioning operation data collection results. These results, used to characterize the operational status, energy-saving optimization effects, and equipment health of the industrial air conditioning system, include precise data collection levels such as energy consumption, equipment operation, optimization effect, fault warning, and data collection accuracy thresholds. The effective data from the correlation screening results are integrated (categorized by energy consumption, equipment operation, optimization effect, and fault warning) and quantified (assigned specific values ​​or ranges) to generate industrial air conditioning operation data collection results, comprehensively characterizing the system's operational status, energy-saving effects, and equipment health.

[0051] During integration, data is categorized into four dimensions: energy consumption, equipment, optimization, and faults. Quantification involves setting numerical ranges to ensure data is quantifiable and comparable. The industrial air conditioning operation data collection results include five core components, as follows: For energy consumption data, the monthly energy consumption of the chiller / heater unit is 8000 kWh (collection accuracy level 1, error 0.8%), the monthly energy consumption of the water pump is 3000 kWh (error 1.2%), and the total monthly energy consumption of Workshop A is 12000 kWh (error 1.5%). Quantified values ​​are set within a reasonable range with reference to the "monthly electricity saving of 12,000 kWh for abnormal production lines," and the accuracy meets the requirement of ≤2%.

[0052] Regarding the equipment operation data, the compressor's average daily operating frequency is 50Hz (fluctuation range 48-52Hz, no abnormalities), the average daily opening degree of the chilled water valve is 60% (fluctuation range 58-62%, within ±5% accuracy), the average daily speed of the water pump is 2000 rpm (stable operation), and the terminal fan's strong cooling mode operation time is 8 hours / day (matching peak production periods). The quantitative data is consistent with the equipment operation data monitoring items.

[0053] Based on the optimization effect data, the monthly energy saving rate is 8% (exceeding the annual benchmark of 3%-5%), the monthly electricity saving of abnormal production line B is 13,000 kWh (within the reference range of 12,000 kWh), and the annual energy efficiency improvement is 4% (meeting the requirement of ≥3%). The quantitative results are in line with the evaluation dimensions of optimization effect data and reflect the effectiveness of energy saving optimization.

[0054] Regarding the fault warning data, there was one abnormal fluctuation in compressor frequency this month (fluctuation ±6Hz, handled within 5 minutes after triggering the warning), 0 abnormal valve openings, and 0 dual cold source redundancy switchings. The fault warning data is consistent with the triggering information, reflecting the health status of the equipment.

[0055] The thresholds for achieving data collection accuracy are as follows: energy consumption data collection error ≤2%, equipment operation data collection delay ≤1 second, optimization effect data evaluation error ≤3%, and fault warning data response time ≤5 minutes. These thresholds integrate energy consumption data collection accuracy matching information with equipment operation data real-time evaluation information, clarify the standards for collection accuracy and timeliness, and ensure that subsequent collections continue to meet the standards.

[0056] S104 processes the collected system operation data and combines it with functional parameters including energy consumption prediction model results and supply and demand balance adjustment parameters to generate data on the correlation between industrial air conditioning energy-saving effect and system operation status.

[0057] In one implementation, the collected system operation data is preprocessed. After data cleaning and standardization, core system operation features are extracted to generate preprocessed system operation feature data. These core features include equipment operating parameter fluctuation characteristics and energy consumption time-period variation characteristics. For the collected compressor frequency data (including some "999Hz" outliers caused by sensor malfunctions) and energy consumption data (containing occasional "0kWh" missing values ​​due to network outages), an "outlier removal + missing value completion" process is used: outliers >70Hz or <30Hz in the compressor frequency are removed, and missing energy consumption data is filled using the "15-minute average completion method" (e.g., the missing water pump energy consumption at 10:15 is filled with the average of 2.5kWh from 10:00 and 10:30), ensuring data validity.

[0058] Normalize the operating parameters of equipment with different dimensions. For example, convert the compressor frequency (30-70Hz) and the cold water valve opening (0%-100%) into standardized data in the 0-1 range using the formula "(actual value - minimum value) / (maximum value - minimum value)" (e.g., the compressor frequency of 50Hz is standardized to 0.5, and the valve opening of 60% is standardized to 0.6) to eliminate the influence of dimensional differences on subsequent analysis.

