A high-efficiency filtration-based domestic ffu air purification system

Abstract

This invention discloses a home FFU air purification system based on high-efficiency filtration, belonging to the field of air purification technology. It includes: an air characteristic sensing module that collects multi-dimensional indoor air pollutant parameters; a filtration strategy generation module that constructs a pollutant concentration-filtration efficiency mapping model and generates a dynamic filtration layer combination strategy; a high-efficiency purification execution module that, according to the dynamic filtration layer combination strategy, drives the FFU fan variable frequency speed control unit and the multi-stage filter switching mechanism to work together to complete air purification and output purified air parameters; and a dynamic feedback adjustment module that calculates the deviation between the purified air parameters and the preset target value, and feeds back to correct the mapping model parameters of the filtration strategy generation module. This invention achieves an intelligent upgrade of the home FFU air purification system from passive response to proactive, precise, efficient, and energy-optimized operation through pollutant trend prediction, scenario-based weight adaptation, and a self-learning optimization model.

Description

Technical Field

[0001] This invention relates to the field of air purification technology, specifically to a household FFU air purification system based on high-efficiency filtration. Background Technology

[0002] With the acceleration of urbanization and the popularization of indoor decoration, air pollutants such as PM2.5, VOCs, and formaldehyde have an increasingly significant impact on human health. Household air purification equipment has become a key means to improve indoor air quality. At present, household air purifiers on the market are mainly divided into filter type, electrostatic type, and photocatalytic type. Among them, filter type purifiers are widely used because of their stable purification effect. Indoor air quality directly affects human health. Home FFU air purification systems are widely used due to their high-efficiency filtration characteristics. However, most existing systems adopt fixed filtration modes and single pollutant detection methods, lacking comprehensive perception of multi-dimensional pollutants, dynamic adaptation filtration strategies, and closed-loop feedback adjustment mechanisms. This leads to an imbalance between purification efficiency and energy consumption, and makes it difficult to adapt to different usage scenarios.

[0003] Achieving integrated purification that combines precise perception, dynamic adjustment, and scene adaptation is a technical challenge that needs to be overcome in this field. Summary of the Invention

[0004] The purpose of this invention is to solve the problems mentioned above by proposing a household FFU air purification system based on high-efficiency filtration.

[0005] The objective of this invention can be achieved through the following technical solutions: A home FFU air purification system based on high-efficiency filtration includes: an air characteristic sensing module, a filtration strategy generation module, a high-efficiency purification execution module, and a dynamic feedback adjustment module; The air feature sensing module is used to collect multi-dimensional air pollutant parameters indoors and normalize the multi-dimensional air pollutant parameters to generate a real-time air quality feature vector. The filtration strategy generation module is electrically connected to the air feature sensing module. The filtration strategy generation module is used to construct a pollutant concentration filtration efficiency mapping model based on the real-time air quality feature vector, combined with the rated parameters of the FFU fan and the filter performance curve, and generate a dynamic filtration level combination strategy. The high-efficiency purification execution module is electrically connected to the filtration strategy generation module. The high-efficiency purification execution module is used to drive the FFU fan frequency conversion speed regulation unit and the multi-stage filter switching mechanism to work together according to the dynamic filtration level combination strategy to complete the air purification process and output the purified air parameters. The dynamic feedback adjustment module is electrically connected to the air feature sensing module and the high-efficiency purification execution module, respectively. The dynamic feedback adjustment module is used to calculate the deviation between the purified air parameters and the preset target values, and to feed back and correct the mapping model parameters of the filtration strategy generation module to perform adaptive high-efficiency purification closed-loop control.

[0006] Preferably, the air feature sensing module includes: The multi-source sensing unit collects PM2.5 concentration, VOCs concentration, formaldehyde concentration, and ambient temperature and humidity by deploying laser dust sensors, gas sensor arrays, and temperature and humidity sensors in different indoor areas, respectively, and obtains multi-dimensional raw sensing data, which is then transmitted to the data preprocessing unit. The data preprocessing unit performs preprocessing on the received multi-dimensional raw sensing data, including: using an outlier removal algorithm to remove abnormal data caused by instantaneous sensor interference; using a filtering algorithm to eliminate high-frequency noise to smooth the data; performing range calibration before data normalization to unify the data benchmark and obtain stable sensing data; and sending the stable sensing data to the feature vector generation unit. The feature vector generation unit receives stable sensing data, performs normalization transformation according to preset pollutant and environmental parameter weight coefficients, maps each parameter to a unified numerical range, and generates a real-time air quality feature vector containing pollutant concentration information and environmental parameter information; and transmits the feature vector to the filtration strategy generation module.

