Wind speed control method, device and equipment of junction station and medium

Abstract

The invention discloses a wind speed control method, device and equipment of a junction station and a medium, and relates to the technical field of wind tunnel experiments, and the junction station is provided with a main and auxiliary dual-channel structure, a main and auxiliary dual-channel adjusting valve, and a sensor used for collecting environmental parameters. Comprising the steps that the initial main channel opening degree of the junction station is determined according to the expected wind speed of the junction station and the mapping relation between the main channel opening degree and the wind speed; predicting a rough main channel opening obtained by searching based on the initial main channel opening and the current environment parameters by using the target prediction model to obtain a target main channel opening interval; searching in the target main channel opening interval to obtain a target main and auxiliary dual-channel opening combination; the current main-auxiliary double-channel adjusting valve is adjusted to control the air speed; the target prediction model is a model fusing a convolutional neural network, a bidirectional long-short-term memory network and an attention mechanism, and training data of the target prediction model are main and auxiliary dual-channel opening combination and flow field dynamic data of the junction station. And the wind speed of the junction station is accurately controlled at low cost in various environments.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wind tunnel experiment, in particular to a wind speed control method and device of a flow concentrator, equipment and a medium. BACKGROUND

[0002] In a modern wind tunnel experiment system, accurate control of airflow parameters is a key prerequisite for achieving high-quality aerodynamic testing. However, the complex nonlinear characteristics of airflow and its coupling relationship with the flow concentrator make it difficult for traditional theoretical calculation methods to accurately depict this physical process.

[0003] The method commonly used in current engineering applications is to obtain the flow concentrator channel characteristic curve through a large number of wind tunnel experiment data, and then to approximately describe the nonlinear mapping relationship between the main and auxiliary channel openings and the airflow parameters, but there are many limitations, a large amount of repeated wind tunnel calibration tests are required, the workload is heavy, the air source consumption is huge, and the calibration period is long. More importantly, the traditional channel characteristic curve is only applicable to specific calibration conditions and lacks self-adaptive ability in varying environmental conditions, making it difficult to flexibly respond to control requirements of different flow field parameters and aerodynamic conditions.

[0004] From the above, it can be seen that how to accurately and low-cost control the wind speed of the flow concentrator under different environments is a problem to be solved in the field. SUMMARY

[0005] Therefore, the purpose of the present application is to provide a wind speed control method, device, equipment and medium of a flow concentrator, which can accurately and low-cost control the wind speed of the flow concentrator under different environments. The specific scheme is as follows: In a first aspect, the present application discloses a wind speed control method of a flow concentrator, the flow concentrator having a main and auxiliary dual-channel structure and comprising main and auxiliary dual-channel regulating valves and sensors for collecting various environmental parameters; the method comprises: determining an initial main channel opening degree of the flow concentrator according to an expected wind speed of the flow concentrator and a target mapping relationship; wherein the target mapping relationship is a mapping relationship between each historical main channel opening degree of the flow concentrator and the wind speed of the flow concentrator; performing wind speed prediction based on the initial main channel opening degree, each rough main channel opening degree obtained by first searching, and the current environmental parameters by using a target prediction model, to obtain a target main channel opening degree interval; performing second searching from each main and auxiliary dual-channel opening degree combination under the target main channel opening degree interval by using the target prediction model, to obtain a target main and auxiliary dual-channel opening degree combination corresponding to the expected wind speed; controlling the ground jet test device to adjust the main and auxiliary dual-channel regulating valves based on the target main and auxiliary dual-channel opening degree combination, to control the wind speed of the flow concentrator to correspond to the expected wind speed; The target prediction model is a deep learning model fusing a convolutional neural network, a bidirectional long short-term memory network, and an attention mechanism, and training data of the target prediction model is a historical main-auxiliary dual-channel opening combination and historical flow field dynamic data of the collector.