[0059] Based on the pre-processed compressor frequency and valve opening data, the fluctuation amplitude and frequency per unit time are calculated to address the fluctuation characteristics of equipment operating parameters. For example, the "maximum fluctuation value of compressor frequency in 1 hour" (when the frequency rises from 45Hz to 55Hz in a certain period, the fluctuation value is 10Hz, and the warning threshold exceeding ±5Hz is marked as "high fluctuation characteristic") and the "number of fluctuations of valve opening in 30 minutes" (3 fluctuations in 1 hour, with each fluctuation amplitude ≤5%, are marked as "stable fluctuation characteristic") are extracted. These characteristics reflect the stability of equipment operation.

[0060] Based on the characteristics of energy consumption changes during different time periods, and combined with production plan data (e.g., production from 8:00 to 20:00 and shutdown from 20:00 to 8:00 the next day), the differences in energy consumption during different time periods are analyzed. For example, the "energy consumption ratio of production period vs. shutdown period" can be extracted (e.g., the total energy consumption of the workshop during the production period is 100 kWh / h, and during the shutdown period it is 30 kWh / h, with a ratio of 3.3:1) and the "duration of peak energy consumption period" (e.g., the peak energy consumption period is from 12:00 to 14:00, lasting for 2 hours, with a peak value of 120 kWh / h). This feature reflects the correlation between energy consumption and production activities, which meets the functional innovation requirement of "dynamic balance between supply and demand".

[0061] The preprocessed system operation characteristic data and the energy consumption prediction model results are correlated. The linear correlation between the two is analyzed using the Pearson correlation coefficient, and the time-series correlation pattern is captured using the sliding window method to generate correlation characteristic information between operation data and energy consumption prediction. The energy consumption prediction model outputs a predicted energy efficiency ratio of 3.2 for a certain period, corresponding to preprocessed system operation characteristic data of "compressor frequency 50Hz, energy consumption during production period 100kWh / h". The Pearson correlation coefficient between the two is calculated. The correlation coefficient between compressor frequency and the predicted energy efficiency ratio is 0.82 (strong positive correlation), indicating that when the frequency increases within a reasonable range (30-70Hz), the energy efficiency ratio is optimized simultaneously, consistent with the design principle of "energy consumption prediction model with energy efficiency ratio as the core".

[0062] The correlation coefficient between the energy consumption variation characteristics during different time periods (production / shutdown energy consumption ratio of 3.3:1) and the predicted energy efficiency ratio is 0.75 (moderately strong positive correlation), indicating that even with reasonable increases in energy consumption during production periods, the energy efficiency ratio can still maintain a high level, validating the effectiveness of the prediction model. Finally, the correlation analysis results are generated, labeling the correlation coefficients and correlation levels between each feature and the prediction results (≥0.8 indicates strong correlation, 0.6-0.79 indicates moderate correlation).

[0063] A sliding window with a duration of 1 hour and a step size of 30 minutes was set to analyze the time-series correlation between "equipment operating parameter fluctuation characteristics" and "energy consumption prediction values" over a continuous 8-hour period. When the compressor frequency fluctuation value within the window is ≤5Hz (stable characteristic), the energy consumption prediction value is stable between 3.1 and 3.3, with a deviation ≤0.2. When the compressor frequency fluctuation value within the window is >8Hz (high fluctuation characteristic), the energy consumption prediction value drops to 2.8-3.0, with a deviation >0.3. Based on this, a time-series correlation pattern of "high fluctuation characteristic → increased prediction value deviation" was generated and incorporated into the operating data-energy consumption prediction correlation characteristic information to provide a basis for subsequent parameter optimization.

[0064] The preprocessed system operation characteristic data and supply-demand balance adjustment parameters are fused together using a weighted fusion algorithm to generate integrated operation data-supply-demand balance adjustment characteristic information. Supply-demand balance adjustment parameters include the output values ​​of the digital twin three-dimensional heat load model (e.g., heat load of 80kW for workshop A and 50kW for workshop B); cross-regional cooling capacity scheduling parameters (e.g., redundant cooling capacity of 20kW in workshop A, scheduled to workshop B); and condensation heat recovery parameters (e.g., the heat value of 15kW for waste heat conversion into 75℃ process hot water).