[0007] Preferably, the air feature sensing module further includes: The time series analysis unit receives the real-time air quality feature vector output by the feature vector generation unit, constructs a data sequence of continuous time periods according to the timestamp, calculates the pollutant concentration change rate and cumulative pollution trend characteristics, and obtains trend feature parameters; the trend feature parameters are then transmitted to the pollution prediction unit. The formula for calculating the pollutant concentration change rate is as follows:

[0008] In the formula, The rate of change of pollutant concentration. For the current moment Measured values ​​of pollutant concentration The measured value of pollutant concentration at time _____. The time interval used for calculation; The formula for calculating the cumulative pollution trend characteristics is as follows:

[0009] In the formula, The cumulative pollution index, To calculate the cumulative number of time windows, These are the measured values ​​of pollutant concentrations at the corresponding times. The pollution prediction unit takes the trend feature parameters output by the time series analysis unit as input, performs prediction calculations through a trained time series prediction neural network, outputs a sequence of predicted pollutant concentration values ​​for a set time period in the future, and synchronizes the prediction sequence to the scene adaptation unit of the filtration strategy generation module. The prediction formula is as follows:

[0010] In the formula, For the future Predicted pollutant concentration at time ( (For predicting step size, e.g., 10 min / step). For a moment The trend characteristic parameter sequence, for The weight parameter matrix of a neural network includes the weights of the input gate, forget gate, cell state, and output gate. for The bias vector of a neural network. for Network forward computation function.

[0011] Preferably, the filtering strategy generation module includes: The basic parameter library unit pre-stores three types of core parameters, including: the rated operating parameters and power characteristics of the FFU fan, the filtration efficiency, resistance characteristics and dust holding capacity parameters of the filter. These parameters are provided as basic data to the efficiency mapping modeling unit. The efficiency mapping modeling unit receives the real-time air quality feature vector output by the air feature sensing module, combines it with the equipment parameters of the basic parameter library unit, constructs a mapping model between pollutant concentration and filtration efficiency through a neural network, calculates the purification efficiency under different fan operating states and filter combinations, and generates a mapping relationship table of fan operating parameters, filter combinations and purification efficiency, which is then transmitted to the strategy optimization unit. The strategy optimization unit uses the mapping table as the basic data and combines it with preset energy consumption constraints to solve the optimal operation scheme using an optimization algorithm, thereby obtaining a dynamic filtration level combination strategy that includes the target operating parameters of the fan and the activation level of the filter; this strategy is then sent to the high-efficiency purification execution module.

[0012] Preferably, the filtering strategy generation module further includes: The scene adaptation unit receives the pollutant concentration prediction value sequence output by the pollution prediction unit, combines it with the scene mode set by the user, including sleep, daily activities and cooking scenes, calls the pre-stored operation constraint parameters and purification priority parameters corresponding to each scene, and passes the scene parameters to the dynamic weight adjustment unit. The dynamic weight adjustment unit adjusts the pollutant weight coefficients in the efficiency mapping model based on the scenario parameters of the scenario adaptation unit, prioritizing the purification efficiency of high-impact pollutants within the scenario, recalculating the mapping relationship table and updating it to the strategy optimization unit, thereby achieving scenario-based filtration strategy optimization.

[0013] Preferably, the high-efficiency purification execution module includes: The variable frequency drive unit receives the target speed of the fan output by the strategy optimization unit, converts the speed command into a motor drive signal through the variable frequency control module, drives the fan motor to run at the target speed, and collects the actual speed of the fan in real time and feeds it back to the cooperative control unit. The multi-level switching unit receives filter level instructions from the dynamic filter level combination strategy, controls valve actions to switch the airflow path of different filter levels, and sends a status signal to the cooperative control unit after the switching is completed. The purification output monitoring unit deploys a group of miniature sensors at the FFU outlet to collect the pollutant concentration in the purified air, calculates the purification efficiency based on the pollutant concentrations before and after purification, and transmits the purified parameters and purification efficiency values ​​to the dynamic feedback adjustment module.

[0014] Preferably, the high-efficiency purification execution module further includes: The equipment status sensing unit detects the fan operating current through a current sensor, measures the pressure difference before and after the filter through a differential pressure sensor, calculates the filter clogging index based on the ratio of the pressure difference to the initial pressure difference, determines whether the filter has reached the state where it needs to be replaced, and sends the fan current, filter differential pressure, and clogging index to the collaborative control unit. The collaborative control unit receives the actual fan speed from the variable frequency drive unit and the path switching status signal from the multi-stage switching unit. During the filter switching process, it outputs a speed reduction command to reduce the fan speed and avoid airflow disturbance during switching. After the switching is completed, it controls the fan to restore the target speed and simultaneously synchronizes the filter clogging index from the equipment status sensing unit to the dynamic feedback adjustment module.

