Fresh air system processing method and system based on intelligent maintenance early warning and self-cleaning

Abstract

The invention provides an intelligent maintenance early warning and self-cleaning fresh air system processing method and system, and relates to the technical field of air purification, equipment operation data and environment data are collected by responding to a processing request of a target fresh air system, and component state risk assessment is carried out to generate system component early warning data; a self-cleaning strategy is matched and executed based on the early warning data, meanwhile, working condition operation data in the self-cleaning operation process are collected, fan health assessment is completed through a preset fan health assessment model in combination with equipment operation and environment data, and a health assessment report is generated. Through data-driven intelligent early warning and automatic cleaning, the problems that an existing fresh air system is low in efficiency due to the fact that maintenance depends on manpower, and the air purification effect and the user experience are affected due to the fact that dynamic adaptation cannot be achieved according to the state of components are solved; and the capability of dynamically monitoring the operation process of the fresh air system and pre-judging faults in advance is lacked. The technical effect of improving the operation reliability of the fresh air system is achieved.

Description

Technical Field

[0001] This invention relates to the technical field of air purification, and in particular to a fresh air system treatment method and system based on intelligent maintenance early warning and self-cleaning. Background Technology

[0002] As people's requirements for indoor air quality and living and working environments continue to increase, fresh air systems, as key equipment for improving indoor ventilation and purifying air, have been widely used in residential, commercial buildings and public facilities. They are of great significance for protecting the health of indoor occupants and improving environmental comfort.

[0003] Currently, although the fresh air systems on the market are equipped with core components such as filters and heat exchange cores to achieve the basic functions of air purification and energy recovery, the maintenance and management of the systems still adopt the traditional manual periodic inspection mode, lacking intelligent status monitoring, early warning and self-cleaning mechanisms.

[0004] However, existing operation and maintenance methods have many shortcomings: on the one hand, manual maintenance is inefficient and costly, and it is impossible to obtain the operating status of each component in the system in real time. It is also difficult to accurately control the maintenance time of filters and heat exchange cores. Often, due to untimely maintenance, filters become clogged and dust accumulates, and heat exchange cores become damp and breed bacteria, which not only reduces the air purification effect but also shortens the service life of components. On the other hand, the system lacks automated cleaning functions and cannot actively perform cleaning operations based on the actual degree of pollution of components, affecting the user experience. At the same time, for key power components such as fans, the existing system lacks the ability to dynamically monitor their operation and predict faults in advance. It is difficult to provide early warnings before faults occur, which can easily cause system operation interruptions and affect the overall stability and reliability. Summary of the Invention

[0005] To address the aforementioned shortcomings in existing technologies, the present invention aims to provide a fresh air system treatment method based on intelligent maintenance early warning and self-cleaning, which improves the operational reliability of the fresh air system.

[0006] The above-mentioned objective of this invention is achieved through the following technical solution: The fresh air system treatment method based on intelligent maintenance early warning and self-cleaning includes: In response to the operation monitoring command of the target fresh air system, acquire the equipment operation data and environmental data of the target fresh air system; The device operation data and the environmental data are used to conduct a component status risk assessment to obtain system component early warning data. The self-cleaning strategy of the system components is obtained by matching the system component early warning data, the equipment operation data and the environmental data. The system performs a self-cleaning operation on the target fresh air system according to the self-cleaning strategy of the system components, and obtains the self-cleaning operation data of the target fresh air system during the self-cleaning operation process. Based on the pre-set fan health assessment model, the fan health is assessed using the equipment operation data, the environmental data, and the self-cleaning operating data, and a health assessment report for the target fresh air system is generated.

[0007] By adopting the above technical solutions, data-driven intelligent early warning and automated cleaning significantly improve maintenance efficiency, reduce labor costs, and allow users to monitor system status in real time. Especially when air quality is poor, the equipment can quickly restore the filter's purification permeability through precise early warning and self-cleaning strategies, avoiding insufficient fresh air volume and decreased purification efficiency caused by filter blockage. This ensures that clean fresh air can be continuously supplied indoors even in harsh air conditions, allowing users to enjoy a fresh and comfortable breathing experience even when outdoor air quality is poor, without worrying about discomfort caused by indoor air pollution. At the same time, it can promptly remove component contamination, prevent bacterial growth, ensure air purification effect, and extend system life. Addressing the deficiency of untimely fan fault diagnosis, multi-dimensional health assessment integrating self-cleaning operating data can capture potential fault characteristics under dynamic fan load, predict faults in advance, and thus ensure stable system operation.

[0008] Preferably, the equipment operation data includes real-time differential pressure data of the filter in the target fresh air system, real-time humidity data of the heat exchange core, cumulative operating time of the filter and cumulative cleaning times of the heat exchange core, and the environmental data includes outdoor air quality index data, indoor and outdoor temperature difference data and current seasonal parameters.

[0009] By adopting the above technical solutions, the key operating parameters of the core components of the fresh air system and the corresponding external environmental parameters are accurately selected. This not only comprehensively covers the wear status and operating load of the filter and heat exchange core, but also takes into account the influence of external conditions such as outdoor air quality and seasons. This provides high-quality data support that is comprehensive and in line with actual operating conditions for component status risk assessment and self-cleaning strategy matching, avoiding the deviation in assessment and strategy adaptation caused by data loss.

[0010] Preferably, the step of using the equipment operating data and the environmental data to perform component status risk assessment and obtain system component early warning data includes: The filter coupling factor is determined based on the outdoor air quality index data, the cumulative operating time of the filter, and the real-time differential pressure data. The filter coupling factor and the preset filter basic early warning threshold are used as inputs to the preset filter multi-factor coupling dynamic threshold model to obtain the filter early warning value; Based on the filter warning value, the status risk is divided to obtain the filter risk level range, which includes a low-risk range, a medium-risk range, and a high-risk range. By comparing the real-time differential pressure data with the filter risk level range, the filter warning level is determined; The heat exchange core coupling factor is determined based on the indoor and outdoor temperature difference data, the current seasonal parameters, and the cumulative number of cleanings of the heat exchange core. The heat exchange core coupling factor and the preset heat exchange core basic early warning threshold are used as inputs to the preset heat exchange core multi-factor coupling dynamic threshold model to obtain the heat exchange core early warning value; Based on the warning value of the heat exchange core, the state risk is divided to obtain the risk level range of the heat exchange core. The risk level range of the heat exchange core includes a low risk range, a medium risk range, and a high risk range. By comparing the real-time humidity data with the risk level range of the heat exchange core, the early warning level of the heat exchange core is determined. The system component warning data includes the filter warning level and the heat exchange core warning level.

[0011] By adopting the above technical solutions, the accuracy of risk assessment can be improved by accurately matching the parameter selection based on component characteristics and combining it with a dynamic threshold model.

[0012] Preferably, determining the filter coupling factor based on the outdoor air quality index data, the cumulative operating time of the filter, and the real-time differential pressure data includes: The quality index weighting factor is determined based on the preset quality index weighting factor range to which the outdoor air quality index data belongs; The ratio of the cumulative running time of the filter to the preset standard cumulative running time is calculated, and the ratio result is rounded down and multiplied by the preset running time coefficient to obtain the running time attenuation coefficient. Based on the preset cleaning evaluation range to which the real-time differential pressure data belongs, determine the cleaning effect correction coefficient; The filter coupling factor includes the quality index weighting factor, the runtime attenuation coefficient, and the cleaning effect correction coefficient.

[0013] By adopting the above technical solution, parameters are decomposed and quantified into coefficients from the core impact dimension of filter contamination, avoiding the one-sidedness of single parameter evaluation, while ensuring the objectivity of coupling factors through standardized calculation.

[0014] Preferably, determining the heat exchange core coupling factor based on the indoor-outdoor temperature difference data, the current seasonal parameters, and the cumulative number of cleaning cycles of the heat exchange core includes: The ratio of the indoor and outdoor temperature difference data to the preset standard temperature difference data is calculated, and the ratio result is rounded down and multiplied by the preset temperature difference coefficient to obtain the temperature difference factor. The seasonal factor to which the current seasonal parameter belongs is used as the seasonal correction coefficient; The ratio of the cumulative number of cleanings of the heat exchange core to the preset standard cumulative number of cleanings is calculated, and the ratio result is rounded down and multiplied by the preset cumulative cleaning coefficient to obtain the cleaning number correction coefficient. The heat exchange core coupling factor includes the temperature difference factor, the seasonal correction factor, and the cleaning frequency correction factor.

[0015] By adopting the above technical solution, based on standardized ratio calculation and seasonal adaptation matching, the influence factors are quantified, avoiding subjective experience evaluation bias, so that the heat exchange core coupling factor can accurately characterize the degree of influence of the core's actual working conditions.

[0016] Preferably, the step of matching the system component early warning data, the equipment operation data, and the environmental data to obtain the system component self-cleaning strategy includes: A first target key is generated using the filter warning level and the outdoor air quality index data; The first target key is used to retrieve the preset filter self-cleaning strategy database and match the filter self-cleaning strategy. A second target key is generated using the aforementioned heat exchange core warning level and the real-time humidity data; The second target key is used to retrieve the preset heat exchange core self-cleaning strategy database and match the heat exchange core self-cleaning strategy; The self-cleaning strategy for system components includes the filter self-cleaning strategy and the heat exchange core self-cleaning strategy.

[0017] By adopting the above technical solution, the system can accurately anchor the strategy retrieval by associating the component warning level with real-time operating data through the target key. At the same time, the system can adapt the cleaning scheme to different components to avoid the adaptation deviation of the unified strategy. This ensures the self-cleaning effect, reduces component wear, and improves the pertinence and efficiency of system operation and maintenance.

[0018] Preferably, the step of performing a fan health assessment based on a pre-set fan health assessment model, using the equipment operation data, the environmental data, and the self-cleaning operating condition data, to generate a health assessment report for the target fresh air system includes: The equipment operation data, the environmental data, and the self-cleaning operation data are used for data preprocessing to obtain a time-series feature dataset. Feature extraction is performed on the time-series feature dataset to obtain wind turbine health status features and wind turbine dynamic load capacity features; The wind turbine health status characteristics and the wind turbine dynamic load capacity characteristics are fused to obtain a fused feature matrix; The fusion feature matrix is ​​used as input to the pre-set fan health assessment model to perform fan health assessment and generate a health assessment report for the target fresh air system.