[0065] Weights are assigned to each parameter (heat load model value weight 0.5, cooling capacity scheduling parameter weight 0.3, condensing heat recovery parameter weight 0.2), and integrated with the pre-processed "energy consumption period variation characteristics" (energy consumption of workshop A during production period is 100 kWh / h). The weighted value is calculated as follows: heat load 80 kW × 0.5 + cooling capacity scheduling 20 kW × 0.3 + condensing heat recovery 15 kW × 0.2 = 40 + 6 + 3 = 49. Combined with the energy consumption period characteristic value (100 kWh / h), the integrated result "energy consumption-heat load matching degree 49%" is generated (if the matching degree is ≥60%, it is marked as "supply and demand balance characteristic"; otherwise, it is "supply and demand imbalance characteristic"). For the "supply and demand imbalance characteristic" (e.g., matching degree 49%), a further adjustment suggestion of "needing to increase cooling capacity scheduling by 10 kW" is generated and incorporated into the operation data-supply and demand adjustment integrated characteristic information, consistent with the functional innovation of "cross-regional cooling capacity scheduling".

[0066] The system categorizes and processes the correlation features between operational data and energy consumption prediction, as well as the integrated features between operational data and supply and demand regulation. It then filters these correlation features using precise energy consumption control parameters to generate correlation data for the energy consumption control dimension. Conversely, it filters the integrated features using system-level energy-saving optimization parameters to generate correlation data for the system energy-saving dimension. Finally, it categorizes and filters the aforementioned correlation and integrated features using both "precise energy consumption control parameters" and "system-level energy-saving optimization parameters" (two types of functional parameter groups) to generate correlation data for both the energy consumption control dimension and the system energy-saving dimension, ensuring a precise match between the data and control objectives.

[0067] The parameters for precise energy consumption control are set based on the goal of "precise energy consumption control," including "single equipment energy consumption threshold (e.g., compressor single unit energy consumption ≤ 20 kWh / h)" and "allowable range of regional energy consumption deviation (±5%)." From the operational data-energy consumption prediction correlation feature information, features such as a compressor frequency-energy efficiency ratio correlation coefficient of 0.82 (strong correlation) and "single equipment energy consumption 18 kWh / h ≤ 20 kWh / h" are selected to generate "equipment energy consumption-predicted energy efficiency ratio compliance correlation data." From the operational data-supply and demand adjustment integration feature information, features such as "regional energy consumption deviation 3% ≤ ±5%" and "supply and demand matching degree improved to 65% after cooling capacity scheduling" are selected to generate "regional energy consumption-cooling capacity scheduling precise control data." These two types of data together constitute the energy consumption control dimension correlation data, focusing on refined energy consumption control at the "single equipment-region" level.

[0068] System-level energy-saving optimization parameters include "annual energy efficiency improvement ≥3%" and "monthly power saving ≥10,000 kWh for abnormal production lines". From the operational data-energy consumption prediction correlation feature information, features such as "correlation coefficient between production period energy consumption and predicted energy efficiency ratio 0.75" and "annual energy efficiency improvement ≥3%" are selected to generate "period energy consumption-system energy efficiency optimization correlation data". From the operational data-supply and demand regulation integration feature information, features such as "power saving of 12,000 kWh / month in Workshop B after cross-regional cooling capacity scheduling ≥10,000 kWh" and "condensing heat recovery reduces boiler costs by 30%" are selected to generate "cooling capacity scheduling-energy-saving effect correlation data". These two types of data together constitute the system energy-saving dimension correlation data, focusing on improving the overall system's energy-saving effectiveness.

[0069] By integrating data related to energy consumption management and system energy saving, and through data splicing and feature aggregation processing, correlation data between industrial air conditioning energy saving effect and system operating status is generated. This correlation data characterizes the inherent correlation and quantitative relationship between industrial air conditioning energy consumption level, equipment operating status, and energy-saving optimization effect. The correlation data from the energy consumption management and system energy saving dimensions are spliced ​​along the time dimension, and core correlation features are summarized to generate correlation data between industrial air conditioning energy saving effect and system operating status. This data must clearly characterize the inherent patterns of energy consumption, equipment operation, and energy saving effect.

[0070] The time dimension is stitched together, with "1 hour" as the time unit. The "equipment energy consumption - predicted energy efficiency ratio compliance correlation data" (e.g., 10:00-11:00, compressor energy consumption 18kWh / h, energy efficiency ratio 3.2) and "cooling capacity scheduling - energy saving effect correlation data" (simultaneous cooling capacity scheduling 20kW, B workshop power saving 100kWh) are stitched together to ensure the consistency of data time sequence.