[0015] Preferably, the dynamic feedback adjustment module includes: The deviation calculation unit receives the purified parameters from the purification output monitoring unit, compares them with the preset air quality target value, calculates the deviation between the two, and triggers the feedback adjustment process when the deviation exceeds the set range, sending the deviation and deviation-related parameters to the feedback analysis unit. The formula for calculating the deviation is as follows:

[0016] In the formula, For deviation degree, For the first The concentration of the pollutants after purification For the first The preset target concentration of each pollutant The number of pollutant types included in the calculation; The feedback analysis unit receives the deviation data from the deviation calculation unit, combines it with the fan operating current and actual speed from the equipment status sensing unit, and determines the type of deviation: if the difference between the actual speed of the fan and the target speed is within the allowable range, it is determined to be a model prediction deviation; otherwise, it is determined to be an equipment execution deviation. The deviation type analysis results are then transmitted to the correction instruction generation unit. The correction instruction generation unit calculates the parameter correction coefficients of the efficiency mapping model for model prediction deviations and calculates the compensation value of the fan speed for equipment execution deviations. The correction coefficients and compensation values ​​are then sent to the efficiency mapping modeling unit of the filtering strategy generation module for strategy correction.

[0017] Preferably, the dynamic feedback adjustment module further includes: The adaptive learning unit periodically summarizes historical deviation data from the deviation calculation unit, correction parameters from the correction instruction generation unit, and purification efficiency data to build a model training sample set. It then uses an online learning algorithm to update the weight parameters of the neural network in the efficiency mapping modeling unit and periodically outputs the optimized model parameters to the efficiency mapping modeling unit. The adjustment parameter optimization unit calculates the deviation change rate based on the deviation change trend, and dynamically adjusts the proportional coefficient of the control module in combination with the change rate. The optimized proportional coefficient is then applied to the compensation value calculation of the correction instruction generation unit to optimize the adjustment effect.

[0018] Preferably, a home FFU air purification system based on high-efficiency filtration further includes: The target setting module receives air quality target parameters input by the user through the interactive interface, converts the target parameters into a data format compatible with the deviation calculation unit, and then sends them to the deviation calculation unit. The intelligent decision-making module receives the pollutant concentration prediction sequence from the air characteristic sensing module and the deviation from the dynamic feedback adjustment module, and combines the two to determine the current purification needs: if the predicted concentration is significantly higher than the target value and the deviation is small, the rapid purification mode is activated (increasing the fan speed and enabling the full-level filter); if the predicted concentration is significantly lower than the target value, the energy-saving mode is switched (reducing the fan speed to reduce energy consumption); the mode switching command is sent to the scene adaptation unit of the filtration strategy generation module to adjust the system operating status.

[0019] Compared with the prior art, the beneficial effects of the present invention are: 1. Based on real-time air quality feature vectors and combined with the performance parameters of FFU fans and filters, a mapping model between pollutant concentration and filtration efficiency is constructed through a neural network. Then, an optimization algorithm is used to solve for the optimal operating scheme, dynamically adjusting the fan speed and filter combination. Compared to a fixed mode, energy consumption can be reduced by 20%–30% while maintaining purification effectiveness.

[0020] 2. By calculating the pollutant change rate and accumulation trend through the time series analysis unit and predicting future concentration changes by combining the LSTM neural network, the system can start targeted purification in advance. At the same time, the dynamic feedback adjustment module continuously corrects the mapping model through deviation calculation and adaptive learning to solve the problem of efficiency decline caused by equipment aging and filter wear, and maintain the stability of purification effect in the long term.

[0021] 3. Based on user-defined scene modes such as sleep and cooking, the system dynamically adjusts the priority of pollutant purification and achieves smooth operation when switching filters through a collaborative control unit, balancing purification effect and scene comfort. Attached Figure Description

[0022] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0023] Figure 1 This is a system framework diagram of the present invention; Figure 2 This is a diagram showing the internal system framework of the air feature sensing module in this invention. Figure 3 This is an internal system framework diagram of the filtering strategy generation module in this invention; Figure 4 This is a diagram showing the internal system framework of the high-efficiency purification execution module in this invention; Figure 5 This is a diagram of the internal system framework of the dynamic feedback adjustment module in this invention. Detailed Implementation