[0019] By adopting the above technical solutions, integrating multi-dimensional operation and environmental data, and strengthening data representation capabilities through a complete link of preprocessing, feature extraction, and fusion, the health assessment of wind turbines becomes more comprehensive and accurate. At the same time, it enables dynamic judgment of operating conditions, providing a scientific basis for predicting faults and optimizing operation and maintenance of fresh air systems.

[0020] Preferably, the pre-set fan health assessment model includes an input adaptation layer, a dual-branch temporal extraction layer, a working condition attention fusion layer, and a multi-task decision layer. The step of using the fused feature matrix as input to the pre-set fan health assessment model to perform fan health assessment and generate a health assessment report for the target fresh air system includes: The fused feature matrix is ​​adjusted in dimension and format by the input preprocessing layer to obtain a standard input feature sequence. The dual-branch time-series extraction layer extracts time-series features from the standard input feature sequence to obtain conventional health time-series features and dynamic load time-series features. The working condition attention fusion layer performs weight allocation and feature fusion on the conventional health time-series features and the dynamic load time-series features to obtain weighted fused health features. The multi-task decision layer performs fault classification, remaining lifetime calculation and dynamic performance scoring on the weighted fusion health features to obtain fault type probability, remaining lifetime and dynamic load score. By integrating the failure type probability, the remaining operating life, and the dynamic load score, a health assessment report for the target fresh air system is generated.

[0021] By adopting the above technical solution, the layered architecture is used to accurately process features, the dual-branch extraction is used to effectively separate the two types of core temporal features, the working condition attention mechanism dynamically strengthens the influence weight of key working conditions, and the multi-task decision-making outputs multi-dimensional evaluation results in a synchronized manner, taking into account the accuracy, comprehensiveness and efficiency of the evaluation.

[0022] Preferably, the self-cleaning operating data includes the fan current fluctuation amplitude, fan speed adjustment delay data, and fan load coefficient.

[0023] By adopting the above technical solutions, the core operating parameters of the wind turbine during self-cleaning are accurately collected, comprehensively covering the three key dimensions of wind turbine operation stability, speed response and load-bearing capacity. This provides real and accurate operating condition data support for wind turbine health assessment and helps to accurately identify potential wind turbine failure risks.

[0024] The second objective of this invention is to provide a fresh air system treatment system based on intelligent maintenance early warning and self-cleaning, which has the characteristic of improving the operational reliability of the fresh air system.

[0025] The second objective of this invention is achieved through the following technical solution: A fresh air system based on intelligent maintenance early warning and self-cleaning includes: The response module is used to respond to the operation monitoring command of the target fresh air system and to acquire the equipment operation data and environmental data of the target fresh air system. The assessment module is used to perform component status risk assessment using the equipment operation data and the environmental data, and obtain system component early warning data; The matching module is used to match the self-cleaning strategy using the system component early warning data, the equipment operation data and the environmental data to obtain the system component self-cleaning strategy; The execution module is used to perform a self-cleaning operation on the target fresh air system according to the self-cleaning strategy of the system components, and to acquire the self-cleaning operation data of the target fresh air system during the self-cleaning operation process. The reporting module is used to perform a health assessment of the fan based on a pre-set fan health assessment model, using the equipment operation data, the environmental data, and the self-cleaning operating data, and generate a health assessment report for the target fresh air system.

[0026] By adopting the above technical solutions, the entire process from data acquisition, risk assessment, strategy matching to cleaning execution and health assessment is automated. This not only improves the adaptability of maintenance warnings and self-cleaning through accurate data support, but also reduces the difficulty of system operation and maintenance through modular design, thus fully ensuring the stable and efficient operation of the fresh air system.

[0027] In summary, the present invention has at least one of the following beneficial technical effects: 1. This invention, through intelligent early warning and automatic cleaning, can ensure fresh indoor air in real time and continuously provide clean fresh air even in inclement weather. At the same time, it effectively improves maintenance efficiency, reduces labor costs, and can detect potential fan malfunctions in advance, extending equipment lifespan and ensuring stable system operation.

[0028] 2. This invention achieves precise matching of cleaning strategies by accurately linking component warning levels with real-time operating data. At the same time, it customizes exclusive cleaning solutions for different components, avoiding one-size-fits-all adaptation biases. This ensures efficient cleaning while reducing component wear and tear, improving the targeted nature and overall efficiency of system maintenance. Attached Figure Description

[0029] Figure 1 This is a flowchart illustrating the steps of a fresh air system treatment method based on intelligent maintenance early warning and self-cleaning provided in Embodiment 1 of the present invention.

[0030] Figure 2 This is a structural block diagram of a fresh air system treatment system based on intelligent maintenance early warning and self-cleaning, provided in Embodiment 2 of the present invention. Detailed Implementation

[0031] This invention provides a method and system for handling fresh air systems based on intelligent maintenance early warning and self-cleaning. It addresses the technical problems of existing fresh air systems, such as low efficiency due to reliance on manual maintenance, inability to dynamically adapt to component status affecting air purification effects and user experience, and lack of dynamic monitoring and early fault prediction capabilities during system operation. It effectively improves the reliability of fresh air system operation.

[0032] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0033] It should be noted that the terms "first," "second," etc., used in this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The implementations described in the following exemplary embodiments do not represent all implementations consistent with this disclosure.

[0034] Furthermore, the term "and / or" in this article only describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects are in an "or" relationship.

[0035] The widespread use of fresh air systems is crucial for improving indoor air quality. However, existing systems rely on regular manual inspections for maintenance and cleaning, lacking intelligent early warning and automated cleaning functions, resulting in low maintenance efficiency and making it difficult for users to monitor the system status in real time.

[0036] In existing technologies, fresh air systems are usually equipped with basic filtration devices, but do not integrate automated maintenance warning and self-cleaning modules.

[0037] High manual maintenance costs and long maintenance cycles may lead to a decline in system performance; users may not receive timely information about system status, affecting the user experience; the lack of efficient self-cleaning function makes it easy for dust to accumulate and bacteria to grow on the filter and heat exchange core, reducing the air purification effect and service life.

[0038] Therefore, the fresh air system treatment method and system based on intelligent maintenance early warning and self-cleaning provided by the present invention adopts a multi-factor coupled dynamic early warning algorithm to achieve accurate state risk prediction of filters and heat exchange cores, and provide early warning of maintenance needs, thereby reducing the cost of manual inspection and improving the level of intelligent system operation and maintenance. Through a self-cleaning strategy database matching mechanism driven by "early warning level + operating condition parameters (i.e., outdoor air quality index data and real-time humidity data)," the filter and heat exchange core are automatically and precisely cleaned. This not only extends the service life of the components but also restores the purification capacity of the components in a timely manner when the air quality is poor, continuously ensuring the air purification effect and significantly improving the stability of indoor air quality. Furthermore, by integrating the improved Bi-LSTM-Attention model (i.e., the pre-set fan health assessment model) with the self-cleaning condition correlation data, the fan is subjected to multi-dimensional health assessment, which makes the fan fault diagnosis more timely and accurate, captures potential faults under dynamic load in advance, and significantly improves the overall stability and reliability of the system. By combining a dynamic self-cleaning strategy with full-process data-driven health management, users can monitor the system status in real time. Especially when air quality fluctuates, the adaptive handling capability of the fresh air system is fully utilized, ensuring continuous improvement in indoor air quality and greatly enhancing the user experience. Example

[0039] Please see Figure 1 The present invention provides a fresh air system treatment method based on intelligent maintenance early warning and self-cleaning, comprising: Step 101: Respond to the operation monitoring command of the target fresh air system and obtain the equipment operation data and environmental data of the target fresh air system.

[0040] The target fresh air system refers to a specific fresh air device that needs to undergo intelligent maintenance and early warning, self-cleaning control and fan health assessment. It can be specific to different types such as residential wall-mounted fresh air system and commercial ceiling-mounted fresh air system. It needs to be pre-configured with data acquisition module, control module and communication module to support data acquisition and command response. Operation monitoring instructions refer to control instructions that trigger the collection of operating data and environmental data of the target fresh air system. There are two ways to generate these instructions: one is to generate them automatically according to a preset cycle, which can be set according to actual needs; the other is to generate them based on specific triggering conditions (such as when a user actively initiates a "system status query" request through a mobile APP, or when the system self-checks and finds that the data cache amount has reached a threshold). The instructions include the time range of data collection and the data type identifier. Equipment operation data refers to a set of parameters that reflect the real-time operating status of the core functional components of the target fresh air system. Specifically, it includes four key data: real-time differential pressure data of the filter, real-time humidity data of the heat exchange core, cumulative operating time of the filter, and cumulative cleaning times of the heat exchange core. The real-time differential pressure data of the filter is the pressure difference between the air inlet and outlet sides of the filter, in Pascals (Pa). It is used to directly characterize the degree of clogging of the filter. This data is collected by micro differential pressure sensors installed on the air inlet and outlet sides of the filter, and the collection frequency is consistent with the operation monitoring command cycle (i.e., once every 1 minute). The real-time humidity data of the heat exchange core is the relative humidity value of the surface of the heat exchange core, in percentage (%RH), which is used to reflect the risk of bacteria or mold growth in the heat exchange core due to moisture. It is collected by a capacitive humidity sensor embedded in the middle of the heat exchange core, and the collection frequency is once every minute. The cumulative runtime of the filter is the cumulative time that the filter has been running with the target fresh air system since its installation or last replacement, in hours (h). It is automatically counted by the timing unit in the control module of the target fresh air system and the data is updated every hour. The cumulative number of cleaning times for the heat exchange core is the total number of times the target fresh air system has performed a self-cleaning operation on the heat exchange core since its installation or last maintenance. It is counted by the counting unit in the control module, and is automatically incremented by 1 after each self-cleaning operation is completed. Environmental data refers to the set of external environmental parameters that affect the operating status of the target fresh air system and the pollution risk of its components. Specifically, it includes three key data: outdoor air quality index (AQI) data, indoor-outdoor temperature difference data, and current seasonal parameters. Outdoor air quality index (AQI) is a dimensionless index that comprehensively reflects the concentration of pollutants such as PM2.5, PM10, and sulfur dioxide in outdoor air. It is used to determine the degree of outdoor air pollution. This data is collected by a laser scattering AQI sensor installed at the fresh air inlet of the target fresh air system. Considering that the outdoor air quality changes relatively slowly, the collection frequency is set to once every 5 minutes. After collection, the data is transmitted to the system data buffer in real time. The indoor-outdoor temperature difference data is the difference between the indoor ambient temperature and the outdoor ambient temperature of the target fresh air system, in degrees Celsius (°C). It is used to evaluate the heat exchange efficiency and condensation risk of the heat exchange core. It is calculated by simultaneously collecting temperature data from temperature and humidity sensors installed at the indoor air outlet and the outdoor air inlet. The collection frequency is consistent with the AQI data (once every 5 minutes). The current season parameter is the current season identifier of the region where the target fresh air system is located. It is divided into four categories: spring (March-May), summer (June-August), autumn (September-November), and winter (December-February). The system control module automatically determines the season by reading the date information from the built-in clock module, and the determination result is updated once a day.