[0071] The core features are summarized, and key quantitative relationships are extracted from the spliced ​​data. For example, "When the compressor frequency is stable at 45-55Hz (fluctuation ≤5Hz), the predicted energy efficiency ratio is ≥3.0, corresponding to a cooling capacity dispatch of 15-25kW, regional energy consumption deviation ≤3%, and monthly electricity savings of 12,000 kWh." "During peak energy consumption periods (120kWh / h), 15kW of waste heat is converted through condensation heat recovery to supplement the process hot water demand, reducing boiler energy consumption by 20% and improving system energy efficiency by 4%."

[0072] The correlation between energy-saving effects of industrial air conditioning and system operation status is specifically manifested as follows: Equipment operation-energy consumption correlation: Compressor frequency 45-55Hz (stable fluctuation) → Energy efficiency ratio 3.0-3.5 → Single equipment energy consumption 15-20kWh / h. Supply and demand adjustment-energy saving correlation: Cross-regional cooling capacity dispatch 15-25kW → Regional supply and demand matching degree ≥65% → Monthly electricity saving of 12,000-15,000 kWh for abnormal production lines. Time period characteristics-system energy efficiency correlation: Energy consumption during production period 100-120kWh / h → Condensation heat recovery 15-20kW → Annual system energy efficiency improvement of 4-5%. This data comprehensively covers the correlation patterns of "equipment-energy consumption-energy saving".

[0073] S105 processes the associated data and parameter settings based on the visualization analysis of the energy management platform and the ISO50001 compliance audit standard to generate industrial air conditioning energy-saving optimization and system reliability evaluation results.

[0074] In one implementation, the visualization and analysis function of the energy management platform is used to perform visualization adaptation processing on the associated data. This involves presenting regional energy consumption differences through energy consumption distribution heatmaps, displaying parameter fluctuation trends through equipment operation curves, and comparing energy-saving magnitudes through optimization effect bar charts. Visual features are annotated on key data nodes to generate a visualization presentation of the associated data. Taking each workshop in the factory as a unit, a heatmap is generated based on the "total energy consumption of each workshop area" (e.g., Workshop A 120kWh / h, Workshop B 80kWh / h, Workshop C 50kWh / h) in the associated data. High-energy-consuming areas (Workshop A, energy consumption ≥100kWh / h) are marked in red, medium-energy-consuming areas (Workshop B, 50-99kWh / h) in yellow, and low-energy-consuming areas (Workshop C, ≤49kWh / h) in green. Key nodes in high-energy-consuming areas are annotated with features, such as "peak energy consumption of 150kWh / h from 12:00 to 14:00 in Workshop A, corresponding to full-load operation of production equipment," which is consistent with the function of "locating high-energy-consuming links," helping to quickly identify key energy-consuming control areas.

[0075] Based on the "compressor frequency and valve opening fluctuation characteristics" in the associated data, equipment operation trend curves are generated. For the compressor frequency curve, the horizontal axis represents time (0-24 hours), and the vertical axis represents frequency (30-70Hz). A daily frequency change curve is plotted, labeled "8:00 frequency rises from 40Hz to 50Hz (production starts, load increases)" and "20:00 frequency drops to 35Hz (production stops, load decreases)". A fluctuation warning threshold (±5Hz) is marked with a dashed line. When the curve exceeds the threshold (e.g., the frequency suddenly rises to 58Hz at 10:00, with a fluctuation of 8Hz), it is marked "abnormal fluctuation, PID control parameters need to be checked", which meets the requirements of "precise control at the edge control execution layer". For the valve opening curve, a chilled water valve opening curve is plotted simultaneously, labeled with the "opening and frequency linkage adjustment" feature (e.g., when the frequency rises to 50Hz, the opening increases from 50% to 60%), reflecting the coordinated operation pattern of the equipment.

[0076] Based on the associated data on "energy saving rate and electricity saving of abnormal production lines," a bar chart comparing the effects before and after optimization is generated. The vertical axis represents energy-saving related indicators (energy saving rate, electricity saving kWh / month), and the horizontal axis represents the optimization stage (before optimization and after optimization). The chart is marked with phrases such as "Monthly energy saving rate increased from 5% to 8% after optimization" and "Monthly electricity saving of abnormal production line B increased from 8,000 kWh to 12,000 kWh." For nodes meeting the standard (electricity saving ≥ 12,000 kWh), the chart is marked with "Meets the annual energy efficiency improvement benchmark of 3%-5%," visually demonstrating the optimization effectiveness of the "energy efficiency closed-loop evolution."

