Multi-area FFU differential pressure collaborative dynamic balance control method

Abstract

The invention discloses a multi-area FFU pressure difference collaborative dynamic balance control method, which belongs to the field of industrial automation and intelligent control, and comprises the following steps: collecting FFU operation parameters, environmental parameters and energy consumption data in multiple clean spaces in real time; constructing an energy consumption evaluation model and a multi-dimensional prediction model based on historical data; generating an FFU group cooperative control strategy according to the predicted energy consumption data, the pressure difference change trend and the human traffic change trend; and a dynamic adjustment instruction is issued, and parameters are continuously collected to carry out adaptive optimization on the model, so that continuous dynamic balance of the multi-region FFU pressure difference is ensured. According to the method, predictive control and multi-dimensional data fusion technologies are adopted, intelligent group cooperative control and energy consumption optimization of the FFU can be achieved, and the pressure difference stability and the operation efficiency of the clean space are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of industrial automation and intelligent control, and in particular to a control method for coordinated dynamic balance of differential pressure of FFUs in multiple zones. BACKGROUND

[0002] Fan filter units (FFUs) are critical equipment in cleanrooms for maintaining air cleanliness and specific differential pressure gradients, widely used in industries with extremely high environmental requirements such as semiconductor manufacturing, biopharmaceuticals, and precision instruments. In large cleanrooms, there are usually multiple interconnected or isolated clean spaces, each of which needs to maintain precise differential pressure to prevent cross-contamination. Therefore, effective control of FFU groups in multiple zones to achieve coordinated dynamic balance of differential pressure in multiple zones is one of the core technologies to ensure the qualification of production environment and the stability of product quality. In the prior art, the control of multiple-zone FFUs usually adopts decentralized independent control or simple linkage control strategy. In decentralized control, the FFU controller of each clean space only adjusts according to the feedback of the differential pressure sensor in the region, lacking coordination between regions. While the simple linkage control considers the mutual influence between regions, it is mostly based on fixed logical rules or static models, for example, after the FFU speed in one region changes, the FFUs in adjacent regions are adjusted according to a preset proportion.

[0003] However, the existing technical solutions have obvious defects. Due to the complexity and dynamics of the cleanroom environment, factors such as personnel flow, equipment door opening and closing, and process exhaust changes will have complex and nonlinear effects on the differential pressure of multiple zones. Decentralized control cannot cope with the coupling effects between regions, often leading to the rise and fall of differential pressure and continuous oscillation. While the simple linkage control based on fixed rules lacks adaptability to dynamic changes and cannot predictively handle environmental disturbances, resulting in lagging control response and low precision, making it difficult to optimize system energy consumption while ensuring stable differential pressure in all regions. SUMMARY

[0004] To solve the above problems, the present application provides a control method for coordinated dynamic balance of differential pressure of FFUs in multiple zones, which adopts a technical solution combining predictive control and adaptive optimization, and can achieve coordinated stability of differential pressure in multiple zones and optimization of energy consumption.

[0005] The above objectives can be achieved by the following solutions: The application discloses a control method for multi-zone FFU differential pressure coordination dynamic balance, comprising the following steps: collecting FFU running parameters, environment parameters and energy consumption data in multiple clean spaces in real time, and obtaining corresponding historical data in the multiple clean spaces, wherein the environment parameters comprise differential pressure data, particle concentration data and passenger flow data; constructing an energy consumption evaluation model according to the historical data, evaluating the energy consumption state of the FFU in each clean space in real time, comparing the energy consumption data collected in real time, and optimizing the parameters of the energy consumption evaluation model; inputting the FFU running parameters and environment parameters collected in real time into the optimized energy consumption evaluation model to obtain predicted energy consumption data in each clean space in a future period of time; based on the FFU running parameters, environment parameters and historical data, a multi-dimensional prediction model is established to generate predicted differential pressure change trends and passenger flow change trends in each clean space in the future period of time; according to the predicted energy consumption data, differential pressure change trends and passenger flow change trends, FFU group coordination control strategies are generated for the multiple clean spaces, wherein the FFU group coordination control strategies are used to guide the working mode and rotating speed of the FFUs in the multiple clean spaces; based on the FFU group coordination control strategies, dynamic adjustment instructions are issued to the FFUs in each clean space; after the dynamic adjustment instructions are issued, the FFU running parameters, environment parameters and energy consumption data are continuously collected, and the multi-dimensional prediction model and the energy consumption evaluation model are adaptively optimized based on the continuously collected parameters, so that the continuous dynamic balance of the multi-zone FFU differential pressure is ensured.

[0006] Optionally, according to the historical data, the energy consumption evaluation model is constructed to evaluate the energy consumption state of the FFU in each clean space in real time, which comprises the following steps: according to the FFU running parameters, environment parameters and energy consumption data contained in the historical data, an initial training is performed on the energy consumption evaluation model to be trained by using a reinforcement learning algorithm, wherein the reinforcement learning algorithm takes the energy consumption data in the historical data as a reward function; the FFU running parameters and environment parameters before the current time are input into the energy consumption evaluation model to obtain a current time energy consumption evaluation result of the FFU; the current time energy consumption evaluation result is compared with the energy consumption data collected in real time to obtain an evaluation difference, and the parameters of the energy consumption evaluation model are optimized by using an error back propagation algorithm according to the evaluation difference.

[0007] Optionally, the parameters of the energy consumption evaluation model are optimized by using the error back propagation algorithm according to the evaluation difference, which comprises the following steps: according to the evaluation difference, the differential pressure data and the passenger flow data, dynamic weights are assigned to the evaluation difference; according to the evaluation difference to which the dynamic weights are assigned, the gradient values of each parameter in the energy consumption evaluation model are calculated by using the error back propagation algorithm, and the parameters of the energy consumption evaluation model are updated based on the gradient values to reduce the evaluation difference.

[0008] Optionally, the establishing the multi-dimensional prediction model comprises: generating, in time sequence, an FFU operation parameter time sequence and an environment parameter time sequence from the FFU operation parameters and the environment parameters contained in the historical data; performing multi-modal feature fusion on the FFU operation parameter time sequence and the environment parameter time sequence to obtain a fusion feature vector; and training the multi-dimensional prediction model to be trained by using a deep neural network based on a long short-term memory network and taking the fusion feature vector as input and predicting a differential pressure change trend and a people flow change trend as output.