[0041] In this embodiment of the invention, after receiving the operation monitoring instruction, the control module of the target fresh air system first parses the data type identifier and time range requirement in the instruction, and then sends data reading requests to the corresponding sensors or functional units: for real-time differential pressure data of the filter and real-time humidity data of the heat exchange core, the latest collected value at the current moment is directly read from the differential pressure sensor and the capacitive humidity sensor; for the cumulative running time of the filter and the cumulative cleaning times of the heat exchange core, the cumulative statistical value is read from the timing unit and counting unit of the control module; for outdoor AQI data and indoor-outdoor temperature difference data, the most recent (within 5 minutes) collection and calculation results are read from the data buffer; for the current seasonal parameters, the latest identifier is read from the seasonal determination result associated with the clock module; after all the read data is format-verified by the control module (such as confirming whether the data unit and numerical range meet the preset standards to avoid abnormal data interference), it is integrated to form a complete set of equipment operation data and environmental data.

[0042] Step 102: Use equipment operation data and environmental data to conduct component status risk assessment and obtain system component early warning data.

[0043] Component status risk assessment refers to the process of quantitatively analyzing the risks currently faced by core components of a fresh air system (such as filters, heat exchange cores, etc.) based on equipment operation data and environmental data, through a pre-set assessment model (such as a multi-factor weighted assessment model) or rule base, including pollution, performance degradation, and potential failures. The assessment dimensions cover the degree of component wear, the intensity of environmental impact on components, and the aggravating effect of equipment operating load on components. The core purpose is to transform scattered raw data into intuitive risk judgment criteria. System component early warning data is a comprehensive data set output after component status risk assessment. It includes key information such as the risk level (low, medium, high), risk type (e.g., filter clogging risk, heat exchange core condensation and mold risk), risk development trend prediction, and risk critical threshold for each system component. It is the core input data for subsequent self-cleaning strategy matching and can directly guide the system to determine whether cleaning is needed, cleaning priority, and cleaning method.

[0044] Furthermore, the equipment operation data includes real-time differential pressure data of the filter in the target fresh air system, real-time humidity data of the heat exchange core, cumulative filter operating time, and cumulative number of cleanings of the heat exchange core; environmental data includes outdoor air quality index data, indoor-outdoor temperature difference data, and current seasonal parameters; and system component warning data includes filter warning level and heat exchange core warning level. Step 102 may include the following sub-steps: S11. Determine the filter coupling factor based on outdoor air quality index data, cumulative filter operating time, and real-time differential pressure data.

[0045] The filter coupling factor is a key composite parameter used to dynamically calculate the real-time adaptation warning threshold of the filter. Its core function is to dynamically adjust the basic warning threshold of the filter by combining three actual operating conditions: outdoor air quality, filter usage time, and filter cleaning effect. This makes the adjusted warning threshold more closely match the actual pollution state and risk level of the filter, providing a precise quantitative basis for the risk assessment of the filter status. The filter coupling factor includes three sub-factors: quality index weighting factor, runtime attenuation coefficient, and cleaning effect correction coefficient. Each sub-factor reflects the impact of operating conditions on the filter warning threshold from different dimensions, working together to achieve dynamic adaptation of the threshold.

[0046] Furthermore, the filter coupling factor includes a quality index weighting factor, a runtime attenuation coefficient, and a cleaning effect correction coefficient. S11 may include the following sub-steps: S111. Determine the weighting factor of the air quality index based on the preset weighting factor interval of the outdoor air quality index data.

[0047] The preset quality index weighting factor range refers to a set of rules pre-set in the target fresh air system control module, which divides the AQI value into different ranges and corresponds to specific weighting factors, based on experimental data on the correlation between filter pollution rate and AQI. The core logic of this range division is: the lower the AQI value, the slower the outdoor pollutants pollute the filter, and the smaller the corresponding weighting factor; conversely, the higher the AQI value, the larger the weighting factor. The quality index weighting factor is a parameter used to quantify the impact of outdoor air quality on the filter warning threshold. Its value ranges from 0.2 to 1.0. This factor can be used to dynamically adjust the filter warning threshold, so that the threshold is appropriately increased in high-pollution environments to provide early warning of the risk of filter clogging, and appropriately decreased in low-pollution environments to avoid over-warning.

[0048] In this embodiment of the invention, the control module of the target fresh air system first reads the latest outdoor air quality index data collected in step 101 from the data cache, and then the control module calls the built-in "preset quality index weighting factor interval" rule for matching: The preset rules are specifically divided into three intervals. The first interval is "AQI≤50" (corresponding to the "Excellent" level in the air quality standard). In this interval, the outdoor pollutant concentration is low, the filter adsorbs fewer pollutants per unit time, and the pollution rate is slow. Therefore, the corresponding quality index weighting factor is 0.2. The second interval is "50<AQI≤100" (corresponding to the "Good" level in the air quality standard). In this interval, the outdoor pollutant concentration is moderate, and the filter pollution rate is at a moderate level. Therefore, the corresponding quality index weighting factor is 0.5. The third interval is "AQI>100" (corresponding to the "Light Pollution" level and above in the air quality standard). In this interval, the outdoor pollutant concentration is high, the filter easily and quickly adsorbs pollutants, and the pollution rate is fast. Therefore, the corresponding quality index weighting factor is 1.0.

[0049] For example, if the AQI value read is 65, it is compared with the three intervals mentioned above one by one to determine that the AQI value belongs to the interval "50 < AQI ≤ 100", and the corresponding quality index weighting factor is determined to be 0.5; if the AQI value read is 30, it is determined to belong to the interval "AQI ≤ 50", and the corresponding weighting factor is 0.2; if the AQI value read is 120, it is determined to belong to the interval "AQI > 100", and the corresponding weighting factor is 1.0. The determined quality index weighting factor... The parameters will be temporarily stored in the parameter cache area, and together with the runtime attenuation coefficient and cleaning effect correction coefficient determined in subsequent steps, they will constitute the filter coupling factor, which will be used to calculate the real-time filter adaptation warning threshold in the next step.

[0050] S112. The ratio of the cumulative running time of the filter to the preset standard cumulative running time is calculated, and the ratio result is rounded down and multiplied by the preset running time coefficient to obtain the running time attenuation coefficient.

[0051] The preset standard cumulative running time refers to the baseline time set in the system in advance to measure the filter usage progress based on the filter life test data of filter material (such as HEPA filter, pre-filter). Its value must match the filter's normal effective filtration cycle. The preset running time coefficient (dimensionless) refers to a proportional parameter that is determined in advance through experimental verification and is used to convert the ratio of the cumulative running time of the filter to the preset standard cumulative running time into the attenuation coefficient. The runtime decay coefficient (dimensionless) refers to the parameter that ultimately quantifies the impact of the filter's cumulative runtime on its warning threshold. To avoid the coefficient being too high and the warning threshold being distorted due to excessively long filter usage, its upper limit is set to 1.0.

[0052] Convert S112 into a formulaic form, specifically as follows:

[0053] In the formula, Indicates the attenuation coefficient over runtime. Indicates the cumulative running time of the filter. This indicates the cumulative runtime of the preset standard. This represents the preset runtime coefficient. This represents the floor function. This represents the minimum value function, used to ensure that the attenuation coefficient does not exceed the upper limit of 1.0.

[0054] In this embodiment of the invention, firstly, the control module of the target fresh air system reads the latest cumulative filter runtime from the timing unit, and simultaneously retrieves the pre-configured preset standard cumulative runtime and preset runtime coefficient from the system preset parameter library; then, the control module performs a ratio calculation according to the formula to calculate... This ratio reflects the multiple of the filter's cumulative runtime relative to the preset baseline runtime, providing an initial indication of the filter's usage progress. Subsequently, the ratio is rounded down to convert non-integer multiples of runtime into integer cycles, avoiding fluctuations in the attenuation coefficient due to decimal parts and better reflecting the actual pattern of gradual attenuation of filter capacity over a complete usage cycle. The rounded-down cycle number is then multiplied by the preset runtime coefficient to obtain a preliminary attenuation coefficient value. This value increases linearly with the number of filter usage cycles, quantifying the degree of attenuation at different usage stages. Finally, a minimum value function is used to limit the upper limit of the preliminary attenuation coefficient value. If the preliminary value does not exceed 1.0, it is directly used as the final runtime attenuation coefficient; if the preliminary value exceeds 1.0, it is fixed at 1.0, ensuring that even with prolonged use, the attenuation coefficient does not deviate from the actual risk range. The calculated runtime attenuation coefficient is temporarily stored in the parameter cache of the control module.

[0055] S113. Determine the cleaning effect correction coefficient based on the preset cleaning assessment range to which the real-time differential pressure data belongs.

[0056] The preset cleaning assessment interval refers to the basic warning threshold of the filter. This refers to the critical pressure difference value for filter clogging calibrated in the laboratory, which is a set of pressure difference intervals set in advance in the target fresh air system control module to divide the cleaning effect level. The interval division is based on the correlation data of "pressure difference reduction and actual filter permeability" in a large number of cleaning experiments. The purpose is to intuitively determine the cleaning effect through pressure difference data. The cleaning effect correction coefficient (dimensionless) is a parameter used to quantify the impact of filter cleaning effect on its warning threshold. Its value is positively correlated with the cleaning effect. The better the cleaning effect, the smaller the coefficient, and the smaller the increase in the warning threshold. A fixed value is set for different cleaning effect levels.