[0024] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0026] Please see Figure 1 As shown, a home FFU air purification system based on high-efficiency filtration includes: an air characteristic sensing module, a filtration strategy generation module, a high-efficiency purification execution module, and a dynamic feedback adjustment module. The air feature sensing module is used to collect multi-dimensional air pollutant parameters indoors and normalize the multi-dimensional air pollutant parameters to generate a real-time air quality feature vector. The filtration strategy generation module is electrically connected to the air feature sensing module. The filtration strategy generation module is used to construct a pollutant concentration filtration efficiency mapping model based on the real-time air quality feature vector, combined with the rated parameters of the FFU fan and the filter performance curve, and generate a dynamic filtration level combination strategy. The high-efficiency purification execution module is electrically connected to the filtration strategy generation module. The high-efficiency purification execution module is used to drive the FFU fan frequency conversion speed regulation unit and the multi-stage filter switching mechanism to work together according to the dynamic filtration level combination strategy to complete the air purification process and output the purified air parameters. The dynamic feedback adjustment module is electrically connected to the air feature sensing module and the high-efficiency purification execution module respectively. The dynamic feedback adjustment module is used to calculate the deviation between the purified air parameters and the preset target values, and to feed back and correct the mapping model parameters of the filtration strategy generation module to perform adaptive high-efficiency purification closed-loop control. The target setting module receives air quality target parameters input by the user through the interactive interface, converts the target parameters into a data format compatible with the deviation calculation unit, and then sends them to the deviation calculation unit. The intelligent decision-making module receives the pollutant concentration prediction sequence from the air characteristic sensing module and the deviation from the dynamic feedback adjustment module, and combines the two to determine the current purification needs: if the predicted concentration is significantly higher than the target value and the deviation is small, the rapid purification mode is activated (increasing the fan speed and enabling the full-level filter); if the predicted concentration is significantly lower than the target value, the energy-saving mode is switched (reducing the fan speed to reduce energy consumption); the mode switching command is sent to the scene adaptation unit of the filtration strategy generation module to adjust the system operating status.

[0027] Please see Figure 2 As shown, the air feature sensing module includes: The multi-source sensing unit collects PM2.5 concentration, VOCs concentration, formaldehyde concentration, and ambient temperature and humidity by deploying laser dust sensors, gas sensor arrays, and temperature and humidity sensors in different indoor areas, respectively, and obtains multi-dimensional raw sensing data, which is then transmitted to the data preprocessing unit. The data preprocessing unit performs preprocessing on the received multi-dimensional raw sensing data, including: using an outlier removal algorithm to remove abnormal data caused by instantaneous sensor interference; using a filtering algorithm to eliminate high-frequency noise to smooth the data; performing range calibration before data normalization to unify the data benchmark and obtain stable sensing data; and sending the stable sensing data to the feature vector generation unit. The feature vector generation unit receives stable sensing data, performs normalization transformation according to preset pollutant and environmental parameter weight coefficients, maps each parameter to a unified numerical range, generates a real-time air quality feature vector containing pollutant concentration information and environmental parameter information, and transmits the feature vector to the filtration strategy generation module. The time series analysis unit receives the real-time air quality feature vector output by the feature vector generation unit, constructs a data sequence of continuous time periods according to the timestamp, calculates the pollutant concentration change rate and cumulative pollution trend characteristics, and obtains trend feature parameters; the trend feature parameters are then transmitted to the pollution prediction unit. The formula for calculating the pollutant concentration change rate is as follows:

[0028] In the formula, The rate of change of pollutant concentration. For the current moment Measured values ​​of pollutant concentration The measured value of pollutant concentration at time _____. The time interval used for calculation; The formula for calculating the cumulative pollution trend characteristics is as follows:

[0029] In the formula, The cumulative pollution index, To calculate the cumulative number of time windows, These are the measured values ​​of pollutant concentrations at the corresponding times. The pollution prediction unit takes the trend feature parameters output by the time series analysis unit as input, performs prediction calculations through a trained time series prediction neural network, outputs a sequence of predicted pollutant concentration values ​​for a set time period in the future, and synchronizes the prediction sequence to the scene adaptation unit of the filtration strategy generation module. The prediction formula is as follows:

[0030] In the formula, For the future Predicted pollutant concentration at time ( (For predicting step size, e.g., 10 min / step). For a moment The trend characteristic parameter sequence, for The weight parameter matrix of a neural network includes the weights of the input gate, forget gate, cell state, and output gate. for The bias vector of a neural network. for Network forward computation function; To address the issues of "single-point sensing and data distortion" in traditional equipment, a multi-region, multi-source sensor array is employed to achieve multi-dimensional, all-space pollutant collection. Data stability is ensured through three preprocessing steps: outlier removal, filtering, and range calibration. A normalized feature vector is then used to unify the data benchmark, providing high-precision, highly consistent input data for subsequent strategy generation. This avoids decision-making errors caused by single-parameter biases and breaks through the limitations of traditional equipment's passive response to pollution. By extracting the dual-trend features of pollutant change rate and cumulative index through time-series analysis, and combining this with an LSTM time-series prediction neural network, future concentration predictions can be achieved. This allows for early identification of rising pollution trends and initiation of intervention, avoiding the lag of purification after pollutant concentrations exceed standards and improving the foresight of indoor air quality management.