[0009] Optionally, the generating the FFU group cooperative control strategy for the plurality of clean spaces according to the predicted energy consumption data, the differential pressure change trend and the people flow change trend comprises: presetting a group cooperative control strategy rule library containing FFU operation modes, rotating speeds and energy-saving control rules; assigning a dynamic weight to each rule in the FFU group cooperative control strategy rule library by using a fuzzy logic algorithm according to the predicted energy consumption data, the differential pressure change trend and the people flow change trend; and matching and generating the FFU group cooperative control strategy from the FFU group cooperative control strategy rule library according to the dynamic weight while meeting the requirements of minimum energy consumption and dynamic balance of differential pressure.

[0010] Optionally, the presetting the group cooperative control strategy rule library containing FFU operation modes, rotating speeds and energy-saving control rules comprises: presetting an FFU operation mode rule set of an FFU standby mode, a low-speed operation mode and a full-speed operation mode; presetting an FFU operation parameter rule set of an FFU wind speed adjustment range and an air volume adjustment range; and presetting an FFU energy-saving control rule set of an FFU differential pressure threshold value, an energy consumption threshold value and a people flow threshold value.

[0011] Optionally, the issuing a dynamic adjustment instruction to each FFU in the clean space based on the FFU group cooperative control strategy comprises: converting the FFU group cooperative control strategy into an FFU dynamic adjustment instruction, wherein the FFU dynamic adjustment instruction comprises an FFU target operation mode and a target rotating speed; and issuing the FFU dynamic adjustment instruction to each FFU in the clean space, and in the issuing process, instructions with higher energy consumption optimization weights and higher differential pressure balance weights are preferentially executed.

[0012] Optionally, the priority execution of the instructions with higher energy consumption optimization weight and higher pressure difference balance weight in the issuing process comprises: evaluating the energy consumption abnormality degree of each FFU in the FFU group cooperative control strategy according to the real-time collected energy consumption data and the predicted energy consumption data, and assigning an energy consumption optimization weight to each FFU according to the energy consumption abnormality degree; evaluating the contribution degree of the FFU to the pressure difference stability in each clean space according to the real-time collected pressure difference data and the pressure difference change trend, and assigning a pressure difference balance weight to each FFU according to the contribution degree; and calculating the comprehensive priority score of the FFU dynamic adjustment instruction by using a weighted summation algorithm according to the energy consumption optimization weight and the pressure difference balance weight, and executing the dynamic adjustment instruction in order from high to low according to the comprehensive priority score.

[0013] Optionally, the adaptive optimization of the multi-dimensional prediction model and the energy consumption evaluation model based on the continuously collected parameters to ensure the continuous dynamic balance of the multi-zone FFU pressure difference comprises: updating the model parameters by performing incremental learning on the multi-dimensional prediction model and the energy consumption evaluation model with the continuously collected FFU operation parameters, environmental parameters and energy consumption data as an incremental data set; performing performance evaluation on the updated model, and if the performance evaluation result is better than the current model, completing the model update; and if the performance evaluation result is not better than the current model, restoring the model to the state before the update.

[0014] Based on the same inventive concept, the application also provides a control system for the cooperative dynamic balance of the multi-zone FFU pressure difference, which comprises: a data collection module for collecting FFU operation parameters, environmental parameters and energy consumption data in a plurality of clean spaces in real time, and obtaining historical data; an energy consumption evaluation module for constructing an energy consumption evaluation model according to the historical data, and evaluating the energy consumption state of the FFU in each clean space in real time; a multi-dimensional prediction module for establishing a multi-dimensional prediction model according to the FFU operation parameters, environmental parameters and historical data, and generating a predicted pressure difference change trend and a predicted human flow change trend in each clean space in a future period of time; a cooperative control strategy module for generating an FFU group cooperative control strategy for a plurality of clean spaces according to the predicted energy consumption data, pressure difference change trend and human flow change trend; an instruction issuing module for issuing a dynamic adjustment instruction to the FFU in each clean space based on the FFU group cooperative control strategy; and an adaptive optimization module for continuously collecting FFU operation parameters, environmental parameters and energy consumption data after the dynamic adjustment instruction is issued, and adaptively optimizing the multi-dimensional prediction model and the energy consumption evaluation model based on the continuously collected parameters.

[0015] Compared with the prior art, the application has the following advantages: 1、The present application can make prospective prediction on the change trend of the flow and the differential pressure itself which affect the differential pressure of the clean space by constructing a multi-dimensional prediction model. This prediction ability enables the change from the traditional lag response to active pre-adjustment, and takes intervention measures before or at the initial stage of environmental disturbance, thereby effectively suppressing the fluctuation of the differential pressure and greatly improving the stability and reliability of the multi-zone clean environment.

[0016] 2、The present application fuses the predicted energy consumption data, the change trend of the differential pressure and the change trend of the flow as the comprehensive decision basis for generating the cooperative control strategy. This multi-objective fusion decision mechanism ensures that the differential pressure stability and energy consumption economy can be intelligently balanced when the control is executed, and under the premise of guaranteeing the cleanliness requirement, a globally optimal energy consumption operation scheme is dynamically found, thereby realizing the cooperative optimization of the differential pressure control and energy saving.

[0017] 3、The present application designs a complete closed-loop adaptive optimization process, and performs incremental learning and performance verification on the energy consumption evaluation model and the multi-dimensional prediction model through continuous acquisition of operation data. This endows the ability of self-evolution and adaptation to environmental changes, and enables the dynamic tracking and adaptation to the long-term working condition changes such as equipment aging and filter screen clogging, thereby guaranteeing the effectiveness and robustness of the control method in the long-term operation process and maintaining the sustained high efficiency of the system performance.

[0018] Other features and advantages of the present application will be set forth in the following description, and in part will become apparent to those skilled in the art from the description, or can be learned by practice of the present application. The objects and other advantages of the present application can be realized and attained by the structure particularly pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0020] Figure 1 is a flow diagram of a multi-zone FFU differential pressure cooperative dynamic balance control method according to an embodiment of the present application.

[0021] Figure 2 is a system performance comparison chart under different control strategies according to an embodiment of the present application.

