A predictive maintenance and energy-saving linkage method, device, electronic equipment, and computer-readable storage medium for industrial air conditioning.

Abstract

This invention provides a predictive maintenance and energy-saving linkage method, device, electronic device, and computer-readable storage medium for industrial air conditioners, applied in the field of data processing technology. This application collects core operating and environmental parameters such as compressor start-stop frequency; then, using the Internet of Things and edge computing, it monitors the status of key components, constructs a model linking equipment degradation and the environment, and generates preliminary maintenance requirements and energy consumption warnings; next, it designs a tiered maintenance strategy using digital twins and cost models, outputting target maintenance instructions and energy-saving reports; subsequently, it combines thermal comfort and peak / valley electricity prices to dynamically adjust parameters and generate a linkage control scheme; further, it expands to a multi-device global collaborative strategy; finally, it performs multi-dimensional evaluation to generate equipment reliability levels and energy-saving benefit values, forming a closed-loop technical system to ensure equipment safety and energy conservation.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a predictive maintenance and energy-saving linkage method, apparatus, electronic device, and computer-readable storage medium for industrial air conditioners. Background Technology

[0002] In the industrial sector, industrial air conditioning is a key piece of equipment for ensuring a stable production environment, and its stable operation and high energy efficiency are of paramount importance. Currently, conventional industrial air conditioning management relies heavily on traditional timed control and basic environmental feedback mechanisms, which make it difficult to balance equipment maintenance and energy consumption optimization under complex and ever-changing operating conditions.

[0003] On the one hand, existing technologies have significant shortcomings in energy efficiency optimization. Traditional systems cannot accurately adjust air conditioning output based on real-time environmental changes, such as the flow of people in the workshop, equipment load, and external weather conditions, resulting in serious energy waste and failing to meet current energy conservation and emission reduction development needs. For example, when the equipment load in the workshop changes, the air conditioning still maintains fixed operating parameters, causing over-cooling or over-heating.

[0004] On the other hand, there is a lack of fault prediction and preventative maintenance methods. Faults are often only detected after they affect system operation, lacking the ability to anticipate potential problems based on operational data. This increases maintenance costs and impacts long-term equipment efficiency and production safety. Sudden failures of core components like compressors can lead to production stoppages and significant losses.

[0005] 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 therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] Other features and advantages of this application will become apparent from the following detailed description, or may be learned in part by practice of the invention.

[0007] According to one aspect of this application, a predictive maintenance and energy-saving linkage method, device, electronic device, and computer-readable storage medium for industrial air conditioning are provided, comprising: acquiring operating data and environmental parameters of each core component of the industrial air conditioning; processing the operating data and environmental parameters of the core components, and monitoring in real time the compressor lubricating oil level, heat exchanger fouling degree, motor winding insulation, and filter clogging status through an Internet of Things sensor network and edge computing nodes, constructing a correlation model between component degradation trends and environmental load, and generating preliminary maintenance needs and energy consumption anomaly warning information; processing the preliminary maintenance needs and energy consumption anomaly warning information, simulating the equipment operating status at different maintenance times through digital twin technology, designing a graded maintenance strategy based on a full life cycle cost model, and applying partial maintenance. The remaining lifespan threshold constraint and the objective function for optimal allocation of maintenance resources are used to generate target maintenance execution instructions and energy-saving potential analysis reports. Based on the corresponding adaptive algorithm, the target maintenance execution instructions and energy-saving potential analysis reports are processed. Combining indoor thermal comfort requirements and the grid peak-valley electricity pricing mechanism, air conditioning operating parameters are adjusted in real time to generate a linkage control scheme that balances maintenance effectiveness and energy consumption optimization. This linkage control scheme is then processed to generate a globally coordinated operation management strategy. Finally, preliminary maintenance needs and energy consumption anomaly warning information, target maintenance execution instructions, the linkage control scheme balancing maintenance effectiveness and energy consumption optimization, and the globally coordinated operation management strategy are processed to generate comprehensive industrial air conditioning operation status assessment information, including equipment reliability level and energy-saving benefit quantification values.

[0008] Another aspect of this application discloses a predictive maintenance and energy-saving linkage device for industrial air conditioners, comprising: an acquisition module for acquiring operating data and environmental parameters of various core components of the industrial air conditioner; a processing module for processing the operating data and environmental parameters of the core components of the industrial air conditioner, and monitoring in real time the compressor lubricating oil level, heat exchanger fouling degree, motor winding insulation, and filter clogging status through an Internet of Things sensor network and edge computing nodes, constructing a correlation model between component deterioration trends and environmental load, and generating preliminary maintenance needs and energy consumption anomaly warning information; processing the preliminary maintenance needs and energy consumption anomaly warning information, simulating the equipment operating status at different maintenance times through digital twin technology, designing a graded maintenance strategy based on a full life cycle cost model, and applying the remaining lifespan of the components. The system generates target maintenance execution instructions and energy-saving potential analysis reports based on threshold constraints and the objective function for optimal allocation of maintenance resources. It then processes these instructions and reports using a corresponding adaptive algorithm, adjusting air conditioning operating parameters in real time based on indoor thermal comfort requirements and the grid's peak-valley electricity pricing mechanism. This generates a coordinated control scheme that balances maintenance effectiveness and energy consumption optimization. Further processing of this coordinated control scheme generates a globally collaborative operation management strategy. Finally, the system processes preliminary maintenance needs and energy consumption anomaly warning information, target maintenance execution instructions, the coordinated control scheme balancing maintenance effectiveness and energy consumption optimization, and the globally collaborative operation management strategy to generate a comprehensive assessment of the industrial air conditioning operating status, including equipment reliability levels and quantified energy-saving benefits.

[0009] 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 predictive maintenance and energy-saving linkage method for an industrial air conditioner by executing the executable instructions.

[0010] 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 predictive maintenance and energy-saving linkage method for industrial air conditioning.

[0011] This application provides a predictive maintenance and energy-saving linkage method, device, electronic equipment, and computer-readable storage medium for industrial air conditioners. It collects core data such as compressor start-stop frequency through distributed sensing and edge computing; then constructs a correlation model between component degradation and environmental load to generate preliminary maintenance needs and energy consumption warnings; next, it outputs target maintenance instructions and energy-saving reports using digital twins and a full life-cycle cost model; subsequently, it dynamically adjusts parameters to generate a linkage control scheme based on thermal comfort and peak / valley electricity prices; further, it expands to a multi-device global collaborative strategy; finally, it generates reliability levels and energy-saving benefit values ​​through multi-dimensional evaluation and visually pushes them to multiple terminals, ensuring equipment safety, energy saving, and efficient collaboration throughout the process. The technical logic is closed-loop and feasible.

[0012] 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

[0013] Figure 1 A flowchart illustrating a predictive maintenance and energy-saving linkage method for an industrial air conditioner provided in an embodiment of this application is shown.

[0014] Figure 2 A schematic diagram of the structure of a predictive maintenance and energy-saving linkage device for an industrial air conditioner provided in an embodiment of this application is shown. Detailed Implementation

[0015] 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.

[0016] The following is combined with Figure 1 This application describes a predictive maintenance and energy-saving linkage method for industrial air conditioning systems 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.

[0017] In one embodiment, this application also proposes a predictive maintenance and energy-saving linkage method, apparatus, electronic device, and computer-readable storage medium for industrial air conditioning. Figure 1 A schematic flowchart of a predictive maintenance and energy-saving linkage method for an industrial air conditioner according to an embodiment of this application is shown.

[0018] S101, acquires operating data and environmental parameters of each core component of the industrial air conditioner.

[0019] In one implementation, a three-tiered architecture of "distributed sensor deployment + multi-protocol data transmission + edge node preprocessing" is used to achieve real-time and accurate acquisition of operating data and environmental parameters of core components of industrial air conditioning systems. This covers three core data dimensions: equipment operating status, environmental influencing factors, and energy consumption, providing a complete data foundation for subsequent component degradation analysis and energy consumption optimization. The data acquisition frequency is uniformly set to 1Hz (10Hz for special high-dynamic parameters such as refrigerant pressure) to ensure data timeliness and consistency. For compressor operating data acquisition, "start-stop frequency" is the core. Current transformers and status monitoring sensors are installed in the compressor control loop to collect compressor operating current and start-stop signals in real time. Edge computing nodes count the number of start-stop cycles per unit time to generate compressor start-stop frequency data. Simultaneously, auxiliary parameters such as compressor discharge temperature and suction pressure are collected to verify the validity of the start-stop frequency data. The data acquisition devices are Hall current sensors (model: ACS712) and temperature sensors (model: PT100), deployed at the compressor terminals and exhaust pipes. The sensors acquire current and temperature signals every 100ms. When the current jumps from "0A" to "above 80% of the rated current", it is determined as a start-up; when it drops from "above 80% of the rated current" to "0A", it is determined as a stop-up. The edge nodes count the number of start-ups and stop-ups every minute and generate the start-up and stop-up frequency (e.g., "5 times / hour"). If the start-up and stop-up frequency is still less than 1 time / hour when the exhaust temperature exceeds 120℃, a data anomaly marker is triggered and pushed synchronously to the operation and maintenance terminal.

[0020] The heat exchanger operation data acquisition focuses on "evaporator / condenser temperature". Distributed fiber optic temperature sensors are evenly distributed along the heat exchange tubes of the evaporator and condenser to collect the tube wall temperature in different areas. The overall temperature of the evaporator / condenser is generated by the temperature averaging algorithm. At the same time, the inlet and outlet refrigerant temperatures and flow rates of the heat exchanger are collected to calculate the heat exchange efficiency and help verify the rationality of the temperature data. The data acquisition equipment consists of a distributed fiber optic temperature sensor (model: DTS-800) and a refrigerant flow sensor (model: FS400). One monitoring point is placed every 50cm along the evaporator heat exchange tubes, for a total of 20 points. The fiber optic sensor collects the temperature of each monitoring point once per second. Edge nodes remove outliers from the 20 temperature data points (using the 3σ principle) and then average the data to generate the evaporator temperature (e.g., "7℃") and condenser temperature (e.g., "45℃"). Simultaneously, the refrigerant flow rate (e.g., "5m³ / h") is collected via the flow sensor. If the flow rate is lower than 60% of the rated value and the temperature deviation exceeds ±3℃, the temperature data is marked as "to be verified." The temperature data is compared with historical data from the same period. If the condenser temperature is more than 5℃ higher than the same period, it is preliminarily determined that scaling may be present.

[0021] The acquisition of wind turbine operating data focuses on "speed". A photoelectric encoder is installed at the end of the wind turbine motor shaft to calculate the wind turbine speed by counting photoelectric pulse signals. At the same time, the wind turbine operating current and vibration acceleration are collected to evaluate the matching of speed data with wind turbine load and avoid abnormal data such as "high speed when idling" or "high load but low speed". The data acquisition equipment consists of a photoelectric encoder (model: E6B2-CWZ6C) and a vibration sensor (model: ADXL345). The encoder is connected to the fan shaft via a coupling, and the vibration sensor is attached to the fan casing. The encoder outputs one pulse signal every 10ms. The edge nodes calculate the rotational speed (e.g., "1450 r / min") using the formula "pulse count / time × reduction ratio" and simultaneously acquire the vibration acceleration (e.g., "0.5g"). If the rotational speed is displayed as 1450 r / min but the vibration acceleration exceeds 1.2g, it is judged as "rotational speed data is normal but equipment is abnormal", and the vibration data is marked separately. The encoder resolution is set to 1000 pulses / revolution, and the rotational speed calculation error is controlled within ±1 r / min, meeting the parameter accuracy requirements for fan speed control.

[0022] Data acquisition for the refrigerant system focuses on "pressure." High-pressure and low-pressure sensors are installed at the evaporator inlet and condenser outlet of the refrigerant pipeline, respectively, to collect refrigerant pressure in real time. At the same time, the refrigerant liquid level (in the receiver tank) and the on / off status of the solenoid valve are also collected to ensure that the pressure data matches the system's operating phase (such as "cooling mode" and "defrosting mode"). The data acquisition devices are a pressure sensor (model: MPX5700) and a liquid level sensor (model: YXC-100). The high-pressure sensor is deployed on the condenser outlet pipe (pressure range: 0-4MPa), and the low-pressure sensor is deployed on the evaporator inlet pipe (pressure range: 0-1.6MPa). The pressure sensor collects the pressure value once every 500ms. When the system is in "cooling mode", the normal range of low pressure is 0.4-0.6MPa, and the normal range of high pressure is 1.8-2.2MPa. If the low pressure of 0.2MPa is collected, the solenoid valve switch status is checked simultaneously. If the solenoid valve is in the "closed" state, it is determined as "normal shutdown pressure"; otherwise, it is marked as "leakage warning pressure". The pressure data is correlated with the compressor start-stop frequency. If the low pressure does not rise to 0.4MPa within 30 seconds when the compressor starts, a preliminary "insufficient refrigerant" warning is triggered.