[0006] Optionally, the target prediction model is acquired, including: The historical main-auxiliary dual-channel opening combination and the historical flow field dynamic data of the collector are determined as training data, and the historical flow field dynamic data includes historical environmental parameters and a historical wind speed, and the environmental parameters include temperature, air pressure, and air density of the collector. An initial prediction model including a convolutional neural network, a bidirectional long short-term memory network, and an attention mechanism is established, and the initial prediction model is iteratively trained using the training data to obtain the target prediction model.

[0007] Optionally, the iteratively training the initial prediction model using the training data to obtain the target prediction model includes: The initial prediction model is determined as a current prediction model; The current prediction model is trained using the training data to obtain a prediction result in a current round, and the prediction result includes a wind speed prediction result and a differential pressure coefficient prediction result; A first mean square error of the wind speed prediction result and a second mean square error of the differential pressure coefficient prediction result are determined using a multi-task loss function, and a weighted mean square error of the first mean square error and the second mean square error is acquired; Parameters of the current prediction model are tuned using the weighted mean square error to obtain a next prediction model; The next prediction model is determined as a new current prediction model, and the training of the current prediction model using the training data to obtain the prediction result in the current round is re-executed until the target prediction model meeting a preset training stop condition is obtained.

[0008] Optionally, the tuning of the parameters of the current prediction model using the weighted mean square error to obtain the next prediction model includes: The parameters of the current prediction model are tuned using the weighted mean square error and an Adam optimizer under a gradient clipping mechanism, and a learning rate of the current prediction model is dynamically adjusted using a scheduler to obtain the next prediction model.

[0009] Optionally, the training of the current prediction model using the training data to obtain the prediction result in the current round includes: The training data is input into the current prediction model; The current prediction model is used to extract a feature vector of the training data by using the convolutional neural network, to obtain a feature sequence containing time sequence dependency based on the feature vector by using the bidirectional long short-term memory network, to obtain a context vector from the feature sequence based on the attention mechanism, and to generate a prediction result in a current round according to a shared feature representation extracted from the context vector.

[0010] Optionally, the wind speed prediction is performed by using a target prediction model, each rough main channel opening degree obtained by the first search based on the initial main channel opening degree, and the current environmental parameters, to obtain a target main channel opening degree interval. The initial main channel opening degree is taken as a direction cutting point, and a first search is performed in a high opening degree direction and a low opening degree direction based on a first preset step size and a preset opening degree constraint condition, to obtain each rough main channel opening degree. The wind speed prediction is performed by using a target prediction model, each rough main channel opening degree, and the current environmental parameters, to obtain a target main channel opening degree interval corresponding to the prediction result and the expected wind speed.

[0011] Optionally, the second search is performed from each main-aided dual-channel opening degree combination under the target main channel opening degree interval by using the target prediction model, to obtain a target main-aided dual-channel opening degree combination corresponding to the expected wind speed. Each main-aided dual-channel opening degree combination under the target main channel opening degree interval is obtained based on a second preset step size; the second preset step size is smaller than the first preset step size. The wind speed prediction result under each main-aided dual-channel opening degree combination is obtained by using the target prediction model, and a wind speed error between each wind speed prediction result and the expected wind speed is determined. The second search is performed in each main-aided dual-channel opening degree combination, to obtain a target main-aided dual-channel opening degree combination with the minimum wind speed error.