[0022] Figure 3 is a FFU dynamic adjustment instruction priority timing chart according to an embodiment of the present application.

[0023] Figure 4is a structural schematic diagram of a multi-zone FFU differential pressure cooperative dynamic balance control system according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in connection with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort belong to the scope of protection of the present application.

[0025] Reference Figure 1 An embodiment of the present application proposes a multi-zone FFU differential pressure cooperative dynamic balance control method, which adopts a technical solution combining predictive control and adaptive optimization, can realize cooperative stability of multi-zone differential pressure and optimization of energy consumption, and significantly improves the stability and operating efficiency of a multi-zone clean environment. The method according to the embodiment specifically includes the following steps. Real-time collection of FFU operating parameters, environmental parameters, and energy consumption data in multiple clean spaces, and acquisition of corresponding historical data in the multiple clean spaces, wherein the environmental parameters include differential pressure data, particle concentration data, and passenger flow data; Specifically, the wind speed, rotating speed, power, and other operating parameters of the FFU in each clean space are acquired in real time. At the same time, the sensors distributed throughout the clean room sense and acquire the current differential pressure data, particle concentration data, and passenger flow data of the environment in real time. The historical data of the FFU operating parameters, environmental parameters, and energy consumption data in each clean space are acquired and stored to form real-time data streams and a historical database.

[0026] According to the historical data, an energy consumption evaluation model is constructed to evaluate the energy consumption state of the FFU in each clean space in real time, and compared with the real-time collected energy consumption data to optimize the parameters of the energy consumption evaluation model; Specifically, an energy consumption evaluation model is constructed, the FFU operating parameters and environmental parameters in the historical data are taken as inputs, the predicted energy consumption of the FFU is calculated, and the predicted result is compared with the energy consumption data in the historical data to train the energy consumption evaluation model. In addition, the FFU operating parameters and environmental parameters before the current time are input into the trained energy consumption evaluation model to obtain the predicted energy consumption data at the current time, the predicted energy consumption data at the current time is compared with the real-time collected energy consumption data, the model parameters are optimized, and the evaluation accuracy is improved.

[0027] The real-time collected FFU operating parameters and environmental parameters are input into the optimized energy consumption evaluation model to obtain the predicted energy consumption data of each clean space in a future period of time; Specifically, the real-time collected FFU operating parameters and environmental parameters are input into the optimized energy consumption evaluation model. According to these inputs, the energy consumption trend of the FFU in the next hour or more can be predicted. For example, when the rotating speed of the FFU continues to rise, the model can predict that the energy consumption of the FFU in the future will also rise accordingly.

[0028] Based on the FFU operating parameters, environmental parameters and historical data, a multi-dimensional prediction model is established to generate the predicted pressure difference change trend and the people flow change trend in each clean space in the future period of time; Specifically, the historical data is used to train the multi-dimensional prediction model. The complex nonlinear relationship between the FFU operating state, environmental changes and the pressure difference and people flow can be learned. In real-time operation, the model can predict the possible pressure difference change and people flow change in each clean space in the future according to the current state. For example, the model can predict the peak period of people flow in the next hour, and predict the downward trend of the pressure difference accordingly.

[0029] According to the predicted energy consumption data, the pressure difference change trend and the people flow change trend, a FFU group cooperative control strategy is generated for multiple clean spaces, wherein the FFU group cooperative control strategy is used to guide the working mode and rotating speed of the FFU in multiple clean spaces; Specifically, after receiving the prediction data from the energy consumption evaluation model and the multi-dimensional prediction model, a group cooperative control strategy is generated for the FFUs in all clean spaces according to the prediction data. The generation process of this strategy considers the stability of the pressure difference and the economy of energy consumption, and is a multi-objective optimization decision. For example, when it is predicted that the pressure difference will drop soon but the energy consumption is high, a cooperative control strategy may be generated, which may include instructions to reduce the rotating speed of some FFUs and increase the rotating speed of other FFUs, so as to maintain the stability of the overall pressure difference while optimizing the energy consumption.

[0030] Based on the FFU group cooperative control strategy, dynamic adjustment instructions are issued to the FFUs in each clean space; Specifically, the generated FFU group cooperative control strategy is converted into FFU dynamic adjustment instructions. The instructions include the target working mode and target rotating speed of each FFU. Through wired or wireless network, these instructions are sent to the FFUs in each clean space, and the FFUs will immediately adjust the fan accordingly after receiving the instructions.

[0031] After the dynamic adjustment instructions are issued, the FFU operating parameters, environmental parameters and energy consumption data are continuously collected, and the multi-dimensional prediction model and the energy consumption evaluation model are adaptively optimized based on the continuously collected parameters, so as to ensure the continuous dynamic balance of the multi-zone FFU pressure difference.

[0032] Specifically, after the FFU executes the instruction, new data is continuously collected. Using these new data, the energy consumption evaluation model and the multi-dimensional prediction model are incrementally learned to continuously optimize their prediction and evaluation capabilities. This closed-loop feedback mechanism ensures that the control method can dynamically adapt to long-term changes such as equipment aging and filter clogging, thereby ensuring the continuous efficiency and robustness of the system performance.

[0033] The technical scheme of real-time multi-dimensional data collection, energy consumption evaluation model, and multi-dimensional prediction model combination can accurately predict the trend of pressure difference and passenger flow under the premise of minimizing energy consumption, and actively adjust the operating state of multi-zone FFUs through FFU group collaborative control strategy, significantly improving the dynamic balance of clean space pressure difference, system operation efficiency, and energy saving effect.

[0034] Optionally, according to the historical data, an energy consumption evaluation model is constructed to evaluate the energy consumption state of the FFU in each clean space in real time, including: According to the FFU operating parameters, environmental parameters, and energy consumption data contained in the historical data, an energy consumption evaluation model to be trained is initially trained using a reinforcement learning algorithm, wherein the reinforcement learning algorithm uses the energy consumption data in the historical data as a reward function. Specifically, during the initial training of the energy consumption evaluation model, the FFU operating parameters, environmental parameters, and energy consumption data in the historical period are used as a historical sample set. The model is trained using a reinforcement learning algorithm, and the goal is to learn an optimal strategy to accurately predict the energy consumption of the FFU given the FFU operating parameters and environmental parameters. The historical energy consumption data is used as a reward function in reinforcement learning, and the model adjusts its parameters to maximize this reward, thereby improving the accuracy of energy consumption prediction. For example, when the model's predicted energy consumption is very close to the actual energy consumption, it will receive a high reward, and vice versa.