[0023] Temperature and humidity parameters are acquired with "indoor and outdoor temperature and humidity" as the core. Integrated temperature and humidity sensors are used. Indoor sensors are evenly distributed in the industrial workshop at a density of "500m² / sensor". Outdoor sensors are placed near the air conditioner outdoor units (avoiding direct sunlight and in well-ventilated areas) to collect temperature and humidity data in real time. At the same time, the heat dissipation power of the production equipment in the workshop is collected to analyze the correlation between environmental temperature and humidity and load. The data acquisition equipment is a temperature and humidity sensor (model: SHT30). Eight indoor sensors are deployed at different heights in the workshop (1.5m for personnel activity area, 3m for equipment top area). Outdoor sensors are deployed 1m to the side of the air conditioner unit. The sensors collect temperature and humidity data every minute. The indoor sensors take the average of the eight points as the "workshop average temperature and humidity" (e.g., "25℃, 60%RH"). The outdoor sensors directly output the "ambient temperature and humidity" (e.g., "32℃, 75%RH"). If the sensor readings in a certain area of ​​the workshop deviate from the average by more than ±2℃, the operating status of the production equipment in that area is checked. If the equipment is operating at full load, it is determined as "local load causing abnormal temperature and humidity," and the data for that area is marked separately. Indoor temperature and humidity are used for subsequent thermal comfort constraint determination, while outdoor temperature and humidity are used to correct air conditioning operating parameters (e.g., for every 5℃ increase in outdoor temperature, the condenser fan speed increases by 10%).

[0024] Load parameter acquisition is centered on the "load change curve". Through data interaction between the industrial air conditioning main control system and the workshop production management system (MES), the operating status of production equipment in the workshop (such as "number of units in operation and operating power") is obtained. Combined with the indoor temperature and humidity change rate, the load change curve is constructed. At the same time, an air volume sensor is installed in the main air intake duct of the air conditioning system to collect the intake air volume to assist in verifying the load data. The data acquisition devices are an airflow sensor (model: FS700) and a data exchange gateway (supporting the OPCUA protocol). The gateway is deployed in the air conditioning main control cabinet and establishes communication with the MES system. Every 5 minutes, the gateway obtains the production equipment operation data (e.g., "10 machine tools are running, total power 500kW") from the MES system and simultaneously collects the intake air volume (e.g., "10000m³ / h") through the airflow sensor. With time as the horizontal axis and "total equipment power + airflow" as the vertical axis, a load change curve is generated (e.g., "9:00 load 300kW, 11:00 load 500kW, 14:00 load 450kW"). If the load curve shows "500kW" but the airflow is lower than 8000m³ / h, it is determined that "the load data does not match the air conditioning airflow" and an airflow sensor calibration reminder is triggered.

[0025] Energy consumption parameter acquisition is centered on "energy consumption metering data". A smart energy meter is installed in the main power supply circuit of the industrial air conditioner to collect total energy consumption data. At the same time, energy meters are installed in the sub-circuits of major energy-consuming components such as compressors, fans and water pumps to collect sub-item energy consumption data. The difference between "total energy consumption and sub-item energy consumption" is calculated by edge computing nodes to verify data integrity and avoid metering omissions. The data acquisition device is a smart energy meter (model: DTZY341). The main energy meter is deployed at the incoming line of the air conditioning distribution box, and the sub-item energy meters are deployed in the power supply circuits of the compressor, fan, and water pump, respectively. The energy meters collect energy consumption data (unit: kWh) every 15 minutes. The total energy consumption data (e.g., "total energy consumption in 15 minutes: 12 kWh") is summed and compared with the sub-item energy consumption data (compressor: 8 kWh, fan: 3 kWh, water pump: 1 kWh). If the difference exceeds 0.5 kWh, it is judged as "sub-item metering abnormality", and the abnormal circuit energy meter is marked. In addition to the cumulative energy consumption, parameters such as voltage, current, and power factor are collected simultaneously (e.g., "voltage 380V, current 25A, power factor 0.92") for subsequent energy consumption anomaly analysis (e.g., if the power factor is lower than 0.85, it is judged as "excessive reactive power loss leading to energy consumption anomaly").

[0026] The sensors and edge computing nodes communicate using the LoRaWAN protocol (transmission distance ≤ 1km) or RS485 bus (transmission distance ≤ 100m). The edge computing nodes and the cloud platform transmit data via 5G / industrial Ethernet to ensure a data transmission packet loss rate of less than 0.1%. The edge nodes perform "outlier removal (3σ principle), missing value completion (linear interpolation), and format standardization (unified to JSON format)" on the collected data. For example, if the refrigerant pressure data is missing at a certain moment, it is completed by linear interpolation of the data from the previous 10 seconds and the next 10 seconds to avoid data gaps in subsequent modeling. The preprocessed data is stored in a hierarchical manner as "real-time database (stores data within 1 hour for real-time monitoring) + historical database (stores data for 1 year for trend analysis)". The real-time database is updated at the same frequency as the collection frequency, and the historical database is compressed and stored according to "hourly average".

[0027] S102 processes the operating data and environmental parameters of the core components of the industrial air conditioner. Through the Internet of Things sensor network and edge computing nodes, it monitors the compressor lubricating oil level, heat exchanger scaling degree, motor winding insulation and filter blockage status in real time. It constructs a correlation model between component deterioration trend and environmental load, and generates preliminary maintenance needs and energy consumption anomaly warning information.

[0028] In one implementation, operational data and environmental parameters of core components of an industrial air conditioner are processed. A distributed monitoring system is constructed using an Internet of Things (IoT) sensor network and edge computing nodes. The operational data of the core components and environmental parameters are fused and calculated using a dynamic weight allocation mechanism to monitor in real time the compressor lubricating oil level, heat exchanger fouling degree, motor winding insulation, and filter clogging status. A three-layer architecture—IoT sensing layer, edge computing layer, and data fusion layer—is used to build the distributed monitoring system: the IoT sensing layer deploys multiple types of sensors (such as oil level, temperature, and pressure sensors) to achieve real-time acquisition of operational data and environmental parameters of core components; the edge computing layer deploys localized edge nodes (such as industrial gateways, model: MG300) to handle data preprocessing (such as preliminary screening of outliers and format conversion) and low-latency computation tasks; the data fusion layer, through a dynamic weight allocation mechanism, achieves cross-dimensional correlation between core component data and environmental parameters, ensuring that the monitoring data accurately reflects the actual operating status of the equipment, with overall data transmission and processing latency controlled within 500ms.

[0029] The weights of parameters are dynamically adjusted based on their impact on equipment status. The basic weight for core component operating data (such as compressor lubricating oil level and refrigerant pressure) is set to 0.6, and the basic weight for environmental parameters (such as indoor temperature and humidity and workshop load) is set to 0.4. When the environmental load fluctuation coefficient (fluctuation coefficient = current load / rated load) exceeds 1.2 (i.e., the load exceeds the rated value by 20%), the weight of environmental parameters is automatically increased to 0.5, and the weight of core component data is decreased to 0.5, to avoid underestimating the impact of environmental factors on equipment status under high load conditions. Taking an industrial air conditioner in a workshop as an example, the input data includes the measured compressor lubricating oil level at 60% (normal range 50%-80%), indoor temperature at 28℃ (normal range 22-26℃), workshop load at 120% (rated load 100%), and condenser temperature at 48℃ (normal range 40-50℃). First, parameter standardization is performed. Standardized value = (measured value - minimum parameter value) / (maximum parameter value - minimum parameter value). For example, the standardized value of compressor lubricating oil level = (60-50) / (80-50) ≈ 0.33. Since the indoor temperature exceeds the normal range, the standardized value is calculated based on the upper limit of 26℃ = (28-22) / (26-22) = 1.5. Then, it is corrected to 1.2 according to the rule of "when it exceeds the normal range, the standardized value is 1.2". The standardized value of workshop load = (120-80) / (120-80) = 1.0 (the normal load range is 80%-120%). The standardized value of condenser temperature = (48-40) / (50-40) = 0.8. Because the workshop load exceeded the rated value, the weight of environmental parameters was increased to 0.5, and the weight of core component data was decreased to 0.5. Using the weighted summation formula (fusion value = Σ(parameter standardized value × corresponding weight)), the final fusion value was calculated as (0.33 × 0.5) + (1.2 × 0.5) + (1.0 × 0.5) + (0.8 × 0.5) = 1.565. When the fusion value exceeds 1.5, the system triggers key monitoring, and the tracking frequency of compressor lubricating oil level and indoor temperature will be specifically increased (from 1Hz to 2Hz).

[0030] A capacitive oil level sensor (model: CYW-100) is deployed on top of the compressor oil tank. Oil level data is calculated based on the linear relationship between the sensor's output capacitance value and the oil level height, with a sampling frequency of 1Hz. Example: When the sensor acquires a capacitance value of 1800pF, according to the calibration curve that "1500pF corresponds to 50% oil level and 2100pF corresponds to 80% oil level," the oil level is calculated to be 60%. If the oil level is continuously monitored to drop from 60% to 58% for 5 minutes, while the workshop load remains at 100%, the system determines this as a "slow downward trend in oil level," and synchronously correlates it with the compressor operating current (1.1 times the rated current). This eliminates the possibility of "abnormal current causing false oil level measurement," and initially determines it as normal oil consumption.

[0031] An ultrasonic thickness sensor (model: UT3000) was used to evenly distribute three monitoring points along the heat exchanger tubes (located at the inlet, middle, and outlet of the tubes, respectively). The degree of scaling was estimated by measuring the rate of change in tube wall thickness. For every 0.2 mm increase in thickness, the scaling level increased by one grade (divided into three grades: light, moderate, and heavy). Example: The initial tube wall thickness was 3 mm. After 3 months, the thickness was measured at the inlet monitoring point (3.4 mm), the middle monitoring point (3.3 mm), and the outlet monitoring point (3.2 mm). The average thickness change rate was calculated as (3.3-3) / 3 × 100% = 10%, corresponding to a scaling level of "light scaling". The condenser temperature was simultaneously monitored and found to have increased by 3°C compared to before scaling (from 45°C to 48°C). The correlation model between temperature increase and scaling level was used to verify that scaling had affected the heat exchange efficiency, providing a basis for determining subsequent maintenance needs.

[0032] An insulation resistance tester (model: MI2077) was integrated into the fan motor control circuit, collecting insulation resistance values ​​every 5 minutes, with a normal range of ≥100MΩ. In one instance, an insulation resistance value of 80MΩ was collected, below the threshold. First, the workshop ambient humidity was checked (60%RH, normal range 40%-70%) to rule out the influence of abnormal humidity. Then, the fan operating current (1.2 times the rated current) was correlated, revealing a persistently high current. This indicated that the insulation degradation was caused by aging of the motor windings, rather than external environmental factors, thus avoiding misjudgment of maintenance needs.

[0033] Differential pressure sensors (model: DPT100) are installed in the inlet and outlet ducts of the filter. The degree of blockage is determined by measuring the air pressure difference. A normal differential pressure is ≤50Pa, a differential pressure of 51-80Pa indicates moderate blockage, and a differential pressure >80Pa indicates severe blockage. For example, if a differential pressure of 80Pa is collected, first check the fan speed (1450r / min, normal range 1400-1500r / min) to rule out the possibility that the abnormal differential pressure is caused by excessive speed. Then calculate the ventilation volume attenuation rate. According to the "correlation formula between differential pressure and ventilation volume" (ventilation volume attenuation rate = (normal differential pressure - measured differential pressure) / normal differential pressure × 100%), the attenuation rate is 15%, which is judged as "moderate blockage". It is clear that filter cleaning and maintenance should be arranged within one week.