[0077] In accordance with the ISO 50001 compliance audit standard, the parameter settings were verified for compliance. The energy consumption statistics method, energy efficiency calculation standard, and data collection accuracy were checked one by one to determine whether they met the requirements of the ISO 50001 energy management system. Non-compliant items were marked with rectification directions, and parameter setting compliance verification results were generated. ISO 50001 requires that "energy consumption statistics must cover all energy-consuming equipment, distinguishing between production and non-production energy consumption." In verifying the "energy consumption statistics scope" in the parameter settings, if it was found that "non-production energy consumption for workshop lighting (approximately 10 kWh / h) was not separately counted," it was deemed non-compliant. The rectification direction was: "Add a lighting energy consumption collection module to separate non-production energy consumption from total energy consumption, ensuring that the statistical classification meets ISO 50001 requirements," which aligns with the goal of "precise energy consumption control."

[0078] ISO 50001 requires that "energy efficiency calculations must use an industry-standard formula (Efficiency Ratio = Cooling Capacity / Power Consumption)". The "Efficiency Ratio calculation logic of the energy consumption prediction model" in the system-level benchmark data is being verified. If it is found that "the model does not include cooling capacity parameters and only calculates energy efficiency based on power consumption," it is deemed non-compliant. The rectification direction is: "Add cooling capacity collection parameters (e.g., the cooling capacity of workshop A is 360kW collected by the refrigeration meter) to the AI-optimized decision-making layer energy consumption prediction model, and recalculate the Efficiency Ratio according to the standard formula (360kW / 120kWh=3.0)" to ensure that energy efficiency calculations are consistent with the standard.

[0079] The significance of the correlation between energy saving effect and system reliability was verified. Evaluation weights were defined for energy saving indicators and system reliability indicators in the correlated data. A weighted scoring method was used to calculate the overall compliance level, enhancing the objectivity of the evaluation results. The total weight was set at 100 points, with 60 points allocated to energy saving indicators (30 points for energy saving rate and 30 points for electricity saving on abnormal production lines): Energy saving rate ≥ 8% received 30 points, 6-7% received 20 points, and < 6% received 10 points; electricity saving ≥ 12,000 kWh / month received 30 points, 10,000-11,000 kWh received 20 points, and < 10,000 kWh received 10 points.

[0080] The system reliability index has a weight of 40 points (equipment failure rate 20 points, uninterrupted temperature control duration 20 points): failure rate ≤1% / month gets 20 points, 1.1-2% gets 15 points, and >2% gets 10 points; uninterrupted temperature control duration ≥7×24 hours gets 20 points, and 10 points are deducted for any interruption (such as failure of the main cold source and failure to switch in time), in accordance with the requirement of "dual cold source redundancy 7×24-hour temperature control".

[0081] Taking a certain month's related data as an example, for energy-saving indicators: energy saving rate 8% (30 points), electricity saving 12,000 kWh (30 points), total 60 points. For system reliability indicators: equipment failure rate 0.8% (20 points), uninterrupted temperature control duration 7×24 hours (20 points), total 40 points. The comprehensive score is 60+40=100 points, marked "Comprehensive compliance, energy saving and reliability are synergistically optimized". If in a certain month "uninterrupted temperature control is interrupted by 1 hour due to a delay in switching to a backup cold source", then 10 points will be deducted from the reliability indicator, the comprehensive score will be 90 points, and it will be marked "Redundancy switching response speed needs to be optimized".

[0082] By integrating and analyzing the results of visualization of related data, compliance verification of parameter settings, and significant verification of the correlation between energy saving and reliability, an evaluation result for energy saving optimization and system reliability of industrial air conditioning is generated. This evaluation result includes quantitative values ​​for energy consumption optimization, system operational reliability parameters, and ISO50001 compliance levels. The evaluation result characterizes the comprehensive performance of the industrial air conditioning system in terms of energy consumption control, operational stability, and compliance. The visualization results provide intuitive "problem location and effect presentation" (such as high energy consumption areas and abnormal equipment fluctuations), the compliance verification results provide "standard compliance judgment" (such as whether energy consumption statistics and data accuracy are compliant), and the significant verification results provide "comprehensive compliance quantitative basis" (coordinated score of energy saving and reliability). The combination of these three results provides a comprehensive evaluation of system performance from three levels: "phenomenon-standard-quantification".