[0035] The FFU operating parameters and environmental parameters before the current time are input into the energy consumption evaluation model to obtain the current time energy consumption evaluation result of the FFU; Specifically, after the model completes the initial training, the FFU operating parameters and environmental parameters before the current time are input into the energy consumption evaluation model as input features. The energy consumption evaluation model immediately processes these inputs and outputs the current time energy consumption evaluation result of the FFU. Real-time and intelligent evaluation of FFU energy consumption is achieved.

[0036] The current time energy consumption evaluation result is compared with the real-time collected energy consumption data to obtain an evaluation difference, and the parameters of the energy consumption evaluation model are optimized using an error backpropagation algorithm based on the evaluation difference.

[0037] Specifically, the current time energy consumption evaluation result output by the energy consumption evaluation model is compared with the real-time collected actual energy consumption data, and an evaluation difference between the two is calculated. Subsequently, the error backpropagation algorithm is used to transfer the difference as an error signal to the energy consumption evaluation model in a reverse direction, and the internal parameters of the model are adjusted according to the error signal. This continuous closed-loop optimization mechanism ensures that the energy consumption evaluation model can adapt to the changes in energy consumption characteristics caused by long-term factors such as FFU device aging and filter screen clogging, and maintain its evaluation accuracy and robustness. As shown in Figure 2 the overall performance advantages of the present application and the traditional control in differential pressure stability and energy consumption are compared intuitively.

[0038] Optionally, the parameter optimization of the energy consumption evaluation model according to the evaluation difference includes: a dynamic weight is assigned to the evaluation difference according to the evaluation difference, the differential pressure data and the passenger flow data; Specifically, after obtaining the evaluation difference between the FFU energy consumption evaluation result and the actual energy consumption data, a dynamic weight needs to be assigned to the difference. The basis for weight assignment is the current differential pressure data and passenger flow data. When the differential pressure of the clean space deviates from the set value by a large margin or the passenger flow is at a peak period, it indicates that the control environment is in an unstable or high-load state, and the evaluation difference is more important for model optimization at this time, and the assigned dynamic weight is also larger. This ensures that the model can learn and adapt to energy consumption changes in high-priority environments more preferentially.

[0039] According to the evaluation difference assigned with the dynamic weight, the error backpropagation algorithm is used to calculate the gradient value of each parameter in the energy consumption evaluation model, and the parameters of the energy consumption evaluation model are updated based on the gradient value to reduce the evaluation difference.

[0040] Specifically, the evaluation difference assigned with the dynamic weight is taken as the output of the loss function, and the error backpropagation algorithm is used to calculate the gradient value of each parameter in the energy consumption evaluation model. Based on these gradient values, the parameters of the model are updated using optimization algorithms such as gradient descent. Through this process, the model can learn in a targeted manner according to the dynamic weight, more effectively reduce the evaluation difference, and improve the accuracy of energy consumption evaluation and the adaptability of the model to complex working conditions.

[0041] Optionally, the establishment of the multi-dimensional prediction model includes: The FFU operating parameters and environmental parameters contained in the historical data are respectively generated into FFU operating parameter time series and environmental parameter time series in chronological order; Specifically, the FFU operating parameter time series can be represented as , wherein, is the time The vector of the collected FFU wind speed, rotation speed and power parameters. The time series of environmental parameters can be expressed as , wherein, is the vector of the collected differential pressure, particle concentration and passenger flow parameters at time These data vectors are arranged in chronological order to form their respective time series. In order to ensure the accuracy of the model, these time series have undergone data cleaning and preprocessing, such as missing value filling, outlier removal and normalization processing, to eliminate the interference of different dimensions on the operation and ensure the mathematical operation rationality of the data.

[0042] The time series of the FFU operating parameters and the time series of the environmental parameters are fused to obtain a fusion feature vector; Specifically, in order to capture the interaction between the FFU operating state and the environmental state, a multi-modal feature fusion technology is adopted. The fusion process is expressed as: , wherein, is the vector of the collected differential pressure, particle concentration and passenger flow parameters at time The fusion feature vector obtained is, , , is a learnable fusion coefficient, and its specific value is determined through model training. The element-level multiplication is represented by “”. This fusion process combines weighted summation and interaction terms to integrate data information of different modalities into a unified vector, which can more comprehensively describe the dynamic state of the clean space and capture the non-linear correlation between different parameters.

[0043] A deep neural network based on long short-term memory network is adopted, and the fusion feature vector is taken as the input to predict the differential pressure change trend and the passenger flow change trend as the output, thereby training the multi-dimensional prediction model to be trained.

[0044] Specifically, a deep neural network based on long short-term memory network (LSTM) is adopted as the multi-dimensional prediction model. LSTM network is particularly good at processing and predicting time series data, and can capture long-term dependencies in data, such as the influence of historical passenger flow fluctuations on future differential pressure trends. The training process can be expressed as minimizing the loss function: , wherein, and represent the predicted differential pressure and the actual differential pressure of the model at time and represent the predicted passenger flow and the actual passenger flow of the model at time and the predicted human flow and the actual human flow of the model, is the total number of historical data. By minimizing this loss function, the trained model can accurately predict the future differential pressure trend and human flow trend according to the fused feature vector. This training process can use optimization algorithms such as gradient descent, and adjust the hyperparameters through cross-validation to ensure the generalization ability and prediction accuracy of the model under different working conditions.

[0045] Optionally, generating the FFU group collaborative control strategy for the plurality of clean spaces according to the predicted energy consumption data, the differential pressure trend and the human flow trend comprises: presetting a group collaborative control strategy rule library containing FFU working modes, rotating speeds and energy-saving control rules; Specifically, before generating the FFU group collaborative control strategy, a rule library containing multiple control rules needs to be preset. The rules of the rule library can include three categories: FFU working mode rules, FFU operating parameter rules and FFU energy-saving control rules. For example, the working mode rules can define "standby mode", "low-speed running mode" and "full-speed running mode", etc.; the operating parameter rules can define the wind speed adjustment range and the air volume adjustment range of the FFU; the energy-saving control rules can define the control strategy of the FFU under different differential pressure, energy consumption and human flow thresholds. The combination of these rules constitutes the decision space of group collaborative control.