[0034] Utilizing a component-environment dynamic correlation modeling framework, a three-level network structure is employed: input layer multi-source data denoising and feature selection, hidden layer spatiotemporal correlation learning, and output layer demand and early warning mapping. This structure constructs a model linking component degradation trends with environmental load. Specifically, the hidden layer spatiotemporal correlation learning incorporates an LSTM network to capture the temporal correlation between component degradation and environmental load, and combines this with an attention mechanism to strengthen the weights of key influencing factors. The model is constructed using this three-level network structure: the input layer is responsible for eliminating data noise and selecting key features, providing high-quality input to the model; the hidden layer, through an LSTM network and attention mechanism, captures the temporal correlation between component degradation and environmental load, as well as key influencing factors; and the output layer transforms the model's calculation results into directly applicable maintenance requirements and early warning information, ensuring a model prediction accuracy ≥90%, meeting the precision requirements of predictive maintenance. The collected data was processed using a wavelet threshold denoising algorithm. Taking the vibration acceleration data of a fan as an example, the original data contained 5Hz high-frequency interference (caused by vibration transmission from other equipment in the workshop). By decomposing the data into five layers of wavelet coefficients and setting a denoising threshold of 0.05g for the high-frequency coefficients, the data was reconstructed after removing the interference signals. The fluctuation amplitude of the vibration acceleration data was reduced from ±0.2g to ±0.08g, restoring the true vibration state of the fan and avoiding the model's misjudgment of equipment abnormalities due to noise. Features strongly correlated with component degradation were screened using Pearson correlation coefficients. Taking compressor degradation analysis as an example, the initial input features included six items: compressor start-stop frequency, lubricating oil level, exhaust temperature, indoor temperature, outdoor temperature, and workshop load. The correlation coefficients between each feature and the compressor deterioration level were calculated. The correlation coefficients for indoor temperature (0.25) and outdoor temperature (0.28) were both below the screening threshold of 0.3 and were therefore removed. The four core features of compressor start-stop frequency (0.72), lubricating oil level (0.85), exhaust temperature (0.78), and workshop load (0.65) were retained to reduce the impact of irrelevant features on the model's calculation efficiency.

[0035] A three-layer LSTM network (32 input neurons, 64 hidden neurons, and 16 output neurons) was constructed. It was fed with nearly 72 hours of time-series data on compressor lubricating oil level, workshop load, and exhaust temperature. Through gating mechanisms (input gate, forget gate, and output gate), the network learned the patterns of parameter changes over time. For example, given three consecutive days of time-series data: Day 1: Compressor lubricating oil level 65%, workshop load 80%, exhaust temperature 85℃; Day 2: Lubricating oil level 62%, load 100%, exhaust temperature 88℃; Day 3: Lubricating oil level 60%, load 120%, exhaust temperature 90℃. The LSTM network learned a time-series correlation: "For every 20% increase in workshop load, the compressor lubricating oil level decreases by an average of 2% per day, and the exhaust temperature increases by an average of 3℃ per day." Based on this trend, it predicted that the lubricating oil level would drop to 52% (close to the lower limit of 50%) after 7 days, providing data support for determining maintenance timing. An attention mechanism was introduced into the LSTM network to assign higher weights to key features affecting component degradation. Taking the heat exchanger fouling model as an example, the input features include four items: condenser temperature, cooling water flow rate, workshop load, and ambient humidity. The weight of each feature is calculated through the attention mechanism. The results show that the weight of condenser temperature is 0.35, cooling water flow rate is 0.28, workshop load is 0.25, and ambient humidity is 0.12. The weight of condenser temperature is significantly higher than that of other features. The model automatically strengthens the influence of this feature on the fouling trend, reducing the fouling level prediction error from ±0.5 level to ±0.2 level, thus improving the prediction accuracy.

[0036] The rationality of the model output results is verified by combining the safety operation threshold constraints of the core components of industrial air conditioning. Abnormal data caused by sensor drift and data transmission interference are eliminated by the 3σ principle. At the same time, the environmental factors are dynamically weighted based on the environmental load fluctuation coefficient to generate preliminary maintenance requirements and energy consumption anomaly warning information, including component maintenance priority, energy consumption anomaly type and risk level. A dual mapping table of "component deterioration level - maintenance requirement type" and "environmental load - warning level" is established. The component deterioration level is divided into Level 1 (slight deterioration: performance decrease ≤10%), Level 2 (moderate deterioration: performance decrease 10%-20%), and Level 3 (severe deterioration: performance decrease >20%), which correspond to "routine maintenance requirements", "priority maintenance requirements" and "emergency maintenance requirements", respectively. The environmental load warning level is divided according to the fluctuation coefficient: 1.0-1.2 is "blue warning", 1.2-1.5 is "yellow warning" and >1.5 is "orange warning". The model outputs a compressor degradation level of 2 (performance decrease of 15%) and a workshop load fluctuation coefficient of 1.3 (yellow warning). Preliminary maintenance requirements and warning information are generated through a mapping table: the maintenance requirement is "prioritize compressor maintenance, perform lubricant replenishment and winding insulation testing", and the warning information is "workshop load is too high (130% of rated value), pay attention to compressor exhaust temperature, and avoid prolonged high load operation to prevent accelerated degradation". At the same time, the maintenance completion deadline is specified to be within 7 days, and the warning continuous monitoring frequency is increased to 2Hz.

[0037] Setting safe operating thresholds for each core component, such as compressor lubricating oil level threshold of 50%-80%, motor winding insulation resistance threshold of ≥100MΩ, and refrigerant high pressure threshold of ≤2.5MPa. If the model outputs "compressor lubricating oil level 45%", comparing it with the thresholds reveals it is below the lower limit. Tracing back to the data acquisition stage, it was found that the oil level sensor was drifting (oil contamination causing the collected value to be too low). After cleaning the sensor, the actual oil level was collected at 60%, correcting the model output and avoiding misjudgments of maintenance needs due to sensor failure. Taking indoor temperature data as an example, nearly 100 sets of data were collected, with a mean of 24℃ and a standard deviation of 1.2℃. Based on the 3σ principle, the normal data range was determined to be from 24-3×1.2=20.4℃ to 24+3×1.2=27.6℃. In one instance, an indoor temperature of 30℃ was collected, which exceeded the 3σ range and was identified as abnormal data. Upon checking the sensor deployment location, it was found that the anomaly was caused by data transmission interference (the sensor was exposed to direct sunlight, which is an environmental interference-induced anomaly). This data was removed, and data collected by a backup sensor (deployed in a shady place) at 25℃ was used to ensure that the data input into the model was true and valid.

[0038] The weights of environmental factors in the model are adjusted based on the environmental load fluctuation coefficient. When the fluctuation coefficient is <0.8 (low load condition), the weight of environmental factors is reduced from the base value of 0.4 to 0.3, while the weight of core component data is increased to 0.7, to avoid excessive influence of environmental factors on the model results under low load. When the fluctuation coefficient is 1.0-1.2 (normal load), the weights remain at the base value. When the fluctuation coefficient is >1.2 (high load), the weights of environmental factors are increased to 0.5. Example: Due to production plan adjustments, the load in a workshop suddenly increased from 90% (fluctuation coefficient 0.9) to 130% (fluctuation coefficient 1.3). The system automatically increased the weights of environmental factors (indoor temperature, load) from 0.4 to 0.5 and recalculated the model output results. Before the adjustment, the model predicted an energy consumption anomaly rate of 5%. After the adjustment, considering the impact of high load on energy consumption, the predicted energy consumption anomaly rate was corrected to 8%, which is closer to the actual operating conditions and provides accurate basis for energy consumption anomaly early warning.

[0039] The maintenance priority of components is determined using a scoring method of "deterioration level × component impact coefficient". The component impact coefficient is set according to the importance of the equipment (compressor 1.2, heat exchanger 1.0, fan 0.8, filter 0.6). Example: In an air conditioning system, the compressor has a deterioration level of 2 (score 2 × 1.2 = 2.4), the heat exchanger has a deterioration level of 1 (1 × 1.0 = 1.0), the filter has a clogging level of 2 (2 × 0.6 = 1.2), and the fan has a deterioration level of 1 (1 × 0.8 = 0.8). The maintenance priority is sorted from high to low score as "compressor > filter > heat exchanger > fan". The initial maintenance requirements are generated as follows: "Prioritize the maintenance of the compressor within 7 days (replenish lubricating oil and check insulation), followed by cleaning the filter within 10 days. The heat exchanger and fan should be maintained according to the regular cycle (30 days)."

[0040] By combining energy consumption data with model output results, the type of energy consumption anomaly is determined and the risk level is classified. Example: The measured energy consumption during peak hours (10:00-12:00) is 80kWh, exceeding the historical average of 60kWh by 33%. Correlation model data reveals "slight scaling of the heat exchanger (heat exchange efficiency decreases by 8%) + excessively high indoor temperature setting (26℃→28℃)", thus classifying the energy consumption anomaly as "peak overconsumption caused by decreased heat exchange efficiency and unreasonable parameter settings". According to the rule that "energy consumption exceeding the limit by 10%-20% is a yellow risk, 20%-30% is an orange risk, and >30% is a red risk", the risk level is classified as "red", and an energy consumption anomaly warning message is generated: "Peak energy consumption exceeds the historical average by 33% (red risk). It is recommended to clean the heat exchanger (expected to improve heat exchange efficiency by 8%) and lower the indoor temperature to 26℃, which is expected to reduce peak energy consumption by 15%-20%. Optimization measures need to be implemented within 3 days."

[0041] S103 processes preliminary maintenance needs and energy consumption anomaly warning information, simulates equipment operating status at different maintenance times using digital twin technology, designs a graded maintenance strategy based on the full life cycle cost model, applies component remaining lifespan threshold constraints and maintenance resource optimal allocation objective function, and generates target maintenance execution instructions and energy-saving potential analysis reports.

[0042] In one implementation, the component degradation level data in the initial maintenance requirements are categorized, labeled, and prioritized according to maintenance timeliness, generating basic parameters for component maintenance. A three-level degradation standard is used to classify and label the component status: Level 1 (mild degradation) indicates a performance decrease of ≤10% with no risk of failure; Level 2 (moderate degradation) indicates a performance decrease of 10%-20% with potential failure risk; and Level 3 (severe degradation) indicates a performance decrease of >20% with a possible failure in the short term. Example: For a compressor, preliminary monitoring data shows a lubricating oil level of 60% (normal range 50%-80%) and a winding insulation resistance of 90MΩ (normal range ≥100MΩ), resulting in an overall performance decrease of 12%, labeled as "Level 2 degradation"; a heat exchanger experiences an 8% decrease in heat exchange efficiency due to scaling, labeled as "Level 1 degradation"; and a clogged filter leads to a 15% reduction in airflow, labeled as "Level 2 degradation".

[0043] Maintenance timeliness is categorized based on the degree of degradation and the importance of the component. Level 1 degraded components (low risk) have a maintenance timeliness of 30 days; Level 2 degraded components (medium risk) have a timeliness of 7-15 days; and Level 3 degraded components (high risk) have a timeliness of 72 hours. Example: A compressor, as a core component (downtime loss of 500 yuan / hour), although classified as Level 2 degraded, has a timeliness shortened to 10 days; a heat exchanger, being a Level 2 degraded and non-core component, has a timeliness set at 15 days; a filter, although classified as Level 2 degraded but with low replacement cost, has a timeliness set at 12 days.

[0044] Prioritization is performed using a "deterioration level × component importance coefficient" (importance coefficient: compressor 1.2, heat exchanger 1.0, fan 0.8, filter 0.6). Components are sorted from highest to lowest score, and basic maintenance parameters (including deterioration level, maintenance timeframe, priority score, and core maintenance content) are generated simultaneously. Example: Compressor score = 2 × 1.2 = 2.4, heat exchanger score = 1 × 1.0 = 1.0, filter score = 2 × 0.6 = 1.2, fan score = 1 × 0.8 = 0.8, resulting in a ranking of "compressor > filter > heat exchanger > fan". Specific basic parameters are: compressor (level 2 deterioration, maintenance within 10 days, priority 2.4, maintenance content: replenish lubricating oil + test insulation), filter (level 2 deterioration, maintenance within 12 days, priority 1.2, maintenance content: clean / replace filter), etc.

[0045] This process involves identifying anomaly types, decomposing influencing factors, and calculating energy-saving potential from energy consumption anomaly warning information. This generates energy consumption optimization analysis parameters, including feature matching model parameters for energy consumption anomalies, influence coefficients of environmental load / equipment operating conditions / operating parameters on energy consumption, and quantitative calculation rules for energy-saving potential under different anomaly scenarios. Energy consumption fluctuation data is extracted from energy consumption anomaly warning information. Through "anomaly type identification to pinpoint the essence of the problem, influencing factor decomposition to clarify key causes, and energy-saving potential calculation to quantify optimization potential," energy consumption optimization analysis parameters containing feature matching parameters, influence coefficients, and calculation rules are generated, providing a basis for subsequent energy-saving strategy formulation. Anomaly type identification is performed based on energy consumption fluctuation characteristics (such as peak overconsumption, high consumption under no-load conditions, and continuous high consumption) compared with historical data to pinpoint the essence of the problem. Example: During peak hours (10:00-12:00), the energy consumption of the air conditioning in a certain workshop is 80kWh, which is 33% higher than the historical average of 60kWh for the same period. The fluctuation curve shows a "continuous upward" trend (not instantaneous fluctuation). Combined with the heat exchanger scaling data, it is identified as "peak overconsumption caused by decreased heat exchange efficiency". During the nighttime off-load period (23:00-7:00), the energy consumption is 15kWh, which is 87.5% higher than the normal off-load value of 8kWh. Combined with the fan speed data (it is still maintained at 1200r / min during off-load, while it should be 800r / min normally), it is identified as "high off-load consumption caused by unreasonable operating parameter settings".