[0057] In this embodiment of the invention, the control module of the target fresh air system first retrieves the real-time differential pressure data (denoted as ) of the filter after its most recent self-cleaning from the data storage unit. At the same time, it calls the pre-stored filter basic warning threshold and corresponding preset cleaning assessment interval rules in the system: The preset rule is specifically divided into three intervals: the first interval is... Within this range, the pressure difference after filter cleaning is far below 30% of the basic warning threshold, indicating that the filter media contaminants have been thoroughly removed and the permeability has been well restored. Therefore, the cleaning effect is deemed satisfactory, and the corresponding cleaning effect correction factor is [not specified]. ; The second interval is Within this range, the pressure difference after filter cleaning is between 30% and 40% of the basic warning threshold, indicating that some contaminants have been removed from the filter media and the permeability has been partially restored. This is considered a basic achievement of the cleaning effect, corresponding to a cleaning effect correction factor. ; The third interval is If the pressure difference after filter cleaning exceeds 40% of the basic warning threshold within this range, it indicates that the filter media contaminants were not thoroughly removed and the permeability recovery was poor. This is judged as a failure to meet cleaning standards, and a corresponding cleaning effect correction factor is applied. .

[0058] The determined cleaning effect correction factor The parameters will be temporarily stored in the parameter cache area of ​​the control module.

[0059] S12. Input the filter coupling factor and the preset filter basic warning threshold into the preset filter multi-factor coupling dynamic threshold model to obtain the filter warning value.

[0060] The pre-set filter multi-factor coupling dynamic threshold model refers to a mathematical model pre-embedded in the algorithm library of the target fresh air system control module, used to combine filter coupling factors with basic warning thresholds to calculate dynamic warning values. Its core logic is: through... and The effect of quantitative chemical conditions on the risk of filter contamination, through The effectiveness of quantifying cleaning results in mitigating risks is used to ultimately output dynamic early warning thresholds that are adapted to the current working conditions. Filter warning value This refers to the dynamic threshold calculated through a multi-factor coupling model, which is used to compare with the real-time differential pressure data of the filter to determine whether the filter is in a low, medium, or high risk state.

[0061] The pre-set filter multi-factor coupled dynamic threshold model is converted into a formula encapsulation form, specifically:

[0062] In the formula, The filter warning values ​​of 0.1 and 0.05 were determined through extensive testing under various operating conditions. These values ​​are used to balance the influence weights of each coupling factor on the warning value, ensuring that the adjustment range of the warning value matches the actual risk changes.

[0063] In this embodiment of the invention, the control module of the target fresh air system first retrieves the previously stored filter coupling factor from the parameter cache. , , At the same time, it retrieves the preset filter basic warning threshold that matches the current filter material from the system preset parameter library. Next, the control module calls the preset filter multi-factor coupled dynamic threshold model, substitutes the above parameters into the formula, and calculates the filter warning value. .

[0064] S13. Based on the filter warning value, classify the status risk to obtain the filter risk level range, which includes the filter low risk range, the filter medium risk range, and the filter high risk range.

[0065] State risk classification refers to the process by which the target fresh air system control module compares the real-time differential pressure data of the filter with the filter warning value and classifies the current state of the filter into different risk levels according to preset rules. The classification is based on the correlation between the "differential pressure ratio and the actual degree of filter blockage" verified by a large number of experiments. The lower the ratio, the better the filter permeability and the lower the risk of blockage, and vice versa. The filter risk level range refers to a set of filter risk ranges, including low risk, medium risk, and high risk ranges, obtained by classifying the state risk. Each range corresponds to a specific differential pressure ratio range and risk response strategy. The low-risk range of the filter refers to the real-time differential pressure data of the filter (denoted as...). (Unit: Pa) is located in Within this range, although a small amount of pollutants may accumulate on the filter, the permeability is still sufficient to meet the normal fresh air purification needs, and the risk of clogging is low. The risk range in the filter refers to the range in which the real-time differential pressure data of the filter is within a certain range. The filter is in a range where the amount of pollutants accumulated is close to the critical value and the permeability begins to decrease. If not dealt with in time, it may affect the fresh air volume and purification efficiency, and the risk of blockage is moderate. The high-risk range for the filter refers to the real-time differential pressure data of the filter being in a certain range. In this range, the accumulation of pollutants on the filter has reached a critical state, the permeability has decreased significantly, and it may even lead to insufficient fresh air volume, secondary pollution of the filter, and a high risk of clogging.

[0066] In this embodiment of the invention, the control module of the target fresh air system first retrieves the filter warning value from step S12 and the latest real-time filter pressure difference data from the data cache, and then calculates two key ratio thresholds: , The filter is divided into low-risk, medium-risk, or high-risk zones.

[0067] S14. Compare the real-time differential pressure data with the filter risk level range to determine the filter warning level.

[0068] In this embodiment of the invention, the control module of the target fresh air system first retrieves the latest real-time filter differential pressure data from the data acquisition unit. Simultaneously, the filter risk level range and corresponding filter warning value stored in step S13 are retrieved from the system decision parameter area. The low-risk range is The medium-risk range is High-risk areas are Next, the control module initiates the interval comparison logic, which compares the real-time differential pressure data. The data is matched against the numerical ranges of the three risk zones one by one: first, it is determined whether it belongs to the low-risk zone; if not, it is determined whether it belongs to the medium-risk zone; if so, the current warning level of the filter is directly determined to be the medium warning level of the filter. If the collected real-time differential pressure matches the low-risk zone range after comparison, it is determined to be the low warning level of the filter. If it matches the high-risk zone range, it is determined to be the high warning level of the filter.

[0069] S15. Determine the coupling factor of the heat exchange core based on indoor and outdoor temperature difference data, current seasonal parameters, and cumulative cleaning times of the heat exchange core.

[0070] The heat exchange core coupling factor is a composite parameter used in the performance evaluation and condition control of heat exchange equipment to comprehensively quantify the impact of three key factors—indoor and outdoor temperature difference, current season, and cumulative cleaning frequency of the heat exchange core—on the actual working efficiency, energy consumption, and condition stability of the heat exchange core. Its core is to integrate three sub-dimensional parameters—temperature difference factor, seasonal correction coefficient, and cleaning frequency correction coefficient—to transform the independent influence of a single factor into a comprehensive quantitative indicator that can be directly used for calculation and judgment. Ultimately, this provides data support for performance degradation early warning, cleaning cycle determination, and operating parameter optimization of the heat exchange core, ensuring that the heat exchange core always operates in a highly efficient state adapted to actual working conditions.

[0071] Furthermore, the heat exchange core coupling factor includes a temperature difference factor, a seasonal correction factor, and a cleaning frequency correction factor. S15 may include the following sub-steps: S151. The ratio of indoor and outdoor temperature difference data to preset standard temperature difference data is calculated, and the ratio result is rounded down and multiplied by the preset temperature difference coefficient to obtain the temperature difference factor.

[0072] The preset standard temperature difference data is a reference temperature difference value calibrated experimentally based on the design conditions of the heat exchange core. It is used to convert the actual temperature difference into a quantifiable load multiple, and is usually preset to 5℃. The preset temperature difference coefficient (dimensionless) is a proportional parameter determined in advance through a large number of temperature difference-load correlation experiments. It is used to convert the multiple of the actual temperature difference and the standard temperature difference into a quantified degree of temperature difference influence. It is usually set to 0.05. The temperature difference factor (dimensionless) is a sub-parameter that ultimately quantifies the impact of indoor and outdoor temperature differences on the heat exchange core. The larger the value, the stronger the superimposed effect of the current temperature difference on the core load.

[0073] Convert S151 into a formulaic form, specifically:

[0074] In the formula, Indicates the temperature difference factor. This represents the indoor-outdoor temperature difference data. This represents the preset standard temperature difference data. This indicates the preset temperature difference coefficient.

[0075] In this embodiment of the invention, the environmental sensing module of the target fresh air system first initiates the indoor and outdoor temperature synchronous acquisition process: controlling the indoor temperature sensor to collect the current real-time indoor temperature, and simultaneously controlling the outdoor temperature sensor to collect the current real-time outdoor temperature. Then, the system control module calculates the indoor and outdoor temperature difference data according to the formula current real-time indoor temperature - current real-time outdoor temperature, and temporarily stores this data in the environmental data cache area. Next, the system control module retrieves the exclusive parameter matching the current heat exchange core model—the preset standard temperature difference data—from the preset parameter library. With preset temperature difference coefficient This ensures the accuracy and compatibility of parameter calls; subsequently, the system control module initiates the calculation process according to a proprietary mathematical formula: the first step is to cache the environmental data in the buffer. With the retrieved The first step is to perform a ratio calculation to obtain the multiple relationship between the two; the second step is to perform a floor operation on this multiple relationship, taking the largest integer not greater than the multiple, to match the characteristics of the core load changing with the complete standard temperature difference range; the third step is to compare the floor result with the retrieved... Perform a multiplication operation to obtain the final temperature difference factor.

[0076] S152. Match the seasonal factor to which the current seasonal parameter belongs as the seasonal correction coefficient.

[0077] The seasonal factor refers to a dimensionless coefficient stored in the system's preset parameter library, corresponding to different seasons. Its value is determined based on a large amount of experimental data. The core logic is: the greater the impact of the seasonal environment on the heat exchange core, the higher the factor value. For example, the high humidity environment in summer easily leads to the growth of bacteria and rapid accumulation of dirt in the core, so the factor value is higher than that in spring and autumn; the low temperature environment in winter easily leads to frost formation on the core, affecting heat exchange efficiency, so the factor value is next; the temperature and humidity are suitable in spring and autumn, so the impact on the core is the least, and the factor value is the lowest. The seasonal correction factor refers to the final sub-parameter obtained by matching the current seasonal parameters with the seasonal factors. It is directly used to calculate the coupling factor of the heat exchange core and quantifies the comprehensive influence of the season on the core.

[0078] In this embodiment of the invention, the control module of the target fresh air system first sends a date read request to the built-in clock module to obtain the current system date. Then, it calls the preset "date-seasonal parameter mapping rule" to determine the corresponding current seasonal parameter S based on the current date. For example, if the current date is July 10th, it is determined to be summer, corresponding to S=2. Next, the control module retrieves the "seasonal parameter-seasonal factor mapping table" from the system's preset parameter library. This table pre-stores the seasonal factor values ​​corresponding to each seasonal parameter (for example, the preset mapping relationship is: S=1 (spring) corresponds to a seasonal factor of 0.95, S=2 (summer) corresponds to a seasonal factor of 1.1, S=3 (autumn) corresponds to a seasonal factor of 0.95, and S=4 (winter) corresponds to a seasonal factor of 1.05). Subsequently, the control module precisely matches the determined current seasonal parameter S with the seasonal parameters in the mapping table, extracts the corresponding seasonal factor, and directly determines this seasonal factor as the seasonal correction coefficient. .