[0046] According to the predicted energy consumption data, the differential pressure trend and the human flow trend, a fuzzy logic algorithm is used to assign a dynamic weight to each rule in the FFU group collaborative control strategy rule library; Specifically, the fuzzy logic algorithm is used to process the fuzzy and uncertain information of the predicted energy consumption data, the differential pressure trend and the human flow trend. First, these predicted data are taken as input variables of the fuzzy algorithm, such as "predicted energy consumption" which can be defined as a fuzzy set of "low", "medium" or "high". "Differential pressure trend" and "human flow trend" are also processed in a similar way. Then, the fuzzy algorithm assigns a dynamic weight to each rule in the rule library according to the preset fuzzy rules. For example, when the predicted energy consumption is "high" and the differential pressure trend is "down", the rule related to "reducing FFU rotating speed" will get a higher weight. This can be represented as a fuzzy inference function: , wherein, is the dynamic weight assigned to the rule, is the predicted energy consumption data, is the differential pressure trend, is the human flow trend, is the fuzzy logic inference function. Through this process, the fuzzy logic algorithm can convert complex prediction information into clear and quantifiable decision weights.

[0047] According to the dynamic weight, the FFU group cooperative control strategy is matched and generated from the FFU group cooperative control strategy rule base, while meeting the requirements of minimum energy consumption and dynamic balance of pressure difference.

[0048] Specifically, after assigning a dynamic weight to each rule in the rule base, according to the size of the weight, the final FFU group cooperative control strategy is matched and generated. Weighted average or maximum membership degree method can be used to select the rule with the highest weight or the combination of multiple rules from the rule base as the final control strategy. The strategy aims to meet the dual goals of minimum energy consumption and dynamic balance of pressure difference. For example, the generated strategy may instruct part of the FFUs with high energy consumption to enter low-speed mode, while instructing another part of the FFUs that contribute greatly to pressure difference stability to maintain the current speed, so as to achieve the cooperative optimization of energy consumption and pressure difference in the entire clean space.

[0049] Optionally, the preset group cooperative control strategy rule base containing FFU working mode, speed and energy-saving control rules comprises: FFU working mode rule set of FFU standby mode, low-speed running mode and full-speed running mode; Specifically, the FFU working mode rule set defines the basic running mode of the FFU in different states. The standby mode rule means that when there is no one in the clean space, the energy consumption is extremely low, and the pressure difference is maintained above the safety threshold, the FFU enters the sleep or low-power state. The low-speed running mode rule means that when energy consumption optimization is the main consideration and the pressure difference changes gently, the FFU runs at a lower speed. The full-speed running mode rule means that in emergency situations such as rapid pressure difference drop, sharp increase in particle concentration or sudden increase in human flow, the FFU runs at the highest speed to quickly restore the pressure difference and cleanliness.

[0050] FFU running parameter rule set of FFU wind speed adjustment range and air volume adjustment range; Specifically, the FFU running parameter rule set specifies the specific adjustment range of the FFU in different modes. The wind speed adjustment range rule defines the allowed wind speed interval of the FFU in low-speed, medium-speed and high-speed modes, such as in low-speed mode, the wind speed adjustment range is to . The air volume adjustment range rule defines the upper and lower limits of air volume adjustment, ensuring that the FFU can provide the minimum air volume required for cleanliness during operation, while not exceeding the maximum air volume designed to carry. The air volume calculation formula is: , wherein, is the air volume, is the cross-sectional area of the air outlet, is the air velocity.

[0051] FFU energy-saving control rule set of FFU differential pressure threshold, energy consumption threshold and human flow threshold.

[0052] Specifically, the FFU energy-saving control rule set defines the condition threshold that triggers FFU mode switching and speed adjustment. The differential pressure threshold rule specifies the minimum differential pressure value and the maximum differential pressure value that maintains the positive pressure of the clean space. When the differential pressure exceeds this range, adjustment is needed. The energy consumption threshold rule sets the upper limit of the expected energy consumption of the FFU, such as When the actual energy consumption exceeds this threshold, the control strategy will preferentially take energy-saving measures. The human flow threshold rule defines the cleanliness requirement under different human flow, such as when the human flow exceeds , it will switch to a more aggressive control mode to deal with the risk of increased pollution.

[0053] Optionally, based on the FFU group cooperative control strategy, the dynamic adjustment instruction is issued to the FFU in each clean space, comprising: The FFU group cooperative control strategy is converted into FFU dynamic adjustment instruction, wherein the FFU dynamic adjustment instruction includes FFU target working mode and target speed; Specifically, after generating the FFU group cooperative control strategy, it needs to be parsed into executable dynamic adjustment instruction. This process converts abstract control strategies, such as "save energy in area A and maintain stable differential pressure in area B", into specific and identifiable FFU operation commands. For example, the instruction may require the FFU in area A to enter "low-speed running mode" and set the speed to 1000 rpm, while the FFU in area B to enter "full-speed running mode" and set the speed to 2000 rpm.

[0054] The FFU dynamic adjustment instruction is issued to the FFU in each clean space, and during the issuance process, instructions with higher energy consumption optimization weight and higher differential pressure balance weight are preferentially executed.

[0055] Specifically, when issuing dynamic adjustment instructions to FFUs in multiple clean spaces, each instruction will be sorted according to its priority. The priority of the instruction is determined by its energy consumption optimization weight and differential pressure balance weight. For example, an instruction for repairing energy consumption anomalies or an instruction for restoring core area differential pressure will be given a higher priority. This priority mechanism ensures that in network congestion or instruction concurrency, the instructions that are most critical to the overall clean environment stability and energy consumption optimization can be executed by the FFU first, thereby ensuring response speed and reliability. For example Figure 3As shown, the priority scheduling mechanism of the instructions is demonstrated.