[0083] For energy consumption optimization quantification values: integrate visualization and related data, and mark "monthly total energy saving rate of 8% (exceeding the annual benchmark of 3%-5%)", "cumulative monthly electricity saving of 15,000 kWh for abnormal production lines (3,000 kWh in Workshop A and 12,000 kWh in Workshop B)" and "energy consumption of high-energy-consuming areas (Workshop A) reduced by 20%", to reflect the energy-saving effect.

[0084] Regarding system operational reliability parameters: Based on significance verification and compliance checks, the following are marked: "Equipment failure rate 0.8% (≤1% target)", "Fault prediction response time 4 minutes (<5 minutes requirement)" and "Dual cold source redundancy switching success rate 100%", reflecting the system's operational stability.

[0085] Regarding the ISO50001 compliance level: Based on the compliance verification results (only the electricity meters in Workshop A need to be replaced, and the rest are compliant), it is determined that "the compliance level is A (no more than 1 non-compliant item, and the direction of rectification is clear)", and it is noted that "the compliance level can be maintained after rectification".

[0086] The final evaluation results clearly state that "the system accurately identifies high-energy-consuming links and achieves significant energy savings in energy consumption control, its operational reliability meets safety requirements, and it basically complies with the ISO50001 compliance standard," comprehensively characterizing the overall performance of the industrial air conditioning system and providing direction for subsequent optimization.

[0087] In one implementation, such as Figure 2 As shown, this application also provides an industrial air conditioning energy-saving optimization control device based on multi-source data fusion, comprising:

[0088] The acquisition module 201 is used to acquire multi-source data acquisition parameters and system-level benchmark data. The multi-source data acquisition parameters include equipment parameters, environmental variables, user behavior and production plan data selected according to the operation requirements of industrial air conditioning. The system-level benchmark data includes multi-source data fusion layer processing standards, AI optimization decision layer operation benchmarks and edge control execution layer regulation benchmarks.

[0089] The processing module 202 is used to combine multi-source data fusion parameters and control strategy parameters at various levels, synchronously match data processing flow and control command execution logic, establish collaborative rules for multi-source data fusion and system control based on hierarchical processing technology and multimodal integration design, and generate energy-saving optimization constraints for the entire system. It processes multi-source data acquisition parameters, system-level baseline data, collaborative control logic, and dynamic strategy combination constraints to collect energy consumption data, equipment operation data, optimization effect data, and fault warning data. It processes the collected system operation data, combining functional parameters including energy consumption prediction model results and supply-demand balance adjustment parameters to generate correlation data between industrial air conditioning energy-saving effect and system operation status. The functional parameter groups include precise energy consumption control parameters and system-level energy-saving optimization parameters. Finally, it processes the correlation data and parameter settings based on the energy management platform's visualization analysis and ISO50001 compliance audit standards to generate industrial air conditioning energy-saving optimization and system reliability evaluation results.

[0090] This application provides an electronic device, which includes a first processor and a memory for storing executable instructions of the first processor; wherein the first processor is configured to execute the aforementioned energy-saving optimization control method for industrial air conditioning based on multi-source data fusion by executing the executable instructions, and has the same beneficial effects as the method.

[0091] This application provides a computer-readable storage medium storing a computer program. When the computer program is read and executed by a second processor, it implements the aforementioned energy-saving optimization control method for industrial air conditioning based on multi-source data fusion, and has the same beneficial effects as the method adopted, run, or implemented by the application program stored therein.

[0092] The various embodiments in this application are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments for evaluating the energy-saving optimization control method, electronic device, electronic device, and readable storage medium for industrial air conditioning based on multi-source data fusion are basically similar to the embodiments of the energy-saving optimization control method for industrial air conditioning based on multi-source data fusion described above, and are therefore described simply. Relevant parts can be referred to in the descriptions of the embodiments of the energy-saving optimization control method for industrial air conditioning based on multi-source data fusion described above.