[0046] Fishbone diagram analysis was used to break down influencing factors and identify key causes. These factors were categorized into three types: environmental load (e.g., workshop production load), equipment condition (e.g., component deterioration), and operating parameters (e.g., fan speed, set temperature). The influence coefficients of environmental load, equipment condition, and operating parameters on energy consumption were calculated (coefficient range 0-1, with larger coefficients indicating more significant effects). For example, for "peak energy consumption exceeding capacity," the breakdown revealed: environmental load (workshop load 120% of rated value, influence coefficient 0.3), equipment condition (mild fouling of heat exchanger, influence coefficient 0.4), and operating parameters (indoor temperature set at 28℃, normal 26℃, influence coefficient 0.3), totaling an influence coefficient of 1.0, covering all abnormal causes. For "high energy consumption under no-load conditions," the breakdown revealed: operating parameters (excessive fan speed, influence coefficient 0.8) and equipment condition (mild wear of fan bearings, influence coefficient 0.2), clearly identifying fan speed as the primary influencing factor.

[0047] Based on influencing factors and energy consumption fluctuation data, energy-saving potential is calculated to quantify optimization potential, and energy consumption optimization analysis parameters are generated simultaneously (including characteristic matching model parameters for energy consumption anomalies, influence coefficients of environmental load / equipment operating conditions / operating parameters on energy consumption, and quantitative calculation rules for energy-saving potential under different abnormal scenarios). Example: For "peak overconsumption," the quantitative calculation rule for energy-saving potential under different abnormal scenarios is "Energy-saving potential = (Measured energy consumption - Baseline energy consumption) × Σ (Optimization rate of influencing factors × Influence coefficient)", where the baseline energy consumption is 60kWh, the optimization rate after heat exchanger cleaning is 8% (influence coefficient 0.4), the optimization rate after indoor temperature is reduced to 26℃ is 10% (influence coefficient 0.3), and the optimization rate after load is reduced to 100% is 5% (influence coefficient 0.3). The calculated energy-saving potential is = (80-60) × (8% × 0.4 + 10). (%×0.3+5%×0.3)=20×(0.032+0.03+0.015)=1.54kWh / peak segment; The final generated energy consumption optimization analysis parameters include: characteristic matching model parameters of energy consumption anomalies (peak segment energy consumption exceedance >20% and heat exchange efficiency decrease >5%), influence coefficients of environmental load / equipment condition / operating parameters on energy consumption (environmental load 0.3, equipment condition 0.4, operating parameters 0.3), and quantitative calculation rules for energy saving potential under different abnormal scenarios (the above calculation formula).

[0048] Based on digital twin technology, the operating status of equipment at different maintenance times is simulated and extrapolated. This is achieved through pre-maintenance, during-maintenance, and post-maintenance performance simulations, component lifespan degradation simulations at different maintenance intervals, and dynamic calculations of the energy consumption impact of maintenance interventions, generating maintenance timing decision parameters. A digital twin model of an industrial air conditioning system (a 1:1 replica of the equipment structure, operating logic, and environmental interaction) is constructed. By simulating the operating status of equipment at different maintenance times (e.g., immediate maintenance, maintenance after 7 days, maintenance after 15 days), the model outputs data on performance changes, lifespan degradation trends, and energy consumption impacts before, during, and after maintenance, generating maintenance timing decision parameters (including the optimal maintenance window, post-maintenance performance recovery rate, and energy consumption improvement rate). Pre-maintenance, during, and post-maintenance performance simulations are performed based on the digital twin model, inputting current equipment status parameters (e.g., compressor stage 2 degradation, heat exchanger stage 1 degradation), to simulate key performance indicators at different maintenance stages. Example: For a compressor (level 2 degradation, current exhaust temperature 90℃, normal ≤85℃), simulate the "immediate maintenance" scenario: before maintenance, the exhaust temperature is 90℃. During maintenance (replacing lubricating oil + testing insulation), the compressor is shut down for 2 hours. After maintenance, the exhaust temperature drops to 82℃, and the performance recovery rate is 95%. Simulate the "maintenance after 7 days" scenario: before maintenance, the exhaust temperature rises to 93℃ (degradation worsens). After maintenance, it drops to 83℃, and the performance recovery rate is 93%. Simulate the "maintenance after 15 days" scenario: before maintenance, the exhaust temperature rises to 98℃ (close to the fault threshold of 100℃). After maintenance, it drops to 85℃, and the performance recovery rate is 90%. In addition, some aged parts need to be replaced.

[0049] This simulation demonstrates the impact of different maintenance intervals (e.g., 30 days, 60 days, 90 days) on the remaining lifespan of components, generating lifespan decay curves. Example: For a wind turbine (currently estimated remaining lifespan of 180 days), the simulation shows a "30-day maintenance interval": after each maintenance, the lifespan recovers to 90% of its initial value, leaving 120 days remaining after 180 days; a "60-day maintenance interval": the lifespan decays more rapidly, leaving only 80 days remaining after 180 days; a "90-day maintenance interval": a minor fault occurs within 180 days, reducing the remaining lifespan to 50 days. Through comparison, the optimal maintenance interval is determined to be 30 days, corresponding to a lifespan decay coefficient (lifespan recovery rate after each maintenance) of 0.9.

[0050] Dynamically calculating the impact of maintenance intervention on energy consumption, this study measures the energy efficiency improvement effect of maintenance intervention at different maintenance times. Example: For a heat exchanger (Level 1 degradation, current peak energy consumption increases by 8% due to scaling), simulating "immediate maintenance": after maintenance, peak energy consumption decreases from 80 kWh to 75 kWh, an energy efficiency improvement rate of 6.25%; simulating "maintenance after 7 days": due to intensified scaling, peak energy consumption decreases to 76 kWh after maintenance, an energy efficiency improvement rate of 5%; simulating "maintenance after 15 days": scaling progresses to Level 2, peak energy consumption decreases to 78 kWh after maintenance, an energy efficiency improvement rate of 2.5%; combining performance and energy consumption data, maintenance timing decision parameters are generated: the optimal maintenance window is "immediate maintenance - within 7 days", with a performance recovery rate ≥93% and an energy efficiency improvement rate ≥5% after maintenance.

[0051] By combining a full lifecycle cost model, cost composition analysis, benefit return calculation, and resource adaptation planning are performed on the tiered maintenance strategy to generate maintenance strategy optimization parameters. Cost composition analysis is conducted on different maintenance schemes within the tiered maintenance strategy, breaking down each cost and quantifying its amount. Example: For a compressor (level 2 degradation, priority 1), analyze the cost of the "priority maintenance" strategy: maintenance cost (labor 200 yuan + lubricating oil 150 yuan = 350 yuan), downtime loss (2 hours × 500 yuan / hour = 1000 yuan), energy cost (energy consumption decreases by 5% after maintenance, saving 120 yuan per month); analyze the cost of the "routine maintenance" strategy (maintenance after 15 days): maintenance cost (due to increased degradation, additional parts need to be replaced, increasing by 300 yuan, totaling 650 yuan), downtime loss (2 hours × 500 yuan / hour = 1000 yuan), energy cost (approximately 80 yuan more electricity is consumed in 15 days, reducing monthly savings to 80 yuan); the comparison shows that the total cost of "priority maintenance" (1350 yuan - 120 yuan = 1230 yuan) is lower than that of "routine maintenance" (1650 yuan - 80 yuan = 1570 yuan).

[0052] Based on the life-cycle cost model, benefit-return calculations are performed, calculating the ratio (cost-benefit ratio) of the benefits (including fault avoidance benefits and energy savings benefits) to the costs of different maintenance strategies. A ratio greater than 1 indicates that the strategy is feasible. Example: The benefits of the "priority maintenance" strategy include: avoiding compressor failure losses (single failure repair cost of 5000 yuan), monthly energy savings of 120 yuan (annualized 1440 yuan), total benefit = 5000 + 1440 = 6440 yuan; cost-benefit ratio = 6440 yuan / 1230 yuan ≈ 5.24 (>1, feasible); The benefits of the "routine maintenance" strategy include: avoiding fault losses of 5000 yuan, monthly energy savings of 80 yuan (annualized 960 yuan), total benefit = 5960 yuan; cost-benefit ratio = 5960 yuan / 1570 yuan ≈ 3.79 (>1, but lower than priority maintenance).

[0053] Based on the current status of the enterprise's maintenance resources (such as manpower, spare parts, and funds), resource adaptation planning is carried out for the hierarchical maintenance strategy, resource allocation schemes for different strategies are formulated, and maintenance strategy optimization parameters are generated. Example: The enterprise currently has 2 available maintenance personnel and sufficient compressor lubricating oil spare parts. The "priority maintenance" strategy can immediately allocate 1 person + 1 set of spare parts, which can be completed within 3 days. The final maintenance strategy optimization parameters include: optimal maintenance level (compressor priority maintenance, filter routine maintenance, heat exchanger routine maintenance), cost-effectiveness ratio (compressor priority maintenance 5.24, filter routine maintenance 3.12), and resource adaptation scheme (1 maintenance personnel / time, 1 set of lubricating oil spare parts / time, downtime 2 hours / unit).

[0054] Applying a component remaining lifespan threshold constraint and a maintenance resource allocation objective function, the above parameters are integrated and optimized. The feasibility of the maintenance plan is verified by the lower limit of the remaining lifespan, and the maintenance task schedule is adjusted to maximize resource utilization, generating target maintenance execution instructions and an energy-saving potential analysis report. Using "component remaining lifespan threshold constraints (e.g., core component remaining lifespan ≥ 100 days)" and "maintenance resource allocation objective function (e.g., maximizing maintenance resource utilization, target value ≥ 90%)" as dual constraints, the above component maintenance basic parameters, energy consumption optimization analysis parameters, maintenance timing decision parameters, and maintenance strategy optimization parameters are integrated and optimized. The feasibility of the maintenance plan is verified, the task schedule is adjusted, and finally, target maintenance execution instructions (including specific maintenance tasks, time, and resources) and an energy-saving potential analysis report (containing quantitative energy-saving targets and implementation paths) are generated. Inputting the current remaining lifespan data of each component verifies whether the maintenance plan meets the requirement of "remaining lifespan ≥ lower limit" (100 days for core components, 60 days for non-core components). Example: The compressor's current estimated remaining lifespan is 150 days. If the "maintenance after 7 days" plan is followed, the remaining lifespan will be restored to 200 days after maintenance, satisfying the ≥100-day constraint. The fan's current estimated remaining lifespan is 180 days. If the "maintenance interval of 30 days" plan is followed, the remaining lifespan will always be ≥120 days, satisfying the constraint. If a component (such as an old filter) has a current remaining lifespan of only 50 days (below the lower limit of 60 days for non-core components), the maintenance plan should be adjusted to "immediate maintenance" to avoid failure due to insufficient lifespan.

[0055] With the goal of "maximizing maintenance resource utilization (target ≥ 90%)", adjust the maintenance task schedule (e.g., stagger the maintenance of multiple devices to avoid resource idleness). Example: A company currently has 2 maintenance personnel, 1 set of compressor spare parts, and 2 sets of filter spare parts. The initial schedule is "Day 1: Maintain compressor (1 person), Day 3: Maintain filter (1 person)", with a resource utilization rate of only 50%. After optimization, the schedule is "Day 1: Morning: Maintain compressor (1 person), Afternoon: Maintain filter for 2 devices (1 person)", with a personnel utilization rate of 100%, a spare parts utilization rate of 100%, and a resource utilization rate of 100% (exceeding the target value).

[0056] The target maintenance execution instructions are as follows: "1. Compressor: Maintenance level priority, execution time Day 1 morning, maintenance content: replenish lubricating oil (1 set of spare parts) + check winding insulation, 1 maintenance personnel, 2-hour downtime; 2. Filter: Maintenance level routine, execution time Day 1 afternoon, maintenance content: clean / replace (2 sets of spare parts), 1 maintenance personnel, 1-hour downtime per unit; 3. Heat exchanger: Maintenance level routine, execution time Day 7 morning, maintenance content: remove scale, 1 maintenance personnel, 1.5-hour downtime."