[0012] In a second aspect, the present application discloses a wind speed control device of a current collector, the current collector having a main-aided dual-channel structure and containing a main-aided dual-channel regulating valve and a sensor for collecting each environmental parameter; the device comprises: An initial opening degree determination module is configured to determine an initial main channel opening degree of the current collector according to an expected wind speed of the current collector and a target mapping relationship; the target mapping relationship is a mapping relationship between each historical main channel opening degree of the current collector and a wind speed of the current collector. An opening degree interval obtaining module is configured to perform wind speed prediction by using a target prediction model, each rough main channel opening degree obtained by a first search based on the initial main channel opening degree, and current environmental parameters, to obtain a target main channel opening degree interval. An opening combination screening module is configured to perform a second search from each main-aid double-channel opening combination under the target main passage opening interval by using the target prediction model to obtain a target main-aid double-channel opening combination corresponding to the expected wind speed; A wind speed control module is configured to control the ground jet test device to adjust the main-aid double-channel regulating valve based on the target main-aid double-channel opening combination, so that the wind speed of the current flow collector corresponds to the expected wind speed. The target prediction model is a deep learning model integrating a convolutional neural network, a bidirectional long short-term memory network, and an attention mechanism, and the training data of the target prediction model is historical main-aid double-channel opening combination and historical flow field dynamic data of the flow collector.

[0013] In a third aspect, the present application discloses an electronic device, comprising: A memory is configured to save a computer program.

[0014] A processor is configured to execute the computer program to implement the steps of the wind speed control method of the flow collector disclosed above.

[0015] In a fourth aspect, the present application discloses a computer readable storage medium configured to store a computer program; wherein the computer program is executed by a processor to implement the steps of the wind speed control method of the flow collector disclosed above.

[0016] The application has the following beneficial effects: the current collector has a main-aid dual-channel structure and comprises a main-aid dual-channel regulating valve and sensors for collecting various environmental parameters; the method comprises: determining an initial main-channel opening degree of the current collector according to an expected wind speed of the current collector and a target mapping relationship; wherein the target mapping relationship is a mapping relationship between various historical main-channel opening degrees of the current collector and wind speeds of the current collector; performing wind speed prediction based on the initial main-channel opening degree, various rough main-channel opening degrees obtained through first searching, and current environmental parameters by using a target prediction model to obtain a target main-channel opening degree interval; performing second searching from various main-aid dual-channel opening degree combinations under the target main-channel opening degree interval by using the target prediction model to obtain a target main-aid dual-channel opening degree combination corresponding to the expected wind speed; and controlling a ground jet test device to adjust the current main-aid dual-channel regulating valve based on the target main-aid dual-channel opening degree combination to control the wind speed of the current collector to correspond to the expected wind speed; wherein the target prediction model is a deep learning model that fuses a convolutional neural network, a bidirectional long short-term memory network, and an attention mechanism, and the training data of the target prediction model is historical main-aid dual-channel opening degree combinations and historical flow field dynamic data of the current collector. As can be seen, the current collector adopts a main-aid dual-channel structure and a main-aid dual-channel regulating valve, can control the macroscopic wind speed interval through the main channel and finely adjust through the auxiliary channel, and can combine the sensors for collecting environmental parameters to sense data in real time, so as to more accurately control the wind speed compared with a single channel or simple mechanical control. The application determines the initial main-channel opening degree according to the expected wind speed and the target mapping relationship between the historical main-channel opening degrees and the wind speed, which is more efficient than the existing method that relies on empirical calibration points. The target prediction model is used to predict the wind speed based on the initial main-channel opening degree, the rough main-channel opening degrees obtained through rough searching, and the environmental parameters to determine the main-channel opening degree interval, and the main-aid dual-channel opening degree combination corresponding to the expected wind speed is selected from the interval to control the regulating valve. The model has stronger nonlinear feature extraction and key parameter focusing capabilities due to the training data comprising the historical main-aid dual-channel opening degree combinations and the flow field dynamic data, and the fusion of the convolutional neural network, the bidirectional long short-term memory network, and the attention mechanism. The model also has adaptability to environmental parameter changes, can more accurately predict the wind speed compared with traditional models, realizes the corresponding control of the current collector wind speed and the expected wind speed, and improves the control precision, efficiency, and flexibility. BRIEF DESCRIPTION OF DRAWINGS

[0017] 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 prior art description. Obviously, the drawings in the following description are only embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor based on the provided drawings.