Claims

1. An energy-saving optimization control method for industrial air conditioning based on multi-source data fusion, characterized in that, include: Acquire multi-source data acquisition parameters and system-level benchmark data. Multi-source data acquisition parameters include equipment parameters, environmental variables, user behavior and production plan data selected according to the operational needs of industrial air conditioning. System-level benchmark data includes multi-source data fusion layer processing standards, AI optimization decision layer computation benchmarks, and edge control execution layer regulation benchmarks. This paper combines multi-source data fusion parameters and control strategy parameters at each level, synchronously matches the data processing flow with the control command execution logic, and establishes collaborative rules for multi-source data fusion and system control based on hierarchical processing technology and multimodal integration design. It generates energy-saving optimization constraints for the entire system, including correlation analysis of multi-source data fusion parameters and control strategy parameters at each level, and generates adaptation simulation correlation analysis results for data fusion and control strategies. The multi-source data fusion parameters include Kalman filter redundancy removal parameters at the data layer, PCA dimensionality reduction / CCA subspace learning parameters at the feature layer, and parameters at the decision layer. The system generates multi-model voting parameters and dynamic load forecasting LSTM model parameters; it filters the correlation analysis results to generate a list of target simulation variables, including multi-source data fusion hierarchical processing parameters, AI optimization decision parameters, edge control execution parameters, data-control response timeliness parameters, and multi-modal data integration error parameters; it processes the correlation analysis results and the list of target simulation variables to generate parameter combination constraints, including hierarchical collaborative constraints conforming to hierarchical processing technology, time-series synchronization constraints of multi-source data fusion and system control, and functional adaptation constraints of multi-modal integration design. The data is processed based on multi-source data acquisition parameters, system-level baseline data, collaborative control logic, and dynamic strategy combination constraints to collect energy consumption data, equipment operation data, optimization effect data, and fault early warning data. The collected system operation data is processed and combined with functional parameters including energy consumption prediction model results and supply and demand balance adjustment parameters to generate data on the correlation between industrial air conditioning energy-saving effect and system operation status. Among them, the functional parameters are grouped into energy consumption precision control parameters and system-level energy-saving optimization parameters. The associated data and parameter settings are processed in conjunction with the visualization analysis of the energy management platform and the ISO50001 compliance audit standard to generate industrial air conditioning energy-saving optimization and system reliability evaluation results.

2. The method as described in claim 1, characterized in that, Processing is performed based on multi-source data acquisition parameters, system-level baseline data, collaborative control logic, and dynamic strategy combination constraints to collect energy consumption data, equipment operation data, optimization effect data, and fault early warning data, including: Extract and classify multi-source data acquisition parameters, system-level baseline data, collaborative control logic and dynamic strategy combination constraints to generate energy consumption data acquisition dimension information, equipment operation data monitoring item information, optimization effect data evaluation dimension information, and fault early warning data triggering basis information. Based on the technical specifications for monitoring the operation of industrial air conditioning systems, the energy consumption data collection dimension information, equipment operation data monitoring item information, optimization effect data evaluation dimension information, and fault early warning data triggering basis information are processed to generate energy consumption data collection accuracy matching information and equipment operation data real-time evaluation information. Based on the energy consumption data acquisition accuracy matching information and equipment operation data real-time evaluation information, the target data in the multi-source data acquisition parameters and system-level benchmark data are marked and filtered to generate operation data correlation filtering results; The results of correlation screening of operational data are integrated and quantified to generate industrial air conditioning operation data collection results information. The industrial air conditioning operation data collection results information is used to characterize the accurate collection level of the industrial air conditioning system's operating status, energy-saving optimization effect, and equipment health. It includes energy consumption data, equipment operation data, optimization effect data, fault warning data, and data collection accuracy threshold.