[0057] The energy-saving potential analysis report states: "1. Peak excess consumption optimization: Through compressor maintenance (energy consumption improvement rate of 5%) + heat exchanger maintenance (energy consumption improvement rate of 6.25%), peak energy consumption was reduced from 80kWh to 72kWh, resulting in monthly energy savings of approximately 240kWh, equivalent to 192 yuan in electricity costs (peak electricity price of 0.8 yuan / kWh); 2. Idle high consumption optimization: Through fan speed parameter adjustment (influence coefficient of 0.8), idle energy consumption was reduced from 15kWh to 9kWh, resulting in monthly energy savings of approximately 180kWh, equivalent to 54 yuan in electricity costs (off-peak electricity price of 0.3 yuan / kWh); 3. Total energy-saving target: monthly energy savings of 420kWh, annual energy savings of 5040kWh, and energy-saving benefits of approximately 4032 yuan."

[0058] S104 processes the target maintenance execution command and energy-saving potential analysis report based on the corresponding adaptive algorithm, and adjusts the air conditioning operating parameters in real time in combination with indoor thermal comfort requirements and the power grid peak-valley electricity price mechanism to generate a linkage control scheme that takes into account both maintenance effect and energy consumption optimization.

[0059] In one implementation, adaptive feature matching and time-series correlation processing are performed on the maintenance time period requirements, component protection parameters, and energy consumption anomaly data and energy-saving target values ​​in the target maintenance execution instructions to generate maintenance demand features and energy consumption optimization features. Maintenance time period requirements (such as maintenance window, downtime) and component protection parameters (such as component allowable operating temperature, pressure threshold) are extracted from the target maintenance execution instructions. Energy consumption anomaly data (such as peak-period overconsumption values, idle high-consumption duration) and energy-saving target values ​​(such as monthly energy saving kWh, percentage reduction in energy consumption) are extracted from the energy-saving potential analysis report. Through "adaptive feature matching + time-series correlation mapping," unstructured instructions and report data are transformed into standardized maintenance demand features (including time constraints and parameter constraints) and energy consumption optimization features (including anomaly features and target features).

[0060] The target maintenance execution instructions are clear: "The compressor maintenance period is Day 19:00-11:00 (2-hour shutdown), and the component protection parameters are: compressor exhaust temperature ≤90℃ and lubricating oil level ≥55% before and after maintenance; the filter maintenance period is Day 114:00-15:00 (1-hour shutdown), and the component protection parameters are: ventilation volume reduction rate ≤8% after maintenance." Through adaptive feature matching, "maintenance period 9:00-11:00" is matched as "time constraint feature (T1: 9:00-11:00, downtime 2h)", and "exhaust temperature ≤90℃, lubricating oil level ≥55%" is matched as "parameter constraint feature (P1: exhaust temperature ≤90℃, P2: lubricating oil level ≥55%)". Similarly, filter maintenance is matched as "time constraint feature (T2: 14:00-15:00, downtime 1h)" and "parameter constraint feature (P3: ventilation volume attenuation rate ≤8%)", and the maintenance requirement features are integrated to generate: {T1,T2,P1,P2,P3}.

[0061] The energy-saving potential analysis report shows that "peak-hour (10:00-12:00) energy consumption anomaly data is 80kWh (historical average 60kWh, exceeding consumption by 33%), and idle-hour (23:00-7:00) energy consumption anomaly data is 15kWh (normal average 8kWh, exceeding consumption by 87.5%); the energy-saving target value is 240kWh of monthly peak-hour energy saving and 180kWh of idle-hour energy saving". Through adaptive feature matching, "peak-hour overconsumption 33%, idle-hour overconsumption 87.5%" is matched as "energy consumption anomaly characteristics (E1: peak-hour overconsumption > 20%, E2: idle-hour overconsumption > 50%)", and "monthly peak-hour energy saving 240kWh, idle-hour energy saving 180kWh" is matched as "energy-saving target characteristics (G1: peak-hour monthly energy saving 240kWh, G2: idle-hour monthly energy saving 180kWh)", integrating to generate energy consumption optimization features: {E1, E2, G1, G2}.

[0062] Based on maintenance demand characteristics and energy consumption optimization characteristics, and combining the temperature and humidity threshold ranges and human comfort weight coefficients in indoor thermal comfort requirements with the time period division standards and electricity price fluctuation ratios in the power grid peak-valley pricing mechanism, a dual-objective optimization weight is generated for the adaptive algorithm. The maintenance demand characteristics and energy consumption optimization characteristics are correlated based on the time dimension to avoid conflicts between maintenance periods and high-energy-consumption periods. For example, the compressor maintenance period T1 (9:00-11:00) partially overlaps with the peak period (10:00-12:00). After association, the adjusted feature is "T1': 9:00-10:00 (avoiding the core peak period after 10:00), while strengthening P1 (exhaust temperature ≤88℃) to reduce the impact of peak energy consumption". The filter maintenance period T2 (14:00-15:00) is a flat period (electricity price 0.5 yuan / kWh), with no conflict, so the original feature remains unchanged. The final associated feature set is the maintenance demand feature {T1',T2,P1',P2,P3} and the energy consumption optimization feature {E1,E2,G1,G2}.

[0063] With "achieving maintenance effectiveness" and "lowest energy consumption cost" as dual objectives, and combining the temperature and humidity threshold ranges (temperature threshold 22-26℃, humidity threshold 40%-60%) in indoor thermal comfort requirements, the weight coefficient of human comfort (weight 1.0 at 24℃, decreasing by 0.1 for every 1℃ deviation; weight 1.0 at 50% RH, decreasing by 0.1 for every 5% deviation), and the time period division standards in the power grid peak-valley pricing mechanism (valley 23:00-7:00, flat 7:00-10:00, 15:00-23:00, peak 10:00-15:00), and the electricity price fluctuation ratio (peak fluctuation is 167% compared to valley), an adaptive algorithm is generated to optimize the dual objectives (maintenance target weight W1, energy consumption target weight W2, and W1+W2=1) through "constraint quantification + dynamic weight allocation". This ensures that the weight allocation meets both equipment maintenance needs and energy-saving and comfort requirements.

[0064] The current indoor temperature is 25℃ (perceived weight 0.9) and humidity is 55% (perceived weight 0.9). The overall thermal comfort quantification value is (0.9 + 0.9) / 2 = 0.9. The current time period is peak (10:00-12:00), and the electricity price quantification value is 0.8 / 0.3 ≈ 2.67 (the ratio of peak electricity price to off-peak price). The maintenance target base weight W10 = 0.5, and the energy consumption target base weight W20 = 0.5. The quantification values ​​are dynamically adjusted according to the constraints: when the overall thermal comfort quantification value is ≥ 0.8, W10 remains unchanged; if < 0.8, W1 (maintenance weight) is decreased by 0.1, and W2 (energy consumption weight) is increased by 0.1 (prioritizing energy-saving adjustments to improve comfort). The current quantified value is 0.9 ≥ 0.8, so W1 remains unchanged. When the electricity price quantified value is ≥ 2.0 (peak period), W2 is increased by 0.1, and W1 is decreased by 0.1 (prioritizing energy saving). When 1.0 < electricity price quantified value < 2.0 (flat period), the weights remain unchanged. When ≤ 1.0 (valley period), W2 is decreased by 0.1, and W1 is increased by 0.1 (prioritizing maintenance). The current electricity price quantified value is 2.67 ≥ 2.0, so W2 is increased by 0.1, and W1 is decreased by 0.1. The final adaptive algorithm's bi-objective optimization weights are W1 = 0.5 - 0.1 = 0.4 (maintenance target weight) and W2 = 0.5 + 0.1 = 0.6 (energy consumption target weight).

[0065] To verify whether the weight meets the requirement of "maintenance needs not being ignored", the minimum threshold for W1 is set to 0.3. Currently, W1 = 0.4 > 0.3, which meets the requirement. If the thermal comfort quantification value is 0.7 (< 0.8) and the electricity price quantification value is 2.5 (≥ 2.0) during a certain period, then W1 = 0.5 - 0.1 - 0.1 = 0.3 (reaching the minimum threshold), and W2 = 0.7, to avoid equipment damage caused by excessively low maintenance weight.

[0066] Input the initial parameters for iteration: compressor start / stop frequency 4 times / hour (maintenance allowable range 3-5 times / hour), fan speed 1400r / min (maintenance allowable range 1200-1500r / min), indoor set temperature 25℃ (thermal comfort range 22-26℃); the optimization objective function of the adaptive algorithm is: Min((1-W1)×maintenance parameter deviation + W2×energy consumption deviation), where maintenance parameter deviation = (measured component parameters - protection threshold) / protection threshold, and energy consumption deviation = (measured energy consumption - energy saving target value) / energy saving target value. The iterative process is as follows: First iteration: Parameter combination (4 times / hour, 1400 r / min, 25℃), maintenance parameter deviation 0.05 (compressor exhaust temperature 88℃, protection threshold 90℃, deviation (88-90) / 90≈-0.02, take the absolute value 0.02; lubricating oil level 58%, protection threshold 55%, deviation 0.05, comprehensive deviation (0.02+0.05) / 2=0.035), energy consumption deviation 0.1 (peak measured energy consumption 78kWh, energy saving target value 72kWh, deviation (78-72) / 72≈0.083); objective function value = (1-0.4)×0.035+0.6×0. 083≈0.069; 50th iteration (convergence): Parameter combination (3 times / hour, 1300 r / min, 26℃), maintenance parameter deviation 0.02 (exhaust temperature 87℃, deviation 0.033; lubricating oil level 60%, deviation 0.09, comprehensive deviation 0.0615 → corrected to 0.02, because some parameters are better than the threshold, the lower limit of the deviation is taken), energy consumption deviation 0.01 (peak measured energy consumption 72.7kWh, deviation (72.7-72) / 72≈0.0097≈0.01); objective function value = 0.6×0.02+0.6×0.01=0.018≤0.01 (convergence error), iteration ends.

[0067] Based on the parameter iteration output results of the dual-objective optimization weights and adaptive algorithm, the air conditioner operating parameters are adjusted in real time to generate a linkage control scheme that balances maintenance effectiveness and energy consumption optimization. According to the dual-objective optimization weights (maintenance target weight W1=0.4, energy consumption target weight W2=0.6) and the parameter iteration output results of the adaptive algorithm (converged parameters in the 50th iteration: compressor start / stop frequency 3 times / hour, fan speed 1300 r / min, indoor set temperature 26℃), the air conditioner operating parameters are adjusted in real time. Specifically, the compressor start / stop frequency is reduced from 4 times / hour to 3 times / hour (reducing start / stop losses and meeting the maintenance parameter deviation requirement of 0.02), while avoiding the maintenance period T1' (9:00-10:00). The start time has been adjusted to 10:00 to avoid overlap between shutdown and peak hours; the fan speed has been reduced from 1400 r / min to 1300 r / min (reducing energy consumption, with peak energy consumption decreasing from 78 kWh to 72.7 kWh, close to the energy-saving target of 72 kWh); the indoor set temperature has been increased from 25℃ to 26℃ (within the thermal comfort range, reducing cooling capacity consumption and contributing to energy optimization); at the same time, the refrigerant pressure has been adjusted: the high-pressure pressure has been reduced from 2.0 MPa to 1.9 MPa (maintenance allowable range 1.8-2.2 MPa), reducing the compressor load.

[0068] The final generated linkage control scheme, which balances maintenance effectiveness and energy consumption optimization, includes a three-dimensional correlation of "time period-parameter-target" to achieve dynamic coordination between maintenance and energy saving. Specifically, during peak hours (10:00-15:00): compressor start / stop frequency 3 times / hour, fan speed 1300r / min, indoor set temperature 26℃, refrigerant high pressure 1.9MPa, target: maintenance parameter deviation ≤0.02, peak energy consumption ≤73kWh; during maintenance period T1' (9:00-10:00): compressor shutdown, The fan speed is reduced to 1200 r / min (to maintain basic ventilation), and the indoor temperature is set at 24℃ (to avoid excessive temperature fluctuations). The target is: compressor lubricating oil level ≥ 55%, and indoor temperature fluctuation ≤ 2℃ during shutdown. During off-load periods (23:00-7:00): the compressor starts and stops twice per hour (off-peak electricity prices are low, so maintenance weight can be appropriately reduced), the fan speed is 1200 r / min, and the indoor temperature is set at 23℃. The target is: off-load energy consumption ≤ 9.5 kWh (close to the energy-saving target of 9 kWh), and no abnormalities in component parameters. Through dynamic parameter adjustment and time-based target setting, while ensuring equipment maintenance needs (such as meeting component parameter standards and extending lifespan), precise energy consumption control (such as reducing peak-hour energy consumption and reducing off-load losses) is achieved, forming a closed-loop linkage between maintenance and energy saving.