[0031] Please see Figure 3 As shown, the filtering strategy generation module includes: The basic parameter library unit pre-stores three types of core parameters, including: the rated operating parameters and power characteristics of the FFU fan, the filtration efficiency, resistance characteristics and dust holding capacity parameters of the filter. These parameters are provided as basic data to the efficiency mapping modeling unit. The efficiency mapping modeling unit receives the real-time air quality feature vector output by the air feature sensing module, combines it with the equipment parameters of the basic parameter library unit, constructs a mapping model between pollutant concentration and filtration efficiency through a neural network, calculates the purification efficiency under different fan operating states and filter combinations, and generates a mapping relationship table of fan operating parameters, filter combinations and purification efficiency, which is then transmitted to the strategy optimization unit. The strategy optimization unit uses the mapping table as the basic data and combines it with preset energy consumption constraints to solve the optimal operation scheme using an optimization algorithm, thereby obtaining a dynamic filtration level combination strategy that includes the target operating parameters of the fan and the activation level of the filter; the strategy is then sent to the high-efficiency purification execution module. The scene adaptation unit receives the pollutant concentration prediction value sequence output by the pollution prediction unit, combines it with the scene mode set by the user, including sleep, daily activities and cooking scenes, calls the pre-stored operation constraint parameters and purification priority parameters corresponding to each scene, and passes the scene parameters to the dynamic weight adjustment unit. The dynamic weight adjustment unit adjusts the pollutant weight coefficients in the efficiency mapping model according to the scenario parameters of the scenario adaptation unit, prioritizes the purification efficiency of high-impact pollutants in the scenario, recalculates the mapping relationship table and updates it to the strategy optimization unit, thereby realizing scenario-based filtration strategy optimization. Addressing the rigid strategies of traditional equipment with fixed filter combinations and single fan speeds, this invention, for the first time, correlates equipment physical parameters with air quality feature vectors and constructs a multivariate mapping model using a neural network. With the goal of achieving optimal purification efficiency and minimum energy consumption, a customized strategy is solved through optimization algorithms. This enables precise matching of fan and filter combinations under different pollution scenarios, avoiding excessive energy consumption or insufficient purification. It also solves the problem of poor scenario adaptability of traditional equipment. A scenario adaptation unit calls upon constraint parameters for different scenarios, and a dynamic weight adjustment unit specifically adjusts the pollutant purification priority; for example, increasing the formaldehyde weight during cooking and decreasing the fan speed weight during sleep. This ensures that the purification strategy is deeply matched to the user's actual needs, balancing effectiveness and comfort.

[0032] Please see Figure 4 As shown, the high-efficiency purification execution module includes: The variable frequency drive unit receives the target speed of the fan output by the strategy optimization unit, converts the speed command into a motor drive signal through the variable frequency control module, drives the fan motor to run at the target speed, and collects the actual speed of the fan in real time and feeds it back to the cooperative control unit. The multi-level switching unit receives filter level instructions from the dynamic filter level combination strategy, controls valve actions to switch the airflow path of different filter levels, and sends a status signal to the cooperative control unit after the switching is completed. The purification output monitoring unit deploys a group of miniature sensors at the FFU outlet to collect the pollutant concentration in the purified air, calculates the purification efficiency based on the pollutant concentration before and after purification, and transmits the purified parameters and purification efficiency values ​​to the dynamic feedback adjustment module. The equipment status sensing unit detects the fan operating current through a current sensor, measures the pressure difference before and after the filter through a differential pressure sensor, calculates the filter clogging index based on the ratio of the pressure difference to the initial pressure difference, determines whether the filter has reached the state where it needs to be replaced, and sends the fan current, filter differential pressure, and clogging index to the collaborative control unit. The collaborative control unit receives the actual fan speed from the variable frequency drive unit and the path switching status signal from the multi-stage switching unit. During the filter switching process, it outputs a speed reduction command to reduce the fan speed and avoid airflow disturbance during switching. After the switching is completed, it controls the fan to restore the target speed and synchronizes the filter blockage index from the equipment status sensing unit to the dynamic feedback adjustment module. Breaking through the limitations of traditional equipment lacking health monitoring and experiencing turbulent airflow during switching, this device uses a clogging index to determine filter lifespan in real time, preventing a decrease in purification efficiency after filter failure. Simultaneously, it reduces speed during filter switching to minimize noise and pressure fluctuations caused by airflow impact, ensuring stable equipment operation. Furthermore, it synchronizes equipment status data to a feedback module, providing a reference for strategy adjustments based on equipment health dimensions.