[0079] S153. The ratio of the cumulative number of cleanings of the heat exchange core to the preset standard cumulative number of cleanings is calculated, and the ratio result is rounded down and multiplied by the preset cumulative cleaning coefficient to obtain the cleaning number correction coefficient.

[0080] The preset standard cumulative cleaning count is a baseline number calibrated based on experimental data of the heat exchange core material (such as paper, aluminum foil) and cleaning method (such as water washing, air drying). It is used to convert the cumulative cleaning count into a quantifiable "wear cycle". It is usually set as the number of cleaning counts corresponding to a significant decrease in core performance (such as a 5% decrease in heat exchange efficiency) (such as 5 times preset). The preset cumulative cleaning coefficient is a proportional parameter determined in advance through a large number of cleaning-performance correlation experiments. It is used to convert the multiple of the cumulative cleaning times to the standard cleaning times into a quantified degree of performance degradation. The value is usually 0.03 (that is, the core performance decreases by 3% after each standard cleaning cycle). The cleaning frequency correction factor is a sub-parameter that ultimately quantifies the impact of the cumulative cleaning frequency on the heat exchange core. Its value ranges from 0.8 to 1.0 (the smaller the factor, the more severe the performance degradation of the core due to cumulative cleaning). In order to avoid the factor being too low and deviating from reality, a lower limit of 0.8 is set.

[0081] Convert S153 into a formula encapsulation form, specifically as follows:

[0082] In the formula, This represents the correction factor for the number of cleaning cycles. This indicates the cumulative number of times the heat exchange core has been cleaned. This indicates the cumulative number of cleaning cycles according to the preset standard. This indicates the preset cumulative cleaning coefficient.

[0083] In this embodiment of the invention, the control module of the target fresh air system first retrieves the latest cumulative cleaning count of the heat exchange core from the equipment operation log storage unit. The system automatically filters out cleaning records that do not meet the standards or are interrupted, and only counts the valid cleaning counts. Next, the control module retrieves the exclusive parameters that match the current heat exchange core model from the preset parameter library: the preset standard cumulative cleaning count and the preset cumulative cleaning coefficient, to ensure parameter compatibility. Subsequently, the cleaning count correction coefficient is calculated according to the formula.

[0084] S16. Using the heat exchange core coupling factor and the preset heat exchange core basic early warning threshold as inputs to the preset heat exchange core multi-factor coupling dynamic threshold model, the early warning value of the heat exchange core is obtained.

[0085] The preset basic warning threshold for the heat exchange core is the benchmark value calibrated under laboratory standard environment when the core reaches the critical risk state. For example, the critical humidity for mold growth in paper core is 65%RH, and the critical pressure difference for blockage in aluminum foil core is 50Pa. The pre-set heat exchange core multi-factor coupling dynamic threshold model is a unified model verified by a large number of core operating condition experiments. The core logic is to dynamically adjust the pre-set basic early warning threshold of the heat exchange core in layers through the heat exchange core coupling factor, so as to accurately adapt to the core risk status under the current operating condition.

[0086] The pre-set heat exchange core multi-factor coupled dynamic threshold model is converted into a formula encapsulation form, specifically:

[0087] In the formula, This indicates the warning value for the heat exchange core.

[0088] In this embodiment of the invention, the heat exchange core coupling factor and the preset heat exchange core basic early warning threshold are input into the preset heat exchange core multi-factor coupling dynamic threshold model for calculation to obtain the heat exchange core early warning value.

[0089] S17. Based on the early warning value of the heat exchange core, the state risk is divided to obtain the risk level range of the heat exchange core. The risk level range of the heat exchange core includes the low risk range, the medium risk range, and the high risk range of the heat exchange core.

[0090] The risk level range of the heat exchange core refers to the set of low-risk, medium-risk, and high-risk ranges of the heat exchange core after division. Each range corresponds to a specific range of monitoring data proportions and risk response strategies. The low-risk zone of the heat exchange core refers to the real-time humidity data of the heat exchange core. In Within this range, the core risk is extremely low; The risk zone in the heat exchange core refers to In The core within this range is of moderate risk and requires cleaning or maintenance to be carried out in the near future. The high-risk zone of the heat exchange core refers to In The core is in a high-risk zone and requires immediate cleaning or maintenance.

[0091] In this embodiment of the invention, the control module of the target fresh air system retrieves the heat exchange core warning value calculated in step S16 from the core warning parameter cache, and simultaneously retrieves the latest real-time humidity data of the core from the core monitoring unit; then, the control module calculates two key ratio thresholds according to preset risk classification rules: the first threshold is... The second threshold is They are divided into three risk zones for heat exchange cores: low risk zone, medium risk zone, and high risk zone.

[0092] S18. Compare real-time humidity data with the risk level range of the heat exchange core to determine the early warning level of the heat exchange core.

[0093] The heat exchange core warning level refers to the core risk identifier determined by comparing real-time humidity data with the risk level range of the heat exchange core. It corresponds one-to-one with the risk range, namely "low warning level of heat exchange core" (corresponding to low risk range), "medium warning level of heat exchange core" (corresponding to medium risk range), and "high warning level of heat exchange core" (corresponding to high risk range). Each level clarifies the priority of subsequent maintenance operations.

[0094] In this embodiment of the invention, determining whether the real-time humidity data is in the low-risk range, medium-risk range, or high-risk range of the heat exchange core is similar to the comparison process in S14 described above, and will not be repeated here.

[0095] Step 103: Use system component early warning data, equipment operation data and environmental data to match self-cleaning strategies and obtain system component self-cleaning strategies.

[0096] The self-cleaning strategy for system components refers to a self-cleaning operation plan for core system components (such as filters and heat exchange cores) generated through preset strategy matching rules based on system component early warning data (such as early warning levels of filters and heat exchange cores), equipment operation data (such as current air volume and cumulative running time of the fresh air system), and environmental data (such as indoor and outdoor temperature and humidity). Its content usually includes cleaning methods (such as water washing, air drying, and dust removal), cleaning duration, cleaning intensity, and triggering timing. Its core function is to achieve precise adaptation of component self-cleaning and avoid over-cleaning or under-cleaning.

[0097] Further, the system component self-cleaning strategy includes a filter screen self-cleaning strategy and a heat exchange core self-cleaning strategy. Step 103 includes the following sub-steps: S21. Generate a first target key by using the filter screen warning level and outdoor air quality index data.

[0098] The first target key refers to a unique identification string formed by splicing the character code corresponding to the filter screen warning level and the interval code corresponding to the outdoor air quality index data according to a preset fixed coding rule. Its core function is to establish an association index between the filter screen risk state and environmental influencing factors to ensure the accuracy of policy retrieval. In the embodiment of the present invention, first, the control module of the target fresh air system retrieves the determined filter screen warning level from the policy matching database, performs character coding according to a preset rule (the low warning level code of the filter screen is L, the medium warning level code is M, and the high warning level code is H), and at the same time retrieves the real-time outdoor air quality index data from the environmental data buffer area and completes the coding according to a preset interval (AQI≤100 is coded as A, 100<AQI≤200 is coded as B, AQI>200 is coded as C); then, the control module generates a first target key according to the fixed splicing rule of "filter screen warning level code - outdoor air quality index code", and the coding order cannot be reversed to ensure retrieval uniqueness. The generated first target key will be temporarily stored in the policy retrieval buffer area.

[0099] S22. Retrieve the preset filter screen self-cleaning strategy database by using the first target key and match the filter screen self-cleaning strategy.

[0100] The preset filter screen self-cleaning strategy database is a structured data set pre-stored in the fresh air system control module, and is internally stored in the form of a fixed key-value pair of "first target key - filter screen self-cleaning strategy", covering the cleaning methods, cleaning durations, cleaning intensities and auxiliary operation parameters corresponding to different risk levels and air quality conditions. All strategies have been verified and optimized through a large number of working condition experiments. The filter screen self-cleaning strategy is a specific cleaning execution plan customized for the filter screen. The core content includes cleaning methods (such as high-pressure water washing, negative pressure dust removal, ultrasonic cleaning, etc.), cleaning duration, cleaning frequency, and auxiliary rotation speed parameters of the fan during the cleaning process, etc.

[0101] In the embodiment of the present invention, the system control module calls the preset filter screen self-cleaning strategy database, starts the built-in retrieval engine, uses the first target key temporarily stored in step S21 as the retrieval keyword, and performs an accurate matching query in the database; the database will automatically locate the unique policy entry associated with this target key. If there is no matching result due to data anomalies during the retrieval process, the system will automatically call the preset default filter screen cleaning strategy.

[0102] S23. The second target key is generated by using the early warning level of the heat exchange core and real-time humidity data.

[0103] The second target key refers to a unique retrieval identifier string formed by concatenating the exclusive code corresponding to the warning level of the heat exchange core with the interval code corresponding to the real-time humidity data, following the encoding logic of the first target key. The core difference from the first target key lies in the different associated component status parameters and application scenarios, specifically adapted to the strategy retrieval needs of the heat exchange core.

[0104] In this embodiment of the invention, the system control module first retrieves the determined early warning level of the heat exchange core from the decision parameter area, and performs exclusive coding according to preset rules (low early warning level of heat exchange core is coded as Lc, medium early warning level as Mc, and high early warning level as Hc), and then retrieves real-time humidity data from the core monitoring unit. And complete the coding according to the preset humidity range ( Encoded as X, Encoded as Y, (Encoded as Z); then, a second target key is generated according to the fixed rule of "heat exchange core early warning level code - real-time humidity data code".

[0105] S24. Use the second target key to search the database of preset heat exchange core self-cleaning strategies and match the heat exchange core self-cleaning strategies.

[0106] The pre-set heat exchange core self-cleaning strategy database is a structured data collection with the same structure as the pre-set filter self-cleaning strategy database. It is stored internally with "second target key - heat exchange core self-cleaning strategy" as a fixed key-value pair. The strategy content focuses on the core risks of the heat exchange core and covers adaptability parameters such as drying intensity, dehumidification temperature, dust removal method and cleaning frequency. The heat exchange core self-cleaning strategy is a cleaning implementation plan specifically designed for the heat exchange core. Its core includes drying time, dehumidification auxiliary temperature, cleaning method selection, and operating parameters of suitable cleaning tools, taking into account both cleaning effect and core protection.