[0069] S105 processes the linkage control scheme that balances maintenance effectiveness and energy consumption optimization, and generates a globally coordinated operation management strategy.

[0070] In one implementation, the adjustment rules for single-equipment operating parameters, maintenance and energy-saving coordination thresholds, and load allocation requirements and energy supply constraints of multiple air conditioning units in an industrial scenario are processed using global feature normalization and coordination correlation extraction to generate standardized coordination features and global trend features. Taking the core parameters of three air conditioners (A, B, and C) in an industrial workshop as an example, the original data to be normalized includes the adjustment rules for single-equipment operating parameters: compressor start-stop frequency of air conditioner A is 3-5 times / hour, air conditioner B is 4-6 times / hour, and air conditioner C is 2-4 times / hour; maintenance and energy-saving coordination thresholds: energy consumption limit of air conditioner A during maintenance period is 73kWh, air conditioner B is 75kWh, and air conditioner C is 70kWh; load allocation requirements: load of area handled by air conditioner A is 300kW, air conditioner B is 400kW, and air conditioner C is 300kW (total load 1000kW); energy supply constraint data: maximum power supply of the power grid during peak hours is 800kW, during normal hours it is 1000kW, and during off-peak hours it is 1200kW.

[0071] Global feature normalization is performed using the "min-max normalization formula (normalized value = (original value - minimum value) / (maximum value - minimum value))": Compressor start / stop frequency: Air conditioner A (3-5) normalized to (0-0.67), Air conditioner B (4-6) normalized to (0.33-1.0), Air conditioner C (2-4) normalized to (0-0.33); Energy consumption limit during maintenance period: Air conditioner A 73kWh normalized value = (73-7 0) / (75-70) = 0.6, Air conditioner B with 75kWh is 1.0, Air conditioner C with 70kWh is 0.0; Load distribution ratio (after normalization, i.e., load ratio): Air conditioner A 30%, Air conditioner B 40%, Air conditioner C 30%; Energy supply constraints: Peak segment 800kW normalized value = (800-800) / (1200-800) = 0.0, Flat segment 1000kW is 0.5, Valley segment 1200kW is 1.0.

[0072] By calculating the Pearson correlation coefficients between different parameters, the synergistic correlation degree was extracted and processed to explore the synergistic correlation relationship. The correlation degree between the operating parameters of a single device and the load distribution: the correlation coefficient between the start-stop frequency of the compressor of air conditioner A and the regional load is 0.82 (the higher the load, the higher the upper limit of the start-stop frequency), air conditioner B is 0.85, air conditioner C is 0.78, and the overall average correlation degree is 0.82, which is judged as "strong synergistic correlation"; the correlation degree between the maintenance and energy saving synergistic threshold and energy supply: the correlation coefficient between the upper limit of energy consumption during the maintenance period and the peak power supply is -0.75 (the lower the peak power supply, the stricter the energy consumption upper limit), which is judged as "strong negative synergistic correlation".

[0073] The standardized collaborative features are formed by integrating normalized data and collaborative correlation, resulting in {equipment number: [normalized range of start / stop frequency, normalized value of energy consumption limit, load percentage, correlation with load]}, such as air conditioner A: [0-0.67, 0.6, 30%, 0.82], air conditioner B: [0.33-1.0, 1.0, 40%, 0.85], air conditioner C: [0-0.33, 0.0, 30%, 0.78]. The global trend features are extracted based on the collaborative correlation and energy supply data of the three air conditioners, such as "when the peak load is concentrated (A+B+C load exceeds 80%), the upper limit of equipment start / stop frequency needs to be reduced by 10% to match the energy supply limit" and "maintenance time is prioritized in the valley (sufficient energy supply, normalized value 1.0)". These trends are quantified as "peak load - frequency adjustment coefficient 0.9" and "valley maintenance priority weight 1.2".

[0074] Based on standardized collaborative characteristics and global trend characteristics, and combined with the equipment cluster operation efficiency target, total energy consumption limit, and environmental collaborative constraint mechanism in the industrial scenario, the multi-dimensional adaptation weights of the global collaborative algorithm are generated. Equipment cluster operation efficiency target (e.g., total cooling capacity of 3 air conditioners ≥ 200kW): basic weight W10 = 0.3 (core target, impacting the production environment); total energy consumption limit (e.g., daily energy consumption ≤ 1500kWh): basic weight W20 = 0.3 (key to cost control); workshop temperature and humidity regional linkage standard (e.g., temperature difference between areas ≤ 2℃, humidity 40%-60%): basic weight W30 = 0.25 (ensuring production process and personnel comfort); pollutant emission control indicators (e.g., air conditioner refrigerant leakage rate ≤ 0.5% / year): basic weight W40 = 0.15 (compliance requirement, relatively minor impact).

[0075] Weight adjustments were made based on standardized collaborative features and global trend features. Equipment efficiency weight adjustment: The global trend feature shows "large fluctuations in equipment efficiency during peak load periods," and the current peak load ratio exceeds 80%, requiring an increase in equipment efficiency weight. The adjustment coefficient is 1.1, W1 = 0.3 × 1.1 = 0.33. Energy quota weight adjustment: Energy supply restriction data shows "tight power supply during peak periods (normalized value 0.0)," indicating stricter energy quota constraints. The adjustment coefficient is 1.2, W2 = 0.3 × 1.2 = 0.36. Temperature and humidity collaborative weight adjustment: The standardized collaborative feature shows a current temperature difference of 1.5℃ (≤2℃, meeting the constraint) in the areas served by the three air conditioners, requiring no adjustment. W3 = 0.25 × 1.0 = 0.25. Emission control weight adjustment: The current refrigerant leakage rate is 0.3% (≤0.5%, meeting the constraint), and there is no abnormal emission trend. The adjustment coefficient is 0.8, W4 = 0.15 × 0.8 = 0.12.

[0076] The corrected total weights = 0.33 + 0.36 + 0.25 + 0.12 = 1.06, which need to be normalized (normalized weights = corrected weights / total weights): W1 = 0.33 / 1.06 ≈ 0.31 (equipment efficiency weight); W2 = 0.36 / 1.06 ≈ 0.34 (energy quota weight); W3 = 0.25 / 1.06 ≈ 0.24 (temperature and humidity coordination weight); W4 = 0.12 / 1.06 ≈ 0.11 (emission control weight); all weights are ≥ 0.1 (no target is ignored), and the energy quota weight is the highest (consistent with the constraint of peak power supply shortage), followed by equipment efficiency and temperature and humidity weights, and the emission weight is the lowest (consistent with the current situation of compliance and no abnormalities). The multi-dimensional adaptation weight allocation of the global collaborative algorithm is reasonable. Key operating conditions during the forecast period: Peak period (10:00-15:00, power supply 800kW), flat period (7:00-10:00, 15:00-23:00, power supply 1000kW), valley period (23:00-7:00, power supply 1200kW); 3 air conditioners need to be maintained (A air conditioner compressor, B air conditioner filter, C air conditioner heat exchanger), and the total load during the peak period is expected to be 900kW (exceeding the power supply limit).

[0077] Based on the iterative calculation results of multi-dimensional adaptation weights and global collaborative algorithms, a globally collaborative operation management strategy is generated, including a cross-device staggered scheduling scheme for maintenance tasks of multiple air conditioning devices, dynamic energy allocation rules during high-load periods, multi-device linkage control instructions when regional temperature and humidity are abnormal, and the optimal scheduling plan for maintenance resources among cluster devices. Iterative optimization objective function: Max(W1×cluster efficiency + W3×temperature and humidity compliance rate - W2×energy excess rate - W4×emission excess rate), where: Cluster efficiency = actual total cooling capacity / rated total cooling capacity; Energy excess rate = (actual energy consumption - limit) / limit (positive when exceeding the limit, otherwise 0); Peak period: Air conditioner A starts and stops 3 times / hour (original upper limit 5 times / hour, reduced by 40%), Air conditioner B starts and stops 4 times / hour (original upper limit 6 times / hour, reduced by 33%), Air conditioner C starts and stops 2 times / hour (original upper limit 4 times / hour, reduced by 50%), total energy consumption is controlled at 780kW (≤800kW), cluster efficiency is 0.92; Air conditioner A maintenance is scheduled (2:00-4:00), Air conditioner C maintenance is scheduled (4:00-6:00), and Air conditioner B maintenance is scheduled during the off-peak period (16:00-17:00, when the load is low) to avoid overlapping maintenance periods that cause temperature and humidity fluctuations.

[0078] Cross - device peak - load shifting scheduling plan for multi - air - conditioner maintenance tasks: For Air - conditioner A (compressor maintenance), from 12:00 to 4:00 on Day 1 (valley period, sufficient energy); for Air - conditioner B (filter maintenance), from 16:00 to 17:00 on Day 1 (flat period, load 300kW, accounting for 30%); for Air - conditioner C (heat exchanger maintenance), from 4:00 to 6:00 on Day 1 (valley period). The maintenance interval is ≥2 hours to avoid temperature and humidity exceeding the standard due to multiple device outages in a single period. Dynamic energy distribution rules during high - load periods: During the peak period (10:00 - 15:00), distribute energy according to "load ratio × efficiency weight". For Air - conditioner A, 30%×0.31≈28% (224kW); for Air - conditioner B, 40%×0.31≈38% (304kW); for Air - conditioner C, 30%×0.31≈34% (272kW), with a total distribution of 780kW (≤800kW). During the flat / valley period, distribute according to "full - load operation, and give priority to supplementing the cooling capacity gap caused by valley - period maintenance".

[0079] Multi - device linkage control instructions when the regional temperature and humidity are abnormal: If the temperature in the area responsible for Air - conditioner A rises to 28℃ (exceeding the upper limit of 26℃), trigger linkage control - increase the air volume of Air - conditioner B by 10% (from 1300r / min to 1430r / min), lower the set temperature of Air - conditioner C by 0.5℃ (from 25℃ to 24.5℃), and at the same time, temporarily increase the start - stop frequency of Air - conditioner A to 4 times per hour (original 3 times per hour) until the temperature in this area drops back to 26℃. Optimal scheduling plan for maintenance resources among cluster devices: 2 maintenance personnel and 2 sets of spare - part kits (1 set of compressor lubricant, 2 sets of filters, 1 set of heat - exchanger cleaning agent). The scheduling plan is as follows: From 12:00 to 4:00 on Day 1: 1 person + 1 set of lubricant spare parts to maintain Air - conditioner A; from 4:00 to 6:00: the same 1 person + 1 set of cleaning - agent spare parts to maintain Air - conditioner C (saving the round - trip time of personnel); from 16:00 to 17:00: the other 1 person + 1 set of filter spare parts to maintain Air - conditioner B. The resource utilization rate is 100% with no idling.

[0080] S106, process the preliminary maintenance requirements, energy - consumption abnormal warning information, target maintenance execution instructions, linkage control scheme that takes into account both maintenance effects and energy - consumption optimization, and globally coordinated operation management strategies, and generate comprehensive evaluation information on the operating status of industrial air - conditioners, including equipment reliability levels and quantified energy - saving benefits.

[0081] In one implementation, the component risk level and energy consumption anomaly type in the preliminary maintenance needs and energy consumption anomaly early warning information are classified and quantified to generate basic evaluation index data. Each component risk level corresponds to a unique reliability coefficient, and each energy consumption anomaly type corresponds to a target energy-saving loss coefficient. The mapping rules between component risk level and reliability coefficient are set as follows: mild risk (component performance degradation ≤ 10%) corresponds to a unique reliability coefficient of 0.9; moderate risk (performance degradation 10%-20%) corresponds to a unique reliability coefficient of 0.7; and severe risk (performance degradation > 20%) corresponds to a unique reliability coefficient of 0.5. Based on the initial maintenance requirements, the compressor is assessed as having a medium risk due to a lubricating oil level of 60% (normal 50%-80%) and an insulation resistance of 90MΩ (normal ≥100MΩ), corresponding to a unique reliability coefficient of 0.7. Scaling in the heat exchanger leads to an 8% decrease in heat exchange efficiency, assessed as a slight risk, corresponding to a unique reliability coefficient of 0.9. Filter blockage results in a 15% reduction in ventilation volume, assessed as a medium risk, corresponding to a unique reliability coefficient of 0.7. The fan winding insulation is normal, assessed as having no risk, corresponding to a unique reliability coefficient of 1.0. Component risk quantification data are generated as follows: Compressor (0.7), Heat Exchanger (0.9), Fan (1.0), Filter (0.7).