[0018] Figure 1 A wind speed control flowchart of a flow concentrator disclosed in the present application; Figure 2 A specific prediction model structure diagram disclosed in the present application; Figure 3 A specific target primary and secondary dual-channel opening combination acquisition flowchart disclosed in the present application; Figure 4 A specific flow concentrator wind speed control diagram disclosed in the present application; Figure 5 A specific overall loss diagram disclosed in the present application; Figure 6 A specific wind speed loss diagram disclosed in the present application; Figure 7 A specific differential pressure coefficient loss diagram disclosed in the present application; Figure 8 A specific wind speed curve diagram under 7°C disclosed in the present application; Figure 9 A specific wind speed curve diagram under 9°C disclosed in the present application; Figure 10 A specific wind speed curve diagram under 11°C disclosed in the present application; Figure 11 A specific differential pressure coefficient curve diagram under 7°C disclosed in the present application; Figure 12 A specific differential pressure coefficient curve diagram under 9°C disclosed in the present application; Figure 13 A specific differential pressure coefficient curve diagram under 11°C disclosed in the present application; Figure 14 A specific channel parameter search path diagram disclosed in the present application; Figure 15 A specific wind speed change diagram in a search process disclosed in the present application; Figure 16 A specific search error change diagram disclosed in the present application; Figure 17 A specific wind speed heat map disclosed in the present application; Figure 18 A wind speed control device structure diagram of a flow concentrator disclosed in the present application; Figure 19 An electronic device structure diagram disclosed in the present application. DETAILED DESCRIPTION

[0019] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by a person of ordinary skill in the art without creative effort belong to the scope of protection of the present application.

[0020] In a modern wind tunnel test system, accurate control of airflow parameters is a key prerequisite for achieving high-quality aerodynamic testing. However, the complex nonlinear characteristics of airflow and the coupling relationship between the airflow and the flow collector make it difficult for traditional theoretical calculation methods to accurately depict this physical process.

[0021] The method commonly used in current engineering applications is to obtain the flow collector channel characteristic curve through a large amount of wind tunnel test data, and then to approximately describe the nonlinear mapping relationship between the main and auxiliary channel openings and the airflow parameters, but there are many limitations, a large amount of repeated wind tunnel calibration tests are required, the workload is heavy, the air source consumption is huge, and the calibration period is long. More importantly, the traditional channel characteristic curve is only applicable to a specific calibration condition and lacks self-adaptive ability in a variable environment, and it is difficult to flexibly respond to control requirements of different flow field parameters and aerodynamic conditions.

[0022] Therefore, the present application correspondingly provides a wind speed control scheme for a flow collector, which can accurately and low-cost control the wind speed of the flow collector in different environments.

[0023] Referring to Figure 1 The embodiment of the present application discloses a wind speed control method for a flow collector, the flow collector has a main and auxiliary dual-channel structure and includes main and auxiliary dual-channel regulating valves and sensors for collecting various environmental parameters; the method comprises: Step S11: determining the initial main channel opening of the flow collector according to the expected wind speed of the flow collector and the target mapping relationship; wherein the target mapping relationship is the mapping relationship between each historical main channel opening of the flow collector and the wind speed of the flow collector.

[0024] The flow collector has a main and auxiliary dual-channel structure, and the flow collector includes main and auxiliary dual-channel regulating valves and sensors for collecting various environmental parameters, that is, the flow collector in the embodiment is a coaxial intelligent flow collector. The coaxial intelligent flow collector realizes full-speed domain wind speed regulation through its main and auxiliary dual-channel structure, and can realize control of the wind speed range of 0-200 m / s of the ground jet test device by coordinating the output pressure of the front end pressure reducing valve and the opening degree of the main channel regulating valve of the coaxial intelligent flow collector. Accurate control relies on the cooperative work of the main channel regulating valve and the auxiliary channel regulating valve.