[0107] In this embodiment of the invention, the system control module calls the preset heat exchange core self-cleaning strategy database, uses the second target key temporarily stored in step S23 as the search keyword to start a precise search, and the database automatically locates and outputs the corresponding strategy entry; if the search result is empty (such as an abnormal code), the system will automatically enable the default cleaning strategy of the heat exchange core.

[0108] As shown in Table 1 below: Table 1. Target Keys and Matching Tables for System Component Self-Cleaning Strategies Target key type Generate core elements Strategy matching example First target key Filter warning level, outdoor air quality index data First target key "MB" → Negative pressure dust removal + 5-minute water washing + medium-speed fan-assisted drying Second target key Heat exchange core warning level, real-time humidity data Second target key "Mc-Y" → Low-temperature air drying for 10 minutes + light negative pressure dust removal + dehumidification assistance Step 104: Perform self-cleaning operation on the target fresh air system according to the self-cleaning strategy of the system components, and obtain the self-cleaning operation data of the target fresh air system during the self-cleaning operation.

[0109] Furthermore, the self-cleaning operating data includes the fan current fluctuation amplitude, fan speed adjustment delay data, and fan load factor.

[0110] Self-cleaning operating data refers to the set of real-time operating parameters generated by the fresh air system under the specific condition of performing self-cleaning operation. The core parameters include the fan current fluctuation amplitude, fan speed adjustment delay data, and fan load coefficient, which are used to reflect the operating stability and load status of the core components of the equipment during the self-cleaning process. The fan current fluctuation range refers to the difference between the maximum and minimum values ​​of the fan operating current during the self-cleaning operation. It reflects the current stability of the fan during operation. The larger the fluctuation range, the more unstable the fan load is usually or the potential problems such as component jamming exist. The fan speed regulation delay data refers to the time interval from when the system control module issues a fan speed regulation command to when the actual fan speed reaches the command set value and stabilizes. It reflects the response performance of the fan speed regulation. The shorter the delay time, the higher the fan control accuracy. The fan load factor is the ratio of the actual operating load of the fan to the rated load under self-cleaning conditions. It is used to quantify the current load level of the fan. Its value range is usually 0-1.2. When the ratio exceeds 1, it means that the fan is operating under overload and there is a risk of overload.

[0111] In this embodiment of the invention, the control module of the target fresh air system first retrieves the complete self-cleaning strategy for system components from the strategy cache, parses the cleaning sequence and cleaning parameters of each component (such as cleaning duration, cleaning intensity, fan auxiliary speed, etc.) in the strategy, and then sends a start command to the system's execution unit to drive each component to perform self-cleaning operations in a preset order. First, according to the filter self-cleaning strategy, the filter cleaning execution mechanism, such as the high-pressure water washing pump and the negative pressure dust removal device, is started, and the fan is adjusted to the corresponding auxiliary speed. After the filter cleaning is completed, the drying device and the light dust removal component are started according to the heat exchange core self-cleaning strategy to perform core cleaning operations. Throughout the entire self-cleaning operation, the control module synchronously triggers the data acquisition unit to collect three types of self-cleaning operating data in real time through preset sensors and monitoring components: through A high-precision current sensor connected in series in the fan power supply circuit collects the real-time operating current of the fan, continuously records the current data, and calculates the fan current fluctuation amplitude. The actual fan speed is fed back in real time through the speed encoder at the fan shaft end. At the same time, the timestamp of the speed adjustment command issued by the control module and the timestamp of the speed stabilization are recorded. The difference between the two is the fan speed adjustment delay data. Through the load calculation unit built into the control module, the actual operating power, torque and other parameters of the fan are read in real time. Combined with the preset rated load parameters of the fan, the fan load coefficient is calculated according to the formula: fan load coefficient = actual load / rated load. The three types of data collected are filtered by the system's built-in filtering algorithm to remove instantaneous interference noise to ensure data accuracy. Then, they are associated with the time node of the self-cleaning operation and the corresponding cleaning component and stored in the system's operating condition database.

[0112] Step 105: Based on the pre-set fan health assessment model, conduct a fan health assessment using equipment operation data, environmental data, and self-cleaning operation data, and generate a health assessment report for the target fresh air system.

[0113] Furthermore, step 105 may include the following sub-steps: S31. Data preprocessing is performed using equipment operation data, environmental data, and self-cleaning operation data to obtain a time-series feature dataset.

[0114] Data preprocessing refers to a set of standardized operations performed on equipment operation data, environmental data, and self-cleaning operation data. Its core purpose is to remove data noise, correct outliers, standardize data formats, and extract time-related features to ensure that the data meets the input requirements of subsequent analysis models. A time-series feature dataset is a structured dataset that integrates the key features of three types of original data after preprocessing and indexing by timestamps. The data records are arranged in chronological order, and each time node corresponds to a complete set of feature parameters, which can intuitively reflect the changing patterns of each data over time. In this embodiment of the invention, the control module of the target fresh air system first retrieves the corresponding three types of raw data from the equipment operation log database, the environmental data cache, and the self-cleaning condition database, respectively. Then, a preset data preprocessing program is started to align the three types of data according to a unified timestamp to ensure that the equipment status, environmental conditions, and self-cleaning operation parameters at the same time point correspond one-to-one, thus solving the problem of asynchronous data acquisition time. Next, a preset outlier detection algorithm is used to identify and process invalid values, missing values, and extreme outliers in the three types of data. Missing data is supplemented using linear interpolation, and extreme outliers are corrected using truncation to avoid abnormal data affecting subsequent analysis. Then, normalization processing is used to uniformly map parameters of different dimensions to the [0,1] interval to eliminate feature weight bias caused by differences in dimensions. Finally, the processed effective data is sorted according to time sequence, the core feature parameters in each group of data are extracted and associated with the corresponding timestamp, and integrated to form a time-series feature dataset.

[0115] S32. Extract features from the time-series feature dataset to obtain the wind turbine health status features and wind turbine dynamic load capacity features.

[0116] Feature extraction refers to the process of using a preset feature extraction algorithm to filter, calculate and extract key information that can characterize the core operating status of a wind turbine from a time-series feature dataset, eliminate redundant data, and retain feature parameters with decision-making value. The health status characteristics of a wind turbine refer to a set of features that reflect the current operational stability and potential failure risks of the wind turbine. The core features include the stability index of wind turbine current fluctuation, the average speed regulation delay, and the frequency of abnormal fluctuations. The closer the feature values ​​are to the preset standard values, the better the health status of the wind turbine. The dynamic load capacity characteristics of a wind turbine refer to a set of features that characterize the wind turbine's adaptability and adjustment potential in response to load changes under self-cleaning conditions. These features mainly include the wind turbine load coefficient change rate, the duration of the load peak, and the rated load adaptability, which are used to determine whether the wind turbine can stably support self-cleaning operations of different intensities.

[0117] In this embodiment of the invention, the control module of the target fresh air system first retrieves the time-series feature dataset generated in step S31 from the feature data storage area. For fan-related parameters, it first filters out the parameter sequences directly related to the fan from the time-series dataset, including time-series sequences such as fan current fluctuation amplitude, fan speed adjustment delay data, and fan load coefficient. Then, based on this sequence, the fan health status characteristics are calculated. Through statistical analysis, the current fluctuation stability index (taking the standard deviation of the current fluctuation amplitude time-series data), the mean of speed adjustment delay (taking the time average of the speed adjustment delay data), and the frequency of abnormal fluctuations (statistically counting the number of times the current fluctuation amplitude exceeds the preset threshold) are obtained. At the same time, the dynamic load capacity characteristics of the fan are calculated. The fan load coefficient change rate is obtained through time-series difference. The duration of the load coefficient exceeding the rated value is used as the duration of the load peak. The average fit between the load coefficient time-series data and the rated load is calculated as the rated load fit. Subsequently, the two types of features are integrated, and duplicate and redundant feature parameters are removed to form structured fan health status characteristics and fan dynamic load capacity characteristics. Finally, the time period information associated with these two types of features is stored in the fan feature database.

[0118] S33. Perform feature fusion on the health status characteristics and dynamic load capacity characteristics of the wind turbine to obtain the fused feature matrix.

[0119] Feature fusion refers to integrating the health status features of wind turbines, which characterize different operating dimensions of wind turbines, with the dynamic load capacity features of wind turbines. By unifying feature dimensions and allocating feature weights, a feature set with more comprehensive characterization capabilities is formed. The core purpose is to avoid the limitations of a single feature and improve the accuracy of subsequent analysis. The fusion feature matrix refers to a two-dimensional structured data matrix formed after feature fusion. Its row dimension corresponds to the time nodes in the time series feature dataset, and its column dimension corresponds to various core feature parameters after fusion. Each element in the matrix is ​​the quantified value of the corresponding feature at a specific time node, which facilitates subsequent wind turbine status analysis through matrix operations.

[0120] In this embodiment of the invention, the control module of the target fresh air system first retrieves the fan health status features and fan dynamic load capacity features obtained in step S32 from the fan feature database. Then, a preset feature fusion algorithm is started to first unify the dimensions of the two types of features, mapping the quantized values ​​of different features to the same data range and eliminating the numerical scale differences between features. Then, according to the importance of the two types of features to the fan status assessment, preset feature weights are assigned (e.g., the weight of the fan health status feature is 0.5, and the weight of the fan dynamic load capacity feature is 0.5, the weights are determined by the working condition test). Next, the weighted two types of feature parameters are integrated one by one according to the time node, and a two-dimensional fusion feature matrix is ​​constructed with the time node as the row index and the fused feature parameters (current fluctuation stability index, load coefficient change rate, etc.) as the column index. Finally, the fusion feature matrix is ​​associated with the corresponding time period information and stored in the system's fusion feature database.

[0121] S34. Use the fusion feature matrix input to the pre-set fan health assessment model to conduct fan health assessment and generate a health assessment report for the target fresh air system.

[0122] The pre-set fan health assessment model is an improved Bi-LSTM-Attention model. It incorporates an operating condition-adaptive Attention mechanism into the traditional Bi-LSTM model, which solves the problem that a single time series model is difficult to capture bidirectional time series dependencies and key operating condition features. At the same time, it realizes the gradual processing of features and multi-task decision-making through a hierarchical architecture, and finally generates a comprehensive target fresh air system health assessment report.