[0033] Please see Figure 5 As shown, the dynamic feedback adjustment module includes: The deviation calculation unit receives the purified parameters from the purification output monitoring unit, compares them with the preset air quality target value, calculates the deviation between the two, and triggers the feedback adjustment process when the deviation exceeds the set range, sending the deviation and deviation-related parameters to the feedback analysis unit. The formula for calculating the deviation is as follows:

[0034] In the formula, For deviation degree, For the first The concentration of the pollutants after purification For the first The preset target concentration of each pollutant The number of pollutant types included in the calculation; The feedback analysis unit receives the deviation data from the deviation calculation unit, combines it with the fan operating current and actual speed from the equipment status sensing unit, and determines the type of deviation: if the difference between the actual speed of the fan and the target speed is within the allowable range, it is determined to be a model prediction deviation; otherwise, it is determined to be an equipment execution deviation. The deviation type analysis results are then transmitted to the correction instruction generation unit. The correction instruction generation unit calculates the parameter correction coefficients of the efficiency mapping model for model prediction deviations and calculates the compensation value of the fan speed for equipment execution deviations. The correction coefficients and compensation values ​​are then sent to the efficiency mapping modeling unit of the filtering strategy generation module for strategy correction.

[0035] The dynamic feedback adjustment module also includes: The adaptive learning unit periodically summarizes historical deviation data from the deviation calculation unit, correction parameters from the correction instruction generation unit, and purification efficiency data to build a model training sample set. It then uses an online learning algorithm to update the weight parameters of the neural network in the efficiency mapping modeling unit and periodically outputs the optimized model parameters to the efficiency mapping modeling unit. The parameter optimization unit calculates the deviation change rate based on the deviation change trend, dynamically adjusts the proportional coefficient of the control module based on the change rate, and applies the optimized proportional coefficient to the compensation value calculation of the correction instruction generation unit to optimize the adjustment effect. To address the issue of declining accuracy in traditional models over long-term operation, an adaptive learning unit continuously optimizes neural network weights using historical data, enabling the mapping model to adapt to long-term changes such as equipment aging and filter wear. Simultaneously, by dynamically adjusting the control proportional coefficient, overshoot or oscillations during feedback adjustment are avoided, ensuring the stability and accuracy of the system's long-term operation and achieving intelligent characteristics of "self-learning and self-optimization."

[0036] In summary, the advantages of this invention are as follows: By deploying a multi-source sensor array, parameters such as PM2.5, VOCs, formaldehyde, temperature and humidity are collected simultaneously. After preprocessing such as outlier removal, filtering and smoothing, and range calibration, combined with normalized feature vector generation technology, a comprehensive and stable perception of indoor air quality is achieved, providing a precise data foundation for subsequent purification strategies. Based on real-time air quality feature vectors and combined with the performance parameters of FFU fans and filters, a mapping model between pollutant concentration and filtration efficiency is constructed through a neural network. Then, the optimal operating scheme is solved through optimization algorithms, and the fan speed and filter combination are dynamically adjusted. Compared with the fixed mode, energy consumption can be reduced by 20% to 30% while ensuring purification effect. The time-series analysis unit calculates the pollutant change rate and accumulation trend, and combines it with the LSTM neural network to predict future concentration changes, enabling the system to start targeted purification in advance. At the same time, the dynamic feedback adjustment module continuously corrects the mapping model through deviation calculation and adaptive learning to solve the problem of efficiency decline caused by equipment aging and filter wear, and maintains the stability of purification effect in the long term. Combined with user-defined scene modes such as sleep and cooking, the pollutant purification priority is dynamically adjusted. For example, when cooking, the formaldehyde concentration is reduced first, and when sleeping, noise and energy consumption are controlled first. The smooth operation of the filter switching is achieved through the collaborative control unit, taking into account both purification effect and scene comfort. The device status sensing unit monitors the filter clogging index and fan operating status in real time, promptly reminding users to replace the filter to avoid a decrease in purification efficiency due to filter failure. At the same time, the adaptive learning unit continuously optimizes model parameters, reducing the need for manual intervention and lowering user maintenance costs.