3. The method as described in claim 2, characterized in that, The collected system operation data is processed, and combined with functional parameters including energy consumption prediction model results and supply-demand balance adjustment parameters, data relating the energy-saving effect of industrial air conditioning to the system operation status is generated, including: The collected system operation data is preprocessed. After data cleaning and standardization, the core features of system operation are extracted to generate preprocessed system operation feature data. The core features of system operation include equipment operating parameter fluctuation characteristics and energy consumption time period variation characteristics. The preprocessed system operation characteristic data and energy consumption prediction model results are correlated and calculated. The linear correlation between the two is analyzed by Pearson correlation coefficient. The time series correlation pattern is captured by the sliding window method to generate operation data-energy consumption prediction correlation characteristic information. The preprocessed system operation characteristic data and supply and demand balance adjustment parameters are fused together, and a weighted fusion algorithm is used to integrate the data to generate operation data-supply and demand adjustment fused characteristic information. The system classifies and processes the correlation features between operational data and energy consumption prediction, and the integrated features between operational data and supply and demand regulation. It then combines these with energy consumption precision control parameters to filter correlation features and generate correlation data for the energy consumption control dimension. Finally, it combines these with system-level energy-saving optimization parameters to filter integrated features and generate correlation data for the system energy-saving dimension. By integrating data related to energy consumption control and system energy saving, and through data splicing and feature aggregation processing, data on the correlation between industrial air conditioning energy saving effect and system operating status is generated. This data is used to characterize the inherent correlation and quantitative relationship between industrial air conditioning energy consumption level, equipment operating status and energy saving optimization effect.

4. The method as described in claim 1, characterized in that, The associated data and parameter settings are processed based on the visualization analysis of the energy management platform and the ISO50001 compliance audit standard to generate industrial air conditioning energy-saving optimization and system reliability evaluation results, including: Based on the visualization analysis function of the energy management platform, the related data is visualized and adapted. The energy consumption distribution heat map presents the regional energy consumption differences, the equipment operation curve shows the parameter fluctuation trend, the optimization effect bar chart compares the energy saving range, the key data nodes are marked with visualization features, and the results of the visualization of related data are generated. In accordance with the ISO 50001 compliance audit standard, the parameter setting basis is verified for compliance. The energy consumption statistics method, energy efficiency calculation standard, data collection accuracy and other dimensions are checked one by one to determine whether they meet the requirements of the ISO 50001 energy management system. Non-compliant items are marked with rectification directions and parameter setting compliance verification results are generated. The significance of the correlation between energy saving effect and system reliability is verified by defining the evaluation weights of energy saving indicators and system reliability indicators in the correlation data, and the comprehensive compliance degree is calculated by weighted scoring method to enhance the objectivity of the evaluation results. By integrating and analyzing the results of visualization of related data, compliance verification of parameter settings, and significant correlation verification of energy saving and reliability, an evaluation result of energy saving optimization and system reliability of industrial air conditioning is generated, which includes quantitative values ​​of energy consumption optimization, system operation reliability parameters, and ISO50001 compliance level. The evaluation result is used to characterize the comprehensive performance of industrial air conditioning system in terms of energy consumption control, operational stability, and compliance.

5. An industrial air conditioning energy-saving optimization control device based on multi-source data fusion, characterized in that, The apparatus for implementing the method of claim 1 includes: The acquisition module is used to acquire multi-source data acquisition parameters and system-level benchmark data. The multi-source data acquisition parameters include equipment parameters, environmental variables, user behavior and production plan data selected according to the operating requirements of industrial air conditioning. The system-level benchmark data includes multi-source data fusion layer processing standards, AI optimization decision layer calculation benchmarks, and edge control execution layer regulation benchmarks. The processing module combines multi-source data fusion parameters and control strategy parameters at various levels, synchronously matching data processing flows with control command execution logic. Based on hierarchical processing technology and multimodal integration design, it establishes collaborative rules for multi-source data fusion and system control, generating energy-saving optimization constraints for the entire system. It processes multi-source data acquisition parameters, system-level baseline data, collaborative control logic, and dynamic strategy combination constraints, collecting energy consumption data, equipment operation data, optimization effect data, and fault warning data. It processes the collected system operation data, combining functional parameters including energy consumption prediction model results and supply-demand balance adjustment parameters to generate correlation data between industrial air conditioning energy-saving effects and system operating status. The functional parameter groups include precise energy consumption control parameters and system-level energy-saving optimization parameters. Finally, it processes the correlation data and parameter settings based on the energy management platform's visualization analysis and ISO50001 compliance audit standards, generating industrial air conditioning energy-saving optimization and system reliability evaluation results.

6. An electronic device, characterized in that, include: First processor; and memory for storing executable instructions of the first processor; Wherein, the first processor is configured to execute the industrial air conditioning energy-saving optimization control method based on multi-source data fusion as described in any one of claims 1 to 4 by executing the executable instructions.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the second processor, it implements the industrial air conditioning energy-saving optimization control method based on multi-source data fusion as described in any one of claims 1 to 4.

Images