[0123] The structure and working principle of the improved Bi-LSTM-Attention model: The improved Bi-LSTM-Attention model is a deep learning model that is trained on a large amount of wind turbine operating data and then embedded into the fresh air system control module. Its core consists of an input adaptation layer, a two-branch temporal extraction layer (a dual-parallel improved Bi-LSTM sub-network), an operating condition attention fusion layer (fusion attention mechanism), and a multi-task decision layer (a fully connected sub-module cluster) connected in series. The overall working principle is "input standardization adaptation → two-branch bidirectional temporal feature mining → operating condition-oriented attention weighted fusion → multi-task parallel decision output".

[0124] Among them, the improved Bi-LSTM sub-network adds dropout layer and layer normalization operation on the basis of traditional Bi-LSTM to avoid model overfitting, while strengthening the capture of forward and backward dependencies of time series data; the working condition attention fusion layer binds the attention mechanism with the real-time working condition of the fresh air system, dynamically allocates feature weights, and allows the model to focus on key health information under the current working condition; the multi-task decision layer completes three types of evaluation tasks simultaneously through parallel sub-modules, improving evaluation efficiency and comprehensiveness.

[0125] Furthermore, the pre-built wind turbine health assessment model includes an input adaptation layer, a dual-branch time series extraction layer, an operating condition attention fusion layer, and a multi-task decision layer. S34 may include the following sub-steps: S341. The fusion feature matrix is ​​adjusted in dimension and format by the input preprocessing layer to obtain the standard input feature sequence.

[0126] The input adaptation layer is the first layer structure of the improved Bi-LSTM-Attention model, which has the functions of dimension calibration, data normalization and format conversion. It completes data preprocessing adaptation to meet the fixed requirements of Bi-LSTM sub-network on input sequence length and feature dimension. The standard input feature sequence is a fixed-length one-dimensional time sequence obtained after processing by the input adaptation layer. Its sequence length and feature dimension conform to the input specifications of the two-branch Bi-LSTM sub-network, and the numerical range is uniformly mapped to the interval [-1,1]. It can be directly passed into the next layer of the model for feature extraction.

[0127] In this embodiment of the invention, the system control module retrieves the fusion feature matrix from the fusion feature database and inputs it into the input adaptation layer of the improved Bi-LSTM-Attention model. This layer first converts the two-dimensional fusion feature matrix into a one-dimensional temporal structure of [sequence length, feature dimension] through a matrix reshaping operation. The sequence length is truncated or padded according to a fixed value preset during model training (such as 100 time nodes), and the feature dimension is consistent with the number of columns of the fusion feature matrix. Then, the layer normalization algorithm is used to eliminate data distribution differences and avoid the interference of numerical fluctuations on the model training parameters. Finally, the built-in activation function maps the sequence values ​​to the [-1,1] interval, and after standardization, the standard input feature sequence is output and directly fed into the dual-branch temporal extraction layer.

[0128] S342. Temporal features are extracted from the standard input feature sequence through a dual-branch temporal extraction layer to obtain conventional health temporal features and dynamic load temporal features.

[0129] It is worth mentioning that the dual-branch temporal extraction layer performs bi-dimensional temporal feature mining on the standard input feature sequence. This layer consists of two parallel improved Bi-LSTM sub-networks, which focus on capturing the temporal dependencies of the wind turbine's normal health status and dynamic load capacity, respectively, to achieve accurate separation and deep extraction of the two types of core features.

[0130] The dual-branch temporal extraction layer is the core feature extraction layer of the improved Bi-LSTM-Attention model. It includes a regular healthy Bi-LSTM sub-network and a dynamic load Bi-LSTM sub-network. The two sub-networks have the same structure (both contain an input layer, a forward LSTM layer, a backward LSTM layer, a dropout layer, and an output layer), but the weight parameters obtained during training are independent and are adapted to the extraction needs of different dimensional features. The routine health time series features are the feature set output by the routine health Bi-LSTM sub-network. By capturing the forward and backward time series dependencies of parameters such as current fluctuation stability and speed regulation delay mean, they reflect the state pattern of long-term stable operation of the wind turbine. Dynamic load time series characteristics are the feature set output by the dynamic load Bi-LSTM sub-network, focusing on the time series evolution of parameters such as load coefficient change rate and load peak duration, characterizing the dynamic response characteristics of the wind turbine under load change conditions. The improved Bi-LSTM subnetwork is a subnetwork that adds a dropout layer (dropout rate set to 0.2) and layer normalization operation to the traditional Bi-LSTM, which can effectively prevent model overfitting and improve the generalization ability of feature extraction.

[0131] In this embodiment of the invention, the standard input feature sequence is synchronously fed into two parallel improved Bi-LSTM sub-networks of the dual-branch temporal extraction layer. The conventional health Bi-LSTM sub-network captures the future dependencies of the temporal data through the forward LSTM layer and the past dependencies through the backward LSTM layer. After the dropout layer filters out redundant information and the layer normalization process, it outputs conventional health temporal features with uniform dimensions. The dynamic load Bi-LSTM sub-network uses the same operation logic to mine the bidirectional temporal correlation of load-related parameters and outputs dynamic load temporal features. The two types of features are synchronously fed into the working condition attention fusion layer.

[0132] S343. Weighted fusion health features are obtained by weighting and fusing the routine health time-series features and dynamic load time-series features through the working condition attention fusion layer.

[0133] It is worth mentioning that the working condition attention fusion layer combines the attention mechanism with the real-time working condition of the fresh air system, and performs dynamic weight allocation and deep fusion of conventional health time-series features and dynamic load time-series features, which enhances the feature representation capability of key time-series nodes under the current working condition and solves the problem of ignoring the differences in working conditions by a single fusion method.

[0134] The working condition attention fusion layer is the core feature enhancement layer of the improved Bi-LSTM-Attention model. It has a built-in attention weight calculation module and feature fusion module. It dynamically calculates the attention weights of two types of temporal features based on real-time working condition parameters, and then achieves feature fusion through weighted summation. The weighted fusion health feature is a comprehensive feature vector output by the operating condition attention fusion layer. It combines the stability information of conventional health time series features with the adaptive information of dynamic load time series features, and highlights the time series features that play a key role in the health assessment of wind turbines under the current operating conditions. The attention weight calculation module is the core component of the working condition attention fusion layer. It calculates the importance weight of each time-series node feature by using real-time working condition parameters (such as self-cleaning working condition indicators and outdoor air quality index) as attention cues.

[0135] In this embodiment of the invention, the operating condition attention fusion layer first retrieves the current operating condition parameters of the target fresh air system and inputs them into the attention weight calculation module. The global attention weights of the normal health time-series features and the dynamic load time-series features are calculated using the softmax function. For example, under the self-cleaning condition, the weight of the dynamic load time-series feature is set to 0.6, and the weight of the normal health time-series feature is set to 0.4. Under the daily operating condition, the weights of both are set to 0.5. At the same time, local attention weights are assigned to each time-series node to highlight the features of abnormal time-series nodes. Then, through the feature fusion module, the two types of time-series features and their corresponding attention weights are weighted and summed. The feature dimension is then compressed and deeply integrated by the fully connected layer. Finally, a weighted fusion health feature with simplified dimensions and prominent key information is output and fed into the multi-task decision layer.

[0136] S344. Through the multi-task decision layer, the weighted fusion health features are classified for faults, the remaining lifetime is calculated and the dynamic performance is scored, so as to obtain the fault type probability, the remaining operating lifetime and the dynamic load score.

[0137] It should be noted that the multi-task decision layer performs parallel multi-dimensional evaluation of weighted and integrated health characteristics. This layer contains three independent fully connected sub-modules that simultaneously complete fault classification, remaining lifetime calculation, and dynamic performance scoring, achieving a comprehensive assessment of the wind turbine's health status.

[0138] The multi-task decision layer is the output decision layer of the improved Bi-LSTM-Attention model. It consists of a fault classification submodule, a remaining lifetime calculation submodule, and a dynamic load scoring submodule. The three submodules share the weighted fusion health feature input and independently complete the calculation output, thereby improving the evaluation efficiency. The fault type probability is the probability value of the occurrence of various potential faults of the fan (such as bearing wear, motor aging, impeller dirt accumulation, etc.) output by the fault classification submodule through the softmax activation function. The value range is [0,1]. The closer the probability value is to 1, the higher the risk of the corresponding fault. The remaining operating life is the stable operating time of the wind turbine (unit: h) predicted by the remaining operating life calculation submodule based on the time-series characteristic trend prediction, and is obtained by extrapolating the evolution law of the characteristic sequence through a linear regression algorithm. The dynamic load score is a score of the wind turbine's load adaptability output by the dynamic load score submodule according to preset rules. The score ranges from 0 to 100. The higher the score, the stronger the wind turbine's ability to cope with dynamic load changes.

[0139] In this embodiment of the invention, after the weighted and fused health features are input into the multi-task decision layer, they enter three parallel fully connected sub-modules: the fault classification sub-module extracts features through a three-layer fully connected network and outputs the probability of occurrence of various faults through a softmax function; the remaining life calculation sub-module extrapolates the remaining operating life of the wind turbine by combining the slope of the feature sequence change with historical life data through a time-series prediction branch; and the dynamic load scoring sub-module maps the feature quantization value to a scoring range of 0-100 points and outputs the dynamic load score. The three sub-modules complete the calculation simultaneously and output three types of core evaluation results.

[0140] S345. Integrate fault type probability, remaining service life and dynamic load score to generate a health assessment report for the target fresh air system.

[0141] The health assessment report is a standardized document that integrates three types of core assessment results, wind turbine health level determination, and targeted operation and maintenance recommendations. It includes four main modules: basic assessment information, core assessment results, health level, and operation and maintenance guidelines. The health level of a wind turbine is determined based on a combination of dynamic load score and failure type probability, and is divided into three levels: excellent (dynamic load score ≥ 80 points, no high probability failures), good (60 points ≤ dynamic load score < 80 points, low probability failures), and requires maintenance (dynamic load score < 60 points, medium to high probability failures).