[0082] The mapping rules for energy consumption anomaly types and target energy-saving loss coefficients are set as follows: peak energy consumption exceeding the limit by more than 20% corresponds to a target energy-saving loss coefficient of 1.2 (higher loss); idle high energy consumption exceeding the limit by more than 50% corresponds to a target energy-saving loss coefficient of 1.1; continuous high energy consumption exceeding the limit by more than 10% throughout the day corresponds to 1.05; and no anomaly corresponds to 1.0 (no loss). According to the energy consumption anomaly warning information, peak energy consumption (10:00-12:00) of 80kWh (exceeding the historical average by 33%) is judged as peak energy consumption exceeding the limit, corresponding to a coefficient of 1.2; idle energy consumption (23:00-7:00) of 15kWh (exceeding the normal average by 87.5%) is judged as idle high energy consumption, corresponding to a coefficient of 1.1; and there is no continuous high energy consumption, corresponding to a coefficient of 1.0. Quantitative data of energy consumption anomalies are generated: peak energy consumption exceeding the limit (1.2), idle high energy consumption (1.1), and continuous high energy consumption (1.0). Integrate component risk and energy consumption anomaly quantitative data to form basic assessment index data, including "component reliability coefficient set (compressor 0.7, heat exchanger 0.9, fan 1.0, filter 0.7)" and "energy consumption and energy saving loss coefficient set (peak 1.2, no-load 1.1, continuous 1.0)", while also marking the data source (preliminary maintenance needs, energy consumption anomaly early warning information) and assessment cycle (monthly).

[0083] The system integrates and calculates basic assessment index data, maintenance completion rate and resource utilization rate in target maintenance execution instructions, parameter compliance rate in linkage control schemes that balance maintenance effectiveness and energy consumption optimization, and cluster collaboration efficiency in global collaborative operation management strategies. This generates a multi-dimensional assessment matrix encompassing equipment health, energy consumption, and collaboration dimensions. In this matrix, row vectors represent assessment dimensions, and column vectors represent specific quantified values. Based on component reliability coefficients in the basic assessment index data, the average value is calculated as the dimensional quantified value. The standardized formula is "Health Dimension Value = (Σ Component Reliability Coefficient) / Total Number of Components". For example, (0.7 + 0.9 + 1.0 + 0.7) / 4 = 0.825, which is standardized to 0.83 (rounded to two decimal places). Simultaneously, the system extracts "Maintenance Completion Rate" (e.g., planned maintenance of 3 items, actual completion of 3 items, completion rate 100%, standardized to 1.0) and "Resource Utilization Rate" (e.g., 2 maintenance personnel, actual usage of 2 personnel, utilization rate 100%, standardized to 1.0) from the target maintenance execution instructions.

[0084] Based on the energy-saving loss coefficient in the basic assessment index data, calculate the weighted average (peak energy consumption percentage 40%, no-load percentage 20%, continuous percentage 40%), with the formula "Energy consumption dimension value = Σ (energy-saving loss coefficient × time period percentage)". Example: 1.2×0.4+1.1×0.2+1.0×0.4=1.1, standardized to 1.1; extract the "parameter compliance rate" (e.g., all 5 parameters such as compressor start-stop frequency and fan speed meet the standard, compliance rate 100%, standardized to 1.0) from the linkage control scheme that takes into account both maintenance effect and energy consumption optimization. Extract the "cluster collaboration efficiency" (e.g., the overall load distribution balance rate of 3 air conditioners is 90%, the energy dynamic distribution compliance rate is 95%, the comprehensive collaboration efficiency = (90%+95%) / 2=92.5%, standardized to 0.93) from the global collaborative operation management strategy; and simultaneously associate the "cross-equipment parameter collaboration degree" (e.g., temperature and humidity area difference ≤2℃, compliance rate 100%, standardized to 1.0) from the linkage control scheme that takes into account both maintenance effect and energy consumption optimization.

[0085] Using the evaluation dimensions as row vectors and the specific quantified values ​​as column vectors, a multi-dimensional evaluation matrix is ​​constructed as follows: Row 1 (Equipment Health Dimension): Health Dimension Value 0.83, Maintenance Completion Rate 1.0, Resource Utilization Rate 1.0; Row 2 (Energy Consumption Dimension): Energy Consumption Dimension Value 1.1, Parameter Compliance Rate 1.0; Row 3 (Collaboration Dimension): Cluster Collaboration Efficiency 0.93, Cross-Device Parameter Collaboration Degree 1.0; All quantified values ​​in the matrix are in the range of 0-1.2 (the energy consumption dimension may slightly exceed 1.0 due to the loss coefficient), and the calculation basis of each parameter is marked (e.g., the health dimension value comes from the average component reliability coefficient, and the cluster collaboration efficiency comes from the global strategy) to ensure traceability.

[0086] A multi-dimensional evaluation matrix is ​​input into the industrial air conditioning operation status evaluation model for weighted calculation, generating core evaluation results including equipment reliability level and energy-saving benefit quantification. Based on the previously constructed multi-dimensional evaluation matrix (Row 1 (Equipment Health Dimension): Health Dimension Value 0.83, Maintenance Completion Rate 1.0, Resource Utilization Rate 1.0; Row 2 (Energy Consumption Dimension): Energy Consumption Dimension Value 1.1, Parameter Compliance Rate 1.0; Row 3 (Collaboration Dimension): Cluster Collaboration Efficiency 0.93, Cross-Equipment Parameter Collaboration Degree 1.0), it is input into the industrial air conditioning operation status evaluation model for calculation. Weights are set for each dimension in the model: Equipment Health 0.4 (core, affecting equipment safety), Energy Consumption 0.35 (critical, affecting cost), and Collaboration 0.25 (auxiliary, affecting overall efficiency), and quantitative evaluation is achieved through weighted calculation. The weighted calculation formula is: Comprehensive evaluation value = (weighted sum of equipment health dimension) × 0.4 + (weighted sum of energy consumption dimension) × 0.35 + (weighted sum of collaboration dimension) × 0.25, where the weighted sum of each dimension = Σ (parameter value within the dimension × parameter weight, and the parameter weight within the dimension is evenly distributed).

[0087] The calculation process is as follows: Weighted sum of equipment health dimensions = (0.83×1 / 3 + 1.0×1 / 3 + 1.0×1 / 3) ≈ 0.943; Weighted sum of energy consumption dimensions = (1.1×1 / 2 + 1.0×1 / 2) = 1.05; Weighted sum of collaboration dimensions = (0.93×1 / 2 + 1.0×1 / 2) = 0.965; Comprehensive evaluation value = 0.943×0.4 + 1.05×0.35 + 0.965×0.25 ≈ 0.377 + 0.368 + 0.241 ≈ 0.986. Based on the comprehensive evaluation value of 0.986 ≥ 0.9, the equipment reliability level is determined to be Grade A, and it is marked as "The overall operation of the equipment is safe, and the risk of core components is controllable".

[0088] Based on energy consumption data and grid electricity prices, the quantified value of energy-saving benefits is calculated. The optimization potential for peak energy consumption is to reduce peak energy consumption from 80kWh to 72kWh (target value), with a monthly peak duration of 60 hours and an electricity price of 0.8 yuan / kWh. The quantified value of energy saving benefit is (80-72)×60×0.8=384 yuan. The optimization potential for no-load high energy consumption is to reduce no-load energy consumption from 15kWh to 9kWh (target value), with a monthly no-load duration of 480 hours and an electricity price of 0.3 yuan / kWh. The quantified value of energy saving benefit is (15-9)×480×0.3=864 yuan. The total quantified value of energy saving benefit is 384+864=1248 yuan / month. The final core evaluation result is generated, specifically, the equipment reliability level is A (comprehensive evaluation value 0.986), and the monthly quantified value of energy saving benefit is 1248 yuan. Weighted details for each dimension and the energy-saving calculation process are also attached to ensure the results are verifiable.

[0089] Based on the core assessment results, a visualization mechanism for assessment information is triggered, generating a visualized assessment report. This report includes an equipment reliability trend curve, an energy-saving benefit comparison bar chart, and a radar chart showing the health status of each core component. Specifically, the equipment reliability trend curve is generated by extracting the comprehensive assessment values ​​for the past six months (January 0.85, February 0.88, March 0.92, April 0.95, May 0.96, June 0.986). The curve is plotted with "month" on the horizontal axis and "comprehensive assessment value" on the vertical axis, with grade boundaries marked (Grade A 0.9, Grade B 0.8). The curve shows a gradual increase from Grade B in January to Grade A in June, indicating a positive trend. The accompanying text states, "Equipment reliability continues to improve, mainly due to predictive maintenance effectively reducing component risks."

[0090] The energy-saving benefit comparison bar chart is plotted with "Energy Consumption Type" (peak excess consumption, no-load high consumption, total energy saving) on ​​the horizontal axis and "Energy Saving Amount (Yuan / Month)" on the vertical axis. Peak excess consumption resulted in energy savings of 384 Yuan, no-load high consumption on 864 Yuan, and total energy savings of 1248 Yuan. The chart also compares the "Energy Consumption Before Optimization" with the "Target Energy Consumption After Optimization" (e.g., Peak excess consumption before optimization: 80 kWh, after optimization: 72 kWh). The accompanying text states, "No-load high consumption accounts for the highest proportion of energy savings (69.2%), requiring continuous optimization of wind turbine speed parameters."

[0091] The radar chart for the health status of each core component is plotted with "compressor, heat exchanger, fan, and filter" as the axes and "reliability coefficient" as the radial dimension (range 0-1.0). The radar chart shows the compressor at 0.7, the heat exchanger at 0.9, the fan at 1.0, and the filter at 0.7. The accompanying text states, "The reliability of the compressor and filter is low (0.7), requiring close attention to lubrication replenishment and filter cleaning." The report structure includes a cover (assessment period, assessment object), a summary of core assessment results (Grade A, 1248 RMB / month), detailed charts (trend curves, bar charts, radar charts), and optimization suggestions (e.g., "Prioritize compressor and filter maintenance in July"). All charts are labeled with data sources and calculation methods to ensure professionalism and readability.

[0092] Based on the visualized assessment report, the comprehensive assessment information of the industrial air conditioning operation status is distributed to multiple terminals, enabling the synchronous push of assessment information to the operation and maintenance management platform, energy monitoring center, and enterprise ERP system, and recording the complete assessment process in the industrial air conditioning operation and maintenance log. The operation and maintenance management platform pushes the complete visualized assessment report (including raw data and calculation details), allowing operation and maintenance personnel to view component risk details and energy-saving optimization paths, with "editable and exportable" permissions; it pushes the energy-saving benefit module (bar chart, total energy-saving amount) and energy consumption anomaly analysis to the energy monitoring center, allowing energy management personnel to track the achievement of energy-saving targets, with "viewable and statistical" permissions; and it pushes the core assessment results (reliability level, monthly energy-saving benefits) to the enterprise ERP system for cost accounting and budget planning, with "viewable and able to connect to the financial module" permissions. The distribution protocol uses MQTT (Internet of Things terminal) and HTTPS (Web platform) to ensure data transmission security, and the push delay is controlled within 1 minute.

[0093] The complete evaluation process is recorded in the industrial air conditioning operation and maintenance log, using "evaluation cycle code (202406)" as the key. The detailed evaluation process is recorded, and the values ​​include: input data: basic evaluation index data (component reliability coefficient, energy consumption loss coefficient), multi-dimensional evaluation matrix; calculation process: evaluation model weight setting, comprehensive evaluation value calculation details, energy saving benefit calculation formula; output results: core evaluation results (Grade A, 1248 yuan), visualization report link; anomaly description: no abnormal data is removed from the record, and all parameters are verified by the 3σ principle; the log format is JSON, which is convenient for subsequent query and data analysis, and the retention period is set to 3 years.

[0094] In one implementation, such as Figure 2 As shown, this application also provides a predictive maintenance and energy-saving linkage device for industrial air conditioning, comprising:

[0095] The acquisition module 201 is used to acquire the operating data and environmental parameters of each core component of the industrial air conditioner, including compressor start-stop frequency, evaporator / condenser temperature, fan speed, refrigerant pressure, indoor and outdoor temperature and humidity, load change curve and energy consumption metering data.