[0037] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A household FFU air purification system based on high-efficiency filtration, characterized in that, include: Air characteristic sensing module, filtration strategy generation module, high-efficiency purification execution module, dynamic feedback adjustment module; The air feature sensing module is used to collect multi-dimensional air pollutant parameters indoors and normalize the multi-dimensional air pollutant parameters to generate a real-time air quality feature vector. The filtration strategy generation module is electrically connected to the air feature sensing module. The filtration strategy generation module is used to construct a pollutant concentration filtration efficiency mapping model based on the real-time air quality feature vector, combined with the rated parameters of the FFU fan and the filter performance curve, and generate a dynamic filtration level combination strategy. The high-efficiency purification execution module is electrically connected to the filtration strategy generation module. The high-efficiency purification execution module is used to drive the FFU fan frequency conversion speed regulation unit and the multi-stage filter switching mechanism to work together according to the dynamic filtration level combination strategy to complete the air purification process and output the purified air parameters. The dynamic feedback adjustment module is electrically connected to the air feature sensing module and the high-efficiency purification execution module, respectively. The dynamic feedback adjustment module is used to calculate the deviation between the purified air parameters and the preset target values, and to feed back and correct the mapping model parameters of the filtration strategy generation module to perform adaptive high-efficiency purification closed-loop control.

2. A household FFU air purification system based on high-efficiency filtration according to claim 1, characterized in that, The air feature sensing module includes: The multi-source sensing unit collects PM2.5 concentration, VOCs concentration, formaldehyde concentration, and ambient temperature and humidity by deploying laser dust sensors, gas sensor arrays, and temperature and humidity sensors in different indoor areas, respectively, and obtains multi-dimensional raw sensing data, which is then transmitted to the data preprocessing unit. The data preprocessing unit performs preprocessing on the received multi-dimensional raw sensing data, including: using an outlier removal algorithm to remove abnormal data caused by instantaneous sensor interference; using a filtering algorithm to eliminate high-frequency noise to smooth the data; performing range calibration before data normalization to unify the data benchmark and obtain stable sensing data; and sending the stable sensing data to the feature vector generation unit. The feature vector generation unit receives stable sensing data, performs normalization transformation according to preset pollutant and environmental parameter weight coefficients, maps each parameter to a unified numerical range, and generates a real-time air quality feature vector containing pollutant concentration information and environmental parameter information; and transmits the feature vector to the filtration strategy generation module.

3. A household FFU air purification system based on high-efficiency filtration according to claim 1, characterized in that, The air feature sensing module also includes: The time series analysis unit receives the real-time air quality feature vector output by the feature vector generation unit, constructs a data sequence of continuous time periods according to the timestamp, calculates the pollutant concentration change rate and cumulative pollution trend characteristics, and obtains trend feature parameters; the trend feature parameters are then transmitted to the pollution prediction unit. The pollution prediction unit takes the trend feature parameters output by the time series analysis unit as input, performs prediction calculations through a trained time series prediction neural network, and outputs a sequence of predicted pollutant concentration values ​​for a set future time period. This prediction sequence is then synchronized to the scene adaptation unit of the filtration strategy generation module.

4. A household FFU air purification system based on high-efficiency filtration according to claim 1, characterized in that, The filtering strategy generation module includes: The basic parameter library unit pre-stores three types of core parameters, including: the rated operating parameters and power characteristics of the FFU fan, the filtration efficiency, resistance characteristics and dust holding capacity parameters of the filter. These parameters are provided as basic data to the efficiency mapping modeling unit. The efficiency mapping modeling unit receives the real-time air quality feature vector output by the air feature sensing module, combines it with the equipment parameters of the basic parameter library unit, constructs a mapping model between pollutant concentration and filtration efficiency through a neural network, calculates the purification efficiency under different fan operating states and filter combinations, and generates a mapping relationship table of fan operating parameters, filter combinations and purification efficiency, which is then transmitted to the strategy optimization unit. The strategy optimization unit uses the mapping table as the basic data and combines it with preset energy consumption constraints to solve the optimal operation scheme using an optimization algorithm, thereby obtaining a dynamic filtration level combination strategy that includes the target operating parameters of the fan and the activation level of the filter; this strategy is then sent to the high-efficiency purification execution module.

5. A household FFU air purification system based on high-efficiency filtration according to claim 4, characterized in that, The filtering strategy generation module also includes: The scene adaptation unit receives the pollutant concentration prediction value sequence output by the pollution prediction unit, combines it with the scene mode set by the user, including sleep, daily activities and cooking scenes, calls the pre-stored operation constraint parameters and purification priority parameters corresponding to each scene, and passes the scene parameters to the dynamic weight adjustment unit. The dynamic weight adjustment unit adjusts the pollutant weight coefficients in the efficiency mapping model based on the scenario parameters of the scenario adaptation unit, prioritizing the purification efficiency of high-impact pollutants within the scenario, recalculating the mapping relationship table and updating it to the strategy optimization unit, thereby achieving scenario-based filtration strategy optimization.