[0142] In this embodiment of the invention, the system control module collects the three types of results output in step S344 and integrates them in a structured manner according to a preset template: First, it marks the basic information such as the evaluation time and the wind turbine model; then, it lists the top three potential faults with the highest probability of fault type and their corresponding probabilities, and clarifies the remaining operating life and dynamic load score; next, it determines the current health level of the wind turbine according to preset standards; finally, it generates targeted operation and maintenance suggestions based on the evaluation results, such as recommending to check and replace the bearing within 15 days for high-probability bearing wear faults, and recommending to prepare spare wind turbines in advance when the remaining life is less than 500 hours. The above content is integrated to generate a standardized health assessment report, which is stored in the system assessment report database and can also be pushed to the operation and maintenance management terminal to provide operation and maintenance personnel with intuitive and practical operation guidance.

[0143] It is worth mentioning that the pre-built wind turbine health assessment model in this implementation can adopt a machine learning algorithm model, such as a conventional long short-term memory (LSTM) + deep neural network (DNN) wind turbine health assessment model. First, the fused feature matrix is ​​standardized and then input into a standard LSTM layer to extract time-series correlation features. Then, the features are fed into a fully connected DNN layer, and through three independent output branches (fault classification, lifetime prediction, and performance score), the fault type probability, remaining operating lifetime, and dynamic load score are output respectively. Finally, the results are integrated to generate a health assessment report. Example

[0144] Please see Figure 2 The present invention provides a fresh air system treatment system based on intelligent maintenance early warning and self-cleaning, comprising: The response module 201 is used to respond to the operation monitoring command of the target fresh air system and to obtain the equipment operation data and environmental data of the target fresh air system. Assessment module 202 is used to conduct component status risk assessment using equipment operation data and environmental data to obtain system component early warning data; Matching module 203 is used to match self-cleaning strategies using system component early warning data, equipment operation data and environmental data to obtain system component self-cleaning strategies; The execution module 204 is used to perform self-cleaning operation on the target fresh air system according to the self-cleaning strategy of the system components, and to obtain the self-cleaning operation data of the target fresh air system during the self-cleaning operation process. Report module 205 is used to conduct a health assessment of the target fresh air system based on a pre-set fan health assessment model, using equipment operation data, environmental data, and self-cleaning operation data.

[0145] Since the above is a system corresponding to the intelligent maintenance early warning and self-cleaning fresh air system treatment method, its implementation principle is the same as that of the intelligent maintenance early warning and self-cleaning fresh air system treatment method. For the sake of convenience and brevity, those skilled in the art can clearly understand that the specific working process of the system and modules described above can be referred to the corresponding process in the aforementioned method embodiments, and will not be repeated here.

[0146] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A fresh air system treatment method based on intelligent maintenance early warning and self-cleaning, characterized in that, include: In response to the operation monitoring command of the target fresh air system, acquire the equipment operation data and environmental data of the target fresh air system; The device operation data and the environmental data are used to conduct a component status risk assessment to obtain system component early warning data. The self-cleaning strategy of the system components is obtained by matching the system component early warning data, the equipment operation data and the environmental data. The system performs a self-cleaning operation on the target fresh air system according to the self-cleaning strategy of the system components, and obtains the self-cleaning operation data of the target fresh air system during the self-cleaning operation process. Based on the pre-set fan health assessment model, the fan health is assessed using the equipment operation data, the environmental data, and the self-cleaning operating data, and a health assessment report for the target fresh air system is generated.

2. The fresh air system treatment method based on intelligent maintenance early warning and self-cleaning according to claim 1, characterized in that, The equipment operation data includes real-time differential pressure data of the filter in the target fresh air system, real-time humidity data of the heat exchange core, cumulative operating time of the filter and cumulative cleaning times of the heat exchange core. The environmental data includes outdoor air quality index data, indoor and outdoor temperature difference data and current seasonal parameters.

3. The fresh air system treatment method based on intelligent maintenance early warning and self-cleaning according to claim 2, characterized in that, The process of using the equipment operation data and the environmental data to perform component status risk assessment and obtain system component early warning data includes: The filter coupling factor is determined based on the outdoor air quality index data, the cumulative operating time of the filter, and the real-time differential pressure data. The filter coupling factor and the preset filter basic early warning threshold are used as inputs to the preset filter multi-factor coupling dynamic threshold model to obtain the filter early warning value; Based on the filter warning value, the status risk is divided to obtain the filter risk level range, which includes a low-risk range, a medium-risk range, and a high-risk range. By comparing the real-time differential pressure data with the filter risk level range, the filter warning level is determined; The heat exchange core coupling factor is determined based on the indoor and outdoor temperature difference data, the current seasonal parameters, and the cumulative number of cleanings of the heat exchange core. The heat exchange core coupling factor and the preset heat exchange core basic early warning threshold are used as inputs to the preset heat exchange core multi-factor coupling dynamic threshold model to obtain the heat exchange core early warning value; Based on the warning value of the heat exchange core, the state risk is divided to obtain the risk level range of the heat exchange core. The risk level range of the heat exchange core includes a low risk range, a medium risk range, and a high risk range. By comparing the real-time humidity data with the risk level range of the heat exchange core, the early warning level of the heat exchange core is determined. The system component warning data includes the filter warning level and the heat exchange core warning level.

4. The fresh air system treatment method based on intelligent maintenance early warning and self-cleaning according to claim 3, characterized in that, The step of determining the filter coupling factor based on the outdoor air quality index data, the cumulative operating time of the filter, and the real-time differential pressure data includes: The quality index weighting factor is determined based on the preset quality index weighting factor range to which the outdoor air quality index data belongs; The ratio of the cumulative running time of the filter to the preset standard cumulative running time is calculated, and the ratio result is rounded down and multiplied by the preset running time coefficient to obtain the running time attenuation coefficient. Based on the preset cleaning evaluation range to which the real-time differential pressure data belongs, determine the cleaning effect correction coefficient; The filter coupling factor includes the quality index weighting factor, the runtime attenuation coefficient, and the cleaning effect correction coefficient.

5. The fresh air system treatment method based on intelligent maintenance early warning and self-cleaning according to claim 3, characterized in that, The step of determining the heat exchange core coupling factor based on the indoor-outdoor temperature difference data, the current seasonal parameters, and the cumulative number of cleaning cycles of the heat exchange core includes: The ratio of the indoor and outdoor temperature difference data to the preset standard temperature difference data is calculated, and the ratio result is rounded down and multiplied by the preset temperature difference coefficient to obtain the temperature difference factor. The seasonal factor to which the current seasonal parameter belongs is used as the seasonal correction coefficient; The ratio of the cumulative number of cleanings of the heat exchange core to the preset standard cumulative number of cleanings is calculated, and the ratio result is rounded down and multiplied by the preset cumulative cleaning coefficient to obtain the cleaning number correction coefficient. The heat exchange core coupling factor includes the temperature difference factor, the seasonal correction factor, and the cleaning frequency correction factor.

6. The fresh air system treatment method based on intelligent maintenance early warning and self-cleaning according to claim 3, characterized in that, The self-cleaning strategy is obtained by matching the system component early warning data, the equipment operation data, and the environmental data, including: A first target key is generated using the filter warning level and the outdoor air quality index data; The first target key is used to retrieve the preset filter self-cleaning strategy database and match the filter self-cleaning strategy. A second target key is generated using the aforementioned heat exchange core warning level and the real-time humidity data; The second target key is used to retrieve the preset heat exchange core self-cleaning strategy database and match the heat exchange core self-cleaning strategy; The self-cleaning strategy for system components includes the filter self-cleaning strategy and the heat exchange core self-cleaning strategy.

7. The fresh air system treatment method based on intelligent maintenance early warning and self-cleaning according to claim 1, characterized in that, The pre-set fan health assessment model uses the equipment operation data, environmental data, and self-cleaning operating data to perform a fan health assessment, generating a health assessment report for the target fresh air system, including: The equipment operation data, the environmental data, and the self-cleaning operation data are used for data preprocessing to obtain a time-series feature dataset. Feature extraction is performed on the time-series feature dataset to obtain wind turbine health status features and wind turbine dynamic load capacity features; The wind turbine health status characteristics and the wind turbine dynamic load capacity characteristics are fused to obtain a fused feature matrix; The fusion feature matrix is ​​used as input to the pre-set fan health assessment model to perform fan health assessment and generate a health assessment report for the target fresh air system.

8. The fresh air system treatment method based on intelligent maintenance early warning and self-cleaning according to claim 7, characterized in that, The pre-built fan health assessment model includes an input adaptation layer, a dual-branch temporal extraction layer, a working condition attention fusion layer, and a multi-task decision layer. The fan health assessment is performed by inputting the fused feature matrix into the pre-built fan health assessment model, generating a health assessment report for the target fresh air system, including: The fused feature matrix is ​​adjusted in dimension and format by the input preprocessing layer to obtain a standard input feature sequence. The dual-branch time-series extraction layer extracts time-series features from the standard input feature sequence to obtain conventional health time-series features and dynamic load time-series features. The working condition attention fusion layer performs weight allocation and feature fusion on the conventional health time-series features and the dynamic load time-series features to obtain weighted fused health features. The multi-task decision layer performs fault classification, remaining lifetime calculation and dynamic performance scoring on the weighted fusion health features to obtain fault type probability, remaining lifetime and dynamic load score. By integrating the failure type probability, the remaining operating life, and the dynamic load score, a health assessment report for the target fresh air system is generated.

9. The fresh air system treatment method based on intelligent maintenance early warning and self-cleaning according to any one of claims 1-8, characterized in that, The self-cleaning operating data includes the fan current fluctuation amplitude, fan speed adjustment delay data, and fan load coefficient.

10. A fresh air system treatment system based on intelligent maintenance early warning and self-cleaning, characterized in that, include: The response module is used to respond to the operation monitoring command of the target fresh air system and to acquire the equipment operation data and environmental data of the target fresh air system. The assessment module is used to perform component status risk assessment using the equipment operation data and the environmental data, and obtain system component early warning data; The matching module is used to match the self-cleaning strategy using the system component early warning data, the equipment operation data and the environmental data to obtain the system component self-cleaning strategy; The execution module is used to perform a self-cleaning operation on the target fresh air system according to the self-cleaning strategy of the system components, and to acquire the self-cleaning operation data of the target fresh air system during the self-cleaning operation process. The reporting module is used to perform a health assessment of the fan based on a pre-set fan health assessment model, using the equipment operation data, the environmental data, and the self-cleaning operating data, and generate a health assessment report for the target fresh air system.

Images