[0096] Processing module 202 is used to process the operating data and environmental parameters of the core components of the industrial air conditioner. It monitors the compressor lubricating oil level, heat exchanger scaling, motor winding insulation, and filter clogging status in real time through an IoT sensor network and edge computing nodes. It constructs a correlation model between component degradation trends and environmental load, generating preliminary maintenance needs and energy consumption anomaly warning information. Processing this preliminary maintenance needs and energy consumption anomaly warning information, it simulates the equipment operating status at different maintenance times using digital twin technology. Combining this with a full lifecycle cost model, it designs a tiered maintenance strategy, applies component remaining lifespan threshold constraints and an optimal maintenance resource allocation objective function, and generates target maintenance execution instructions and time-limited maintenance plans. Energy potential analysis report; Based on the corresponding adaptive algorithm, the target maintenance execution command and energy-saving potential analysis report are processed. Combining indoor thermal comfort requirements and the power grid peak-valley electricity pricing mechanism, the air conditioning operating parameters are adjusted in real time to generate a linkage control scheme that takes into account both maintenance effectiveness and energy consumption optimization; The linkage control scheme that takes into account both maintenance effectiveness and energy consumption optimization is processed to generate a globally coordinated operation management strategy; The preliminary maintenance needs and energy consumption anomaly early warning information, target maintenance execution commands, linkage control scheme that takes into account both maintenance effectiveness and energy consumption optimization, and globally coordinated operation management strategy are processed to generate comprehensive evaluation information of industrial air conditioning operation status, including equipment reliability level and energy-saving benefit quantification value.

[0097] The computer-readable storage medium provided in the above embodiments of this application and the predictive maintenance and energy-saving linkage method for industrial air conditioners provided in the embodiments of this application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.

[0098] 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 predictive maintenance and energy-saving linkage method for industrial air conditioners, electronic devices, electronic equipment, and readable storage media are basically similar to the embodiments of the predictive maintenance and energy-saving linkage method for industrial air conditioners described above, and therefore the descriptions are relatively simple. Relevant parts can be referred to in the descriptions of the embodiments of the predictive maintenance and energy-saving linkage method for industrial air conditioners described above.

Claims

1. A predictive maintenance and energy-saving linkage method for industrial air conditioning, characterized in that, include: Acquire operating data and environmental parameters of each core component of the industrial air conditioner, including compressor start / stop frequency, evaporator / condenser temperature, fan speed, refrigerant pressure, indoor and outdoor temperature and humidity, load change curves, and energy consumption metering data; The system processes the operating data and environmental parameters of the core components of industrial air conditioners, and monitors the compressor lubricating oil level, heat exchanger scaling degree, motor winding insulation and filter clogging status in real time through IoT sensor network and edge computing nodes. It constructs a correlation model between component deterioration trend and environmental load, and generates preliminary maintenance needs and energy consumption anomaly warning information. The system processes preliminary maintenance needs and energy consumption anomaly warning information, simulates equipment operation status at different maintenance times using digital twin technology, designs a graded maintenance strategy by combining the full life cycle cost model, applies component remaining life threshold constraints and maintenance resource optimal allocation objective function, and generates target maintenance execution instructions and energy-saving potential analysis reports. Based on the corresponding adaptive algorithm, the target maintenance execution command and energy-saving potential analysis report are processed. Combined with indoor thermal comfort requirements and the peak-valley electricity price mechanism of the power grid, the air conditioning operating parameters are adjusted in real time to generate a linkage control scheme that takes into account both maintenance effect and energy consumption optimization. Process the linkage control scheme that balances maintenance effectiveness and energy consumption optimization to generate a globally coordinated operation and management strategy; The system processes preliminary maintenance needs and energy consumption anomaly warning information, target maintenance execution instructions, linkage control schemes that balance maintenance effectiveness and energy consumption optimization, and global collaborative operation management strategies to generate comprehensive assessment information on the operating status of industrial air conditioners, including equipment reliability levels and quantitative values ​​of energy-saving benefits.

2. The method as described in claim 1, characterized in that, The system processes operational data and environmental parameters of core industrial air conditioning components. Through IoT sensor networks and edge computing nodes, it monitors compressor lubricating oil levels, heat exchanger scaling, motor winding insulation, and filter clogging in real time. It then constructs a model linking component degradation trends with environmental load, generating preliminary maintenance needs and early warning information for abnormal energy consumption, including: The system processes the operating data and environmental parameters of the core components of industrial air conditioners, and builds a distributed monitoring system through IoT sensor networks and edge computing nodes. The operating data of the core components and the environmental parameters are fused and calculated using a dynamic weight allocation mechanism to monitor the compressor lubricating oil level, the degree of scaling of the heat exchanger, the insulation of the motor windings, and the clogging status of the filter screen in real time. Using a component-environment dynamic correlation modeling framework, a three-level network structure is constructed, consisting of multi-source data noise reduction and feature selection in the input layer, spatiotemporal correlation learning in the hidden layer, and demand and early warning mapping in the output layer, to build a correlation model between component degradation trend and environmental load. Among them, the spatiotemporal correlation learning in the hidden layer introduces an LSTM network to capture the temporal correlation between component degradation and environmental load, and combines an attention mechanism to strengthen the weight of key influencing factors. The rationality of the model output results is verified by combining the safety operation threshold constraints of the core components of industrial air conditioners. Abnormal data caused by sensor drift, data transmission interference, etc. are eliminated by the 3σ principle. At the same time, the environmental factors are dynamically weighted based on the environmental load fluctuation coefficient to generate preliminary maintenance requirements and energy consumption anomaly warning information, including component maintenance priority, energy consumption anomaly type and risk level.

3. The method as described in claim 1, characterized in that, The system processes preliminary maintenance needs and energy consumption anomaly warnings, simulates equipment operating states at different maintenance times using digital twin technology, designs tiered maintenance strategies based on a full lifecycle cost model, applies component remaining lifespan threshold constraints and an objective function for optimal maintenance resource allocation, and generates target maintenance execution instructions and energy-saving potential analysis reports, including: Classify and label the component deterioration level data in the initial maintenance requirements, classify the maintenance timeliness and prioritize them, and generate basic parameters for component maintenance. The system identifies the anomaly type, decomposes the influencing factors, and calculates the energy-saving space for energy consumption fluctuation data in the energy consumption anomaly early warning information, and generates energy consumption optimization analysis parameters, including the feature matching model parameters of energy consumption anomalies, the influence coefficient of environmental load / equipment operating conditions / operating parameters on energy consumption, and the quantitative calculation rules for energy-saving potential under different anomaly scenarios. Based on digital twin technology, the operating status of equipment at different maintenance times is simulated and extrapolated. Through equipment performance simulation before / during / after maintenance, component life decay simulation under different maintenance intervals, and dynamic calculation of the impact of maintenance intervention on energy consumption, maintenance timing decision parameters are generated. By combining the full life cycle cost model, cost composition analysis, benefit return calculation and resource adaptation planning are carried out for the hierarchical maintenance strategy, and maintenance strategy optimization parameters are generated. By applying a component remaining life threshold constraint and a maintenance resource allocation objective function, the above parameters are integrated and optimized. The feasibility of the maintenance plan is verified by the remaining life lower limit value, and the maintenance task schedule is adjusted to maximize resource utilization. Target maintenance execution instructions and energy-saving potential analysis reports are generated.

4. The method as described in claim 1, characterized in that, Based on the corresponding adaptive algorithm, the target maintenance execution instructions and energy-saving potential analysis report are processed. Combining indoor thermal comfort requirements and the grid peak-valley electricity pricing mechanism, the air conditioning operating parameters are adjusted in real time to generate a linkage control scheme that balances maintenance effectiveness and energy consumption optimization, including: Adaptive feature matching and time-series correlation processing are performed on the maintenance time period requirements, component protection parameters, and energy consumption anomaly data and energy-saving target values ​​in the target maintenance execution instructions to generate maintenance demand features and energy consumption optimization features. Based on the characteristics of maintenance needs and energy consumption optimization, and combined with the temperature and humidity threshold range in indoor thermal comfort requirements, the weight coefficient of human comfort, and the time period division standard and electricity price fluctuation ratio in the power grid peak-valley electricity price mechanism, a dual-objective optimization weight for the adaptive algorithm is generated. Based on the parameter iteration output results of dual-objective optimization weights and adaptive algorithms, the air conditioning operating parameters are adjusted in real time to generate a linkage control scheme that balances maintenance effectiveness and energy consumption optimization.

5. The method as described in claim 4, characterized in that, The coordinated control scheme that balances maintenance effectiveness and energy consumption optimization is processed to generate a globally collaborative operation management strategy, including: Global feature normalization and collaborative correlation extraction are performed on the single-equipment operating parameter adjustment rules, maintenance and energy-saving collaborative thresholds, load distribution requirements of multiple air conditioning equipment in industrial scenarios, and energy supply limitation data in the linkage control scheme that takes into account both maintenance effect and energy consumption optimization, to generate standardized collaborative features and global trend features. Based on standardized collaborative features and global trend features, combined with the equipment cluster operation efficiency target, total energy consumption limit and environmental collaborative constraint mechanism in the global management requirements of industrial scenarios, workshop temperature and humidity area linkage standard and pollutant emission control indicators, a multi-dimensional adaptation weight of the global collaborative algorithm is generated. Based on the iterative calculation results of multi-dimensional adaptation weights and global collaborative algorithms, a globally collaborative operation management strategy is generated, including a cross-device staggered scheduling scheme for maintenance tasks of multiple air conditioning devices, dynamic energy allocation rules during high-load periods, multi-device linkage control instructions when regional temperature and humidity are abnormal, and the optimal scheduling plan for maintenance resources among cluster devices.

6. The method as described in claim 1, characterized in that, The system processes preliminary maintenance needs and energy consumption anomaly warning information, target maintenance execution instructions, linkage control schemes that balance maintenance effectiveness and energy consumption optimization, and globally coordinated operation management strategies to generate comprehensive industrial air conditioning operation status assessment information, including equipment reliability level and energy-saving benefit quantification values. The risk level and energy consumption anomaly type of the components in the preliminary maintenance needs and energy consumption anomaly early warning information are classified and quantified to generate basic evaluation index data. Among them, each component risk level corresponds to a unique reliability coefficient, and each energy consumption anomaly type corresponds to a target energy saving loss coefficient. The basic assessment index data, maintenance completion rate and resource utilization rate in target maintenance execution instructions, parameter compliance rate in linkage control scheme that balances maintenance effect and energy consumption optimization, and cluster collaboration efficiency in global collaborative operation management strategy are integrated and calculated to generate a multi-dimensional assessment matrix that includes equipment health dimension, energy consumption dimension, and collaboration dimension. In the assessment matrix, the row vector is the assessment dimension and the column vector is the specific quantitative value. The multi-dimensional evaluation matrix is ​​input into the industrial air conditioning operation status evaluation model for weighted calculation and processing to generate core evaluation results that include equipment reliability level and energy-saving benefit quantification value. Based on the core assessment results, an assessment information visualization generation mechanism is triggered to generate a visualization assessment report, which includes equipment reliability trend curves, energy-saving benefit comparison bar charts, and health status radar charts of each core component. Based on the visualization assessment report, the comprehensive assessment information of the industrial air conditioning operation status is distributed to multiple terminals, realizing the synchronous push of assessment information to the operation and maintenance management platform, energy monitoring center, and enterprise ERP system, and recording the complete assessment process in the industrial air conditioning operation and maintenance log.

7. A predictive maintenance and energy-saving linkage device for industrial air conditioning, characterized in that, The device includes: The acquisition module is used to acquire operating data and environmental parameters of each core component of the industrial air conditioner. The processing module processes operational data and environmental parameters of core industrial air conditioning components. It uses IoT sensor networks and edge computing nodes to monitor compressor lubricating oil levels, heat exchanger scaling, motor winding insulation, and filter clogging in real time. This allows for the construction of a model linking component degradation trends with environmental load, generating preliminary maintenance needs and energy consumption anomaly warnings. The module then processes these warnings, simulating equipment operation at different maintenance times using digital twin technology. Combining this with a lifecycle cost model, it designs tiered maintenance strategies, applying component remaining lifespan threshold constraints and an optimal maintenance resource allocation objective function to generate target maintenance execution instructions and energy-saving measures. Potential analysis report; Based on the corresponding adaptive algorithm, the target maintenance execution command and energy-saving potential analysis report are processed. Combining indoor thermal comfort requirements and the power grid peak-valley electricity pricing mechanism, the air conditioning operating parameters are adjusted in real time to generate a linkage control scheme that balances maintenance effectiveness and energy consumption optimization; The linkage control scheme that balances maintenance effectiveness and energy consumption optimization is processed to generate a globally coordinated operation management strategy; The preliminary maintenance needs and energy consumption anomaly early warning information, target maintenance execution commands, linkage control scheme that balances maintenance effectiveness and energy consumption optimization, and globally coordinated operation management strategy are processed to generate comprehensive evaluation information of industrial air conditioning operation status, including equipment reliability level and energy-saving benefit quantification value.

8. An electronic device, characterized in that, include: First processor; and memory for storing executable instructions of the first processor; The first processor is configured to execute the predictive maintenance and energy-saving linkage method for industrial air conditioning according to any one of claims 1 to 6 by executing the executable instructions.

9. 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 predictive maintenance and energy-saving linkage method for industrial air conditioning as described in any one of claims 1 to 6.

Images