6. A household FFU air purification system based on high-efficiency filtration according to claim 1, characterized in that, The high-efficiency purification execution module includes: The variable frequency drive unit receives the target speed of the fan output by the strategy optimization unit, converts the speed command into a motor drive signal through the variable frequency control module, drives the fan motor to run at the target speed, and collects the actual speed of the fan in real time and feeds it back to the cooperative control unit. The multi-level switching unit receives filter level instructions from the dynamic filter level combination strategy, controls valve actions to switch the airflow path of different filter levels, and sends a status signal to the cooperative control unit after the switching is completed. The purification output monitoring unit deploys a group of miniature sensors at the FFU outlet to collect the pollutant concentration in the purified air, calculates the purification efficiency based on the pollutant concentrations before and after purification, and transmits the purified parameters and purification efficiency values ​​to the dynamic feedback adjustment module.

7. A household FFU air purification system based on high-efficiency filtration according to claim 6, characterized in that, The high-efficiency purification execution module also includes: The equipment status sensing unit detects the fan operating current through a current sensor, measures the pressure difference before and after the filter through a differential pressure sensor, calculates the filter clogging index based on the ratio of the pressure difference to the initial pressure difference, determines whether the filter has reached the state where it needs to be replaced, and sends the fan current, filter differential pressure, and clogging index to the collaborative control unit. The collaborative control unit receives the actual fan speed from the variable frequency drive unit and the path switching status signal from the multi-stage switching unit. During the filter switching process, it outputs a speed reduction command to reduce the fan speed and avoid airflow disturbance during switching. After the switching is completed, it controls the fan to restore the target speed and simultaneously synchronizes the filter clogging index from the equipment status sensing unit to the dynamic feedback adjustment module.

8. A household FFU air purification system based on high-efficiency filtration according to claim 1, characterized in that, The dynamic feedback adjustment module includes: The deviation calculation unit receives the purified parameters from the purification output monitoring unit, compares them with the preset air quality target value, calculates the deviation between the two, and triggers the feedback adjustment process when the deviation exceeds the set range, sending the deviation and deviation-related parameters to the feedback analysis unit. The feedback analysis unit receives the deviation data from the deviation calculation unit, combines it with the fan operating current and actual speed from the equipment status sensing unit, and determines the type of deviation: if the difference between the actual speed of the fan and the target speed is within the allowable range, it is determined to be a model prediction deviation; otherwise, it is determined to be an equipment execution deviation. The deviation type analysis results are then transmitted to the correction instruction generation unit. The correction instruction generation unit calculates the parameter correction coefficients of the efficiency mapping model for model prediction deviations and calculates the compensation value of the fan speed for equipment execution deviations. The correction coefficients and compensation values ​​are then sent to the efficiency mapping modeling unit of the filtering strategy generation module for strategy correction.

9. A household FFU air purification system based on high-efficiency filtration according to claim 8, characterized in that, The dynamic feedback adjustment module also includes: The adaptive learning unit periodically summarizes historical deviation data from the deviation calculation unit, correction parameters from the correction instruction generation unit, and purification efficiency data to build a model training sample set. It then uses an online learning algorithm to update the weight parameters of the neural network in the efficiency mapping modeling unit and periodically outputs the optimized model parameters to the efficiency mapping modeling unit. The adjustment parameter optimization unit calculates the deviation change rate based on the deviation change trend, and dynamically adjusts the proportional coefficient of the control module in combination with the change rate. The optimized proportional coefficient is then applied to the compensation value calculation of the correction instruction generation unit to optimize the adjustment effect.

10. A household FFU air purification system based on high-efficiency filtration according to claim 1, characterized in that, Also includes: The target setting module receives air quality target parameters input by the user through the interactive interface, converts the target parameters into a data format compatible with the deviation calculation unit, and then sends them to the deviation calculation unit. The intelligent decision-making module receives the pollutant concentration prediction sequence from the air feature sensing module and the deviation from the dynamic feedback adjustment module, and combines the two to determine the current purification needs: if the predicted concentration is significantly higher than the target value and the deviation is small, the rapid purification mode is activated; if the predicted concentration is significantly lower than the target value, the energy-saving mode is switched, and the mode switching command is sent to the scene adaptation unit of the filtration strategy generation module to adjust the system operating status.

Images