[0056] Optionally, in the issuing process, the instructions with higher energy consumption optimization weight and higher pressure difference balance weight are preferentially executed. According to the real-time collected energy consumption data and the predicted energy consumption data, the energy consumption abnormality degree of each FFU in the FFU group cooperative control strategy is evaluated, and the energy consumption optimization weight of each FFU is allocated according to the energy consumption abnormality degree. Specifically, the allocation of the energy consumption optimization weight is based on the energy consumption abnormality degree of the FFU. The calculation formula of the energy consumption abnormality degree is as follows: , Among them, is the energy consumption abnormality degree of the FFU, is the real-time collected energy consumption data, is the energy consumption data predicted by the energy consumption evaluation model. When the energy consumption abnormality degree of the FFU exceeds the preset threshold, the energy consumption optimization weight of the FFU will be increased accordingly. This ensures that when the energy consumption is abnormal, the control instructions of these high abnormal FFUs can be preferentially processed to achieve the purpose of rapid energy saving.

[0057] According to the real-time collected pressure difference data and the pressure difference change trend, the contribution degree of each FFU to the pressure difference stability in the clean space is evaluated, and the pressure difference balance weight of each FFU is allocated according to the contribution degree. Specifically, the allocation of the pressure difference balance weight is based on the contribution degree of the FFU to the pressure difference stability. The evaluation formula of the contribution degree is as follows: , Among them, is the contribution degree of the FFU to the pressure difference stability, is the real-time collected pressure difference data, is the slope of the pressure difference change trend, indicating the degree of change of the pressure difference. When the pressure difference of the clean space where the FFU is located changes dramatically and the adjustment of the FFU can significantly improve this change, the pressure difference balance weight of the FFU will be increased accordingly. This ensures that when the pressure difference fluctuates, the FFU that contributes most to the pressure difference stability can be preferentially executed to quickly restore the pressure difference balance.

[0058] According to the energy consumption optimization weight and the pressure difference balance weight, a weighted summation algorithm is used to calculate the comprehensive priority score of the FFU dynamic adjustment instruction, and the dynamic adjustment instruction is executed in descending order according to the comprehensive priority score.

[0059] Specifically, the comprehensive priority score formula of the FFU dynamic adjustment instruction is as follows: , Among them, is a comprehensive priority score, and is a preset global weight coefficient for adjusting the relative importance of energy consumption optimization and differential pressure balance in decision-making. As in the energy consumption optimization-oriented working condition, may be set to be higher than After obtaining the comprehensive priority score, the dynamic adjustment instruction is issued to the corresponding FFU in order from high to low score, ensuring that critical instructions can be executed in priority.

[0060] Optionally, the adaptive optimization of the multi-dimensional prediction model and the energy consumption evaluation model based on the continuously collected parameters ensures the continuous dynamic balance of the multi-zone FFU differential pressure, comprising: The continuously collected FFU operating parameters, environmental parameters and energy consumption data are used as an incremental data set for incremental learning of the multi-dimensional prediction model and the energy consumption evaluation model, and the model parameters are updated; Specifically, after the FFU executes the instruction, new data is continuously collected. These new data will form an incremental data set and be used for incremental learning of the multi-dimensional prediction model and the energy consumption evaluation model. Unlike traditional batch learning, incremental learning does not require retraining of the entire model, but fine-tunes the model parameters using the new data set, so that it can dynamically adapt to changes in long-term working conditions such as equipment aging and filter clogging. For example, when the filter of the FFU is clogged, causing increased wind resistance, its operating parameters will change. Through incremental learning, the energy consumption evaluation model can learn this new energy consumption pattern and update its parameters accordingly.

[0061] The performance of the updated model is evaluated, and if the performance evaluation result is better than the current model, the model updating is completed; if the performance evaluation result is not better than the current model, the model is restored to the state before updating.

[0062] Specifically, after the model parameters are updated, the performance of the updated model is evaluated immediately, and the error rate, accuracy and other indicators on the validation data set are used to judge the performance of the new model. If the evaluation result shows that the performance of the updated model is better than the current model, the new model will be formally deployed. Otherwise, if the performance of the new model is not good, in order to prevent unstable models from affecting the control accuracy, the model will be rolled back to the state before updating. This adaptive optimization process with performance verification and rollback mechanism ensures the continuous stability and reliability of the system.

[0063] Based on the same inventive concept, as shown in Figure 4 The present application also provides a control system for multi-zone FFU differential pressure collaborative dynamic balance, comprising: A data acquisition module is used to collect FFU operating parameters, environmental parameters and energy consumption data in real time in a plurality of clean spaces, and to obtain historical data;​ an energy consumption evaluation module, configured to construct an energy consumption evaluation model according to the historical data, and evaluate the energy consumption state of the FFU in each clean space in real time; a multi-dimensional prediction module, configured to establish a multi-dimensional prediction model according to the FFU operation parameters, the environmental parameters and the historical data, and generate a predicted pressure difference change trend and a predicted human flow change trend in each clean space in a future period of time; a cooperative control strategy module, configured to generate a FFU group cooperative control strategy for multiple clean spaces according to the predicted energy consumption data, the pressure difference change trend and the human flow change trend; an instruction issuing module, configured to issue a dynamic adjustment instruction to the FFU in each clean space based on the FFU group cooperative control strategy; an adaptive optimization module, configured to continuously collect the FFU operation parameters, the environmental parameters and the energy consumption data after the dynamic adjustment instruction is issued, and perform adaptive optimization on the multi-dimensional prediction model and the energy consumption evaluation model based on the continuously collected parameters.

[0064] Embodiment 1: To verify the feasibility of the present application, the present application is applied to the clean room environment of a large-scale semiconductor manufacturing plant. The plant has multiple interconnected clean areas, and has extremely strict requirements on internal pressure difference, particle concentration and other environmental parameters. The traditional FFU control method is difficult to ensure stable pressure difference while taking into account high operating energy consumption. The plant hopes to use the method of the present application to achieve intelligent cooperative control of the FFU group in multiple areas, and achieve the dual goals of dynamic balance of pressure difference and optimal operating energy consumption.

[0065] In this embodiment, the semiconductor plant deploys the control method of the present application in the A zone, B zone and C zone, three core production areas. The control method collects FFU operation parameters, environmental parameters and energy consumption data through sensors deployed in each area. Using the historical data of the past 6 months, the energy consumption evaluation model is initially trained offline using the gradient boosting decision tree algorithm. At the same time, a multi-dimensional prediction model is constructed using a long short-term memory network to predict the pressure difference and human flow change trend in the next 15 minutes.

[0066] In the continuous operation test in April, it demonstrated its dynamic optimization and forward-looking control capabilities. At 9:00 am on April 15, during the production shift handover, the multi-dimensional prediction model successfully predicted that the C area human flow would surge from 5 to 25 people in the next 10 minutes based on historical patterns, and predicted that the pressure difference would have a rapid downward trend of -2.5 Pa. At the same time, the energy consumption evaluation model combined with the real-time parameters of the FFU output the real-time energy consumption evaluation result of the C area FFU as 4.5 kW, while the real-time energy consumption collected by the power monitoring device was 4.7 kW, resulting in an energy consumption difference of 0.2 kW. Since both the human flow and the pressure difference trend pointed to high risk, the FFU instruction assigned a higher dynamic weight, with a weight of , accelerating the online parameter fine-tuning of the energy consumption evaluation model to make it more accurate in reflecting the current working conditions.

[0067] Based on the above prediction information, the fuzzy logic decision algorithm assigned dynamic weights to the rules in the collaborative control strategy rule base. At this time, the rule of "full-speed running mode" aiming to "rapidly increase the wind speed to stabilize the pressure difference" obtained the highest weight. Then the collaborative control strategy was generated and converted into specific FFU dynamic adjustment instructions. A "target working mode: full-speed running, target speed: 1800 RPM" instruction for the C area obtained a comprehensive priority score of 0.95 because it had the highest contribution to maintaining pressure difference stability, and was prioritized for issuance. At the same time, the A and B areas were stable, and their FFU energy-saving adjustment instructions had priority scores of only 0.2-0.3, so they were temporarily suspended. After the instructions were issued, the pressure difference in the C area fluctuated by less than ±0.5 Pa during the peak period, far better than the ±3.0 Pa fluctuation range under traditional control, effectively avoiding the risk of cross-contamination due to pressure difference imbalance.

[0068] In the one-month test, the invention reduced the overall energy consumption by about 18% compared to the control group, while maintaining all area pressure difference qualified rate above 99.8%. The prediction accuracy of air leakage reached more than 92%, and through optimized scheduling, the average response time and intervention time for high-risk events were shortened from 2 minutes under traditional control to less than 10 seconds.

[0069] Table 1 Multi-dimensional prediction model prediction accuracy data table Table 2 Comparison of system performance under different control strategies Table 3 FFU dynamic adjustment instruction priority example The above Tables 1 to 3 show the actual application data of the invention in a semiconductor factory cleanroom, which details the superior performance in prediction accuracy, collaborative control effect, and intelligent scheduling.

[0070] The data in Table 1 shows that the multi-dimensional prediction model has a high prediction accuracy for the trend of human flow and pressure difference, such as in the peak period of human flow in area C, the prediction accuracy is more than 90%, which provides a reliable basis for taking forward-looking intervention measures.

[0071] The data in Table 2 clearly shows the great advantage of the present application in energy consumption. After using the present application, the average fluctuation of pressure difference is significantly reduced from ±3.2Pa to ±0.8Pa, the qualified rate is increased to 99.8%, and more than 18% of energy saving effect is achieved, successfully solving the contradiction between pressure difference stability and energy saving.

[0072] Table 3 directly shows the operation process of the instruction priority scheduling mechanism. When facing the urgent risk of pressure difference imbalance in area C, the priority score of the corresponding instruction is calculated as high as 0.95 and is immediately executed, while the non-urgent energy-saving instructions of other areas are put on the waiting queue, ensuring that the control resources are preferentially used to deal with the most critical problem, greatly improving the response speed and robustness.

[0073] It should be noted that the electrical connection between the above-mentioned units does not necessarily represent the direct connection of the line, and the indirect connection mode can also be applied to the embodiments of the present application as long as the purpose of the present application is achieved. The above-mentioned is only an exemplary embodiment of the present application, and cannot limit the scope of the present application.

[0074] That is, any equivalent changes and modifications made according to the teachings of the present application are still within the scope of the present application. Other embodiments of the present application will be readily apparent to those skilled in the art upon considering the specification and practice of the true principles disclosed herein. The present application is intended to cover any variations, uses or adaptive changes to the present application following the general principles of the present application and including commonly known or conventional technical means in the art not disclosed by the present application.

Claims

1. A control method for coordinated dynamic balance of pressure differentials in multi-region FFUs, characterized in that, The method includes: Real-time acquisition of FFU operating parameters, environmental parameters, and energy consumption data in multiple clean spaces, and acquisition of corresponding historical data in multiple clean spaces, wherein the environmental parameters include differential pressure data, particle concentration data, and personnel flow data; Based on the historical data, an energy consumption assessment model is constructed to assess the energy consumption status of each FFU in the clean space in real time, and the model is compared with the real-time collected energy consumption data to optimize the parameters of the energy consumption assessment model. The real-time collected FFU operating parameters and environmental parameters are input into the optimized energy consumption assessment model to obtain the predicted energy consumption data for each clean space in the future. Based on the FFU operating parameters, environmental parameters and historical data, a multidimensional prediction model is established to generate the predicted pressure difference change trend and the flow of people in each clean space over a future period of time. Based on the predicted energy consumption data, pressure difference change trend, and personnel flow change trend, an FFU group collaborative control strategy is generated for multiple clean spaces. The FFU group collaborative control strategy is used to guide the working mode and rotation speed of FFUs in multiple clean spaces. Based on the FFU group collaborative control strategy, dynamic adjustment instructions are issued to each FFU in the clean space; After the dynamic adjustment command is issued, FFU operating parameters, environmental parameters and energy consumption data are continuously collected, and the multidimensional prediction model and energy consumption assessment model are adaptively optimized based on the continuously collected parameters to ensure the continuous dynamic balance of FFU pressure difference in multiple regions.

2. The control method for coordinated dynamic balance of multi-region FFU pressure differential as described in claim 1, characterized in that, Based on the historical data, an energy consumption assessment model is constructed to evaluate the energy consumption status of each FFU in the clean space in real time, including: Based on the FFU operating parameters, environmental parameters and energy consumption data contained in the historical data, a reinforcement learning algorithm is used to initially train the energy consumption evaluation model to be trained, wherein the reinforcement learning algorithm uses the energy consumption data in the historical data as the reward function. Input the FFU operating parameters and environmental parameters from the current moment to the energy consumption assessment model to obtain the current energy consumption assessment result of the FFU. The energy consumption assessment result at the current moment is compared with the energy consumption data collected in real time to obtain the assessment difference. Based on the assessment difference, the parameters of the energy consumption assessment model are optimized using the error backpropagation algorithm.

3. The control method for coordinated dynamic balance of multi-region FFU pressure differential as described in claim 2, characterized in that, The optimization of the energy consumption assessment model parameters using the backpropagation algorithm based on the assessment difference includes: Based on the assessment difference, pressure difference data, and pedestrian flow data, dynamic weights are assigned to the assessment difference; Based on the evaluation difference with dynamically assigned weights, the backpropagation algorithm is used to calculate the gradient value of each parameter in the energy consumption evaluation model, and the parameters of the energy consumption evaluation model are updated based on the gradient value to reduce the evaluation difference.

4. The control method for coordinated dynamic balance of multi-region FFU pressure differential as described in claim 1, characterized in that, The establishment of the multidimensional prediction model includes: The FFU operating parameters and environmental parameters contained in the historical data are used to generate FFU operating parameter time series and environmental parameter time series respectively in chronological order; The time series sequences of FFU operating parameters and environmental parameters are fused using multimodal features to obtain a fused feature vector; A deep neural network based on long short-term memory is used, with the fused feature vector as input, and the predicted pressure difference change trend and traffic flow change trend as output, to train the multidimensional prediction model to be trained.

5. The control method for coordinated dynamic balance of multi-region FFU pressure differential as described in claim 1, characterized in that, Based on the predicted energy consumption data, pressure difference change trends, and personnel flow change trends, a collaborative control strategy for FFU groups is generated for multiple clean spaces, including: A pre-defined group collaborative control strategy rule base includes FFU operating modes, speed, and energy-saving control rules; Based on the predicted energy consumption data, pressure difference change trend and population flow change trend, a fuzzy logic algorithm is used to assign dynamic weights to each rule in the FFU group collaborative control strategy rule base. Based on the dynamic weights, an FFU group collaborative control strategy is matched and generated from the FFU group collaborative control strategy rule base, which simultaneously meets the requirements of minimizing energy consumption and dynamic pressure balance.

6. The control method for coordinated dynamic balance of multi-region FFU pressure differential as described in claim 5, characterized in that, The preset group collaborative control strategy rule base, which includes FFU operating modes, speed, and energy-saving control rules, includes: FFU operating mode rule set for FFU standby mode, low-speed operation mode and full-speed operation mode; FFU operating parameter rule set for FFU wind speed adjustment range and air volume adjustment range; FFU energy-saving control rule set based on FFU differential pressure threshold, energy consumption threshold, and passenger flow threshold.

7. The control method for coordinated dynamic balance of multi-region FFU pressure differential as described in claim 1, characterized in that, Based on the aforementioned FFU group collaborative control strategy, dynamic adjustment instructions are issued to each FFU in the clean space, including: The FFU group collaborative control strategy is transformed into FFU dynamic adjustment instructions, wherein the FFU dynamic adjustment instructions include the FFU target operating mode and target speed; The FFU dynamic adjustment command is issued to each FFU in the clean space, and during the issuance process, the command with higher energy consumption optimization weight and higher pressure differential balance weight is executed first.

8. The control method for coordinated dynamic balance of multi-region FFU pressure differential as described in claim 7, characterized in that, During the distribution process, instructions with higher energy consumption optimization weight and higher pressure differential balance weight are prioritized for execution, including: Based on the real-time collected energy consumption data and predicted energy consumption data, the degree of energy consumption anomaly of each FFU in the FFU group collaborative control strategy is evaluated, and energy consumption optimization weights are assigned to each FFU according to the degree of energy consumption anomaly. Based on the real-time collected differential pressure data and differential pressure change trends, the contribution of each FFU in the clean space to differential pressure stability is evaluated, and a differential pressure balance weight is assigned to each FFU according to the contribution. Based on the energy consumption optimization weight and pressure difference balance weight, a weighted summation algorithm is used to calculate the comprehensive priority score of the FFU dynamic adjustment command, and the dynamic adjustment command is executed in descending order of the comprehensive priority score.

9. The control method for coordinated dynamic balance of multi-region FFU pressure differential as described in claim 1, characterized in that, The adaptive optimization of the multidimensional prediction model and energy consumption assessment model based on continuously collected parameters to ensure the continuous dynamic balance of FFU pressure differentials in multiple regions includes: The continuously collected FFU operating parameters, environmental parameters, and energy consumption data are used as incremental datasets to incrementally learn the multidimensional prediction model and energy consumption assessment model, and update the model parameters. The updated model is evaluated for performance. If the performance evaluation result is better than the current model, the model update is completed; if the performance evaluation result is not better than the current model, the model is restored to its state before the update.

10. A control system for coordinated dynamic balancing of multi-region FFU differential pressure is applied to the control method for coordinated dynamic balancing of multi-region FFU differential pressure as described in any one of claims 1-9, characterized in that, The system includes: The data acquisition module is used to collect FFU operating parameters, environmental parameters, and energy consumption data in multiple clean spaces in real time, and to acquire historical data; The energy consumption assessment module is used to construct an energy consumption assessment model based on the historical data and to assess the energy consumption status of each FFU in the clean space in real time. The multidimensional prediction module is used to establish a multidimensional prediction model based on the FFU operating parameters, environmental parameters and historical data, and generate the predicted pressure difference change trend and personnel flow change trend in each clean space over a future period of time. The collaborative control strategy module is used to generate FFU group collaborative control strategies for multiple clean spaces based on the predicted energy consumption data, pressure difference change trends and people flow change trends. The instruction issuing module is used to issue dynamic adjustment instructions to each FFU in the clean space based on the FFU group collaborative control strategy; The adaptive optimization module is used to continuously collect FFU operating parameters, environmental parameters and energy consumption data after the dynamic adjustment command is issued, and to adaptively optimize the multidimensional prediction model and energy consumption assessment model based on the continuously collected parameters.

Images