[0025] In modern wind tunnel test systems, accurate control of airflow parameters is a key prerequisite for achieving high-quality aerodynamic testing. However, the complex nonlinear characteristics of airflow and its coupling relationship with the coaxial intelligent flow collector make it difficult for traditional theoretical calculation methods to accurately depict this physical process. The method commonly used in current engineering applications is to obtain the channel characteristic curve through a large number of wind tunnel test data and use iterative correction to approximate the nonlinear mapping relationship between the main and auxiliary channel openings and the airflow parameters. This method relies on empirical solutions and has many limitations, including the need for a large amount of repetitive wind tunnel calibration tests, heavy workload, huge air source consumption, long calibration period, and the lack of self-adaptive ability in varying environmental conditions, making it difficult to flexibly respond to control requirements of different flow field parameters and aerodynamic conditions.

[0026] In this embodiment, the channel calibration module establishes the target mapping relationship between the wind speed and the main channel opening of the coaxial intelligent flow collector. Specifically, the target mapping relationship between each historical main channel opening and the corresponding wind speed of the flow collector at nine key calibration points can be established, such as the target mapping relationship between historical main channel opening O1 and wind speed S1, the target mapping relationship between historical main channel opening O2 and wind speed S2,..., and the target mapping relationship between historical main channel opening O9 and wind speed S9.

[0027] After determining the expected wind speed of the flow collector, the channel calibration module determines the initial main channel opening of the flow collector based on the expected wind speed of the flow collector, the target mapping relationship, and the extrapolation algorithm. Specifically, after obtaining the expected wind speed of the flow collector, if the expected wind speed is within the range of the existing calibration point wind speed, the corresponding opening can be estimated using the mapping relationship of the adjacent calibration points through the interpolation algorithm; if the expected wind speed exceeds the calibration point range, the extrapolation algorithm is used, such as calculating the change rate of the boundary calibration point based on a linear relationship, combining the change rate with the difference between the expected wind speed and the boundary calibration point wind speed, and calculating the opening of the part exceeding the range, thereby obtaining the initial main channel opening of the flow collector, providing a basic parameter for subsequent control adjustment.

[0028] Step S12: Perform wind speed prediction using the target prediction model, each rough main channel opening obtained based on the initial main channel opening, and the current environmental parameters to obtain a target main channel opening interval. The target prediction model is a deep learning model that integrates convolutional neural networks, bidirectional long short-term memory networks, and attention mechanisms. The training data for the target prediction model is the historical main and auxiliary double-channel opening combinations and historical flow field dynamic data of the flow collector.

[0029] In this embodiment, the target prediction model is obtained, including: determining the collected historical main and auxiliary dual-channel opening combination of the collector and historical flow field dynamic data as training data; wherein the historical flow field dynamic data includes historical environmental parameters and historical wind speed, and the environmental parameters include temperature, air pressure and air density of the collector; an initial prediction model containing a convolutional neural network, a bidirectional long short-term memory network and an attention mechanism is established, and the initial prediction model is iteratively trained using the training data to obtain the target prediction model.

[0030] For example Figure 2 A specific prediction model structure diagram is shown, the historical main and auxiliary dual-channel opening combination of the collector and the historical flow field dynamic data are collected, the historical flow field dynamic data includes historical environmental parameters and historical wind speed, and the environmental parameters include temperature, air pressure and air density of the collector, wherein the air pressure is the difference between the dynamic pressure and the static pressure, that is, the historical main channel opening (Main valve opening), the historical auxiliary channel opening (Needle valve opening), the temperature (Temperature) of the collector, the air density (Air density), the dynamic pressure (Dynamic pressure) and the static pressure (Static Pressure) are determined as training data.

[0031] An initial prediction model containing a convolutional neural network (Convolutional Neural Networks, CNN), a bidirectional long short-term memory network (BiLSTM) and an attention mechanism (Attention) is established, and the prediction model is a multi-output deep learning model (Multi Output-CNN-BiLSTM-Attention).

[0032] The initial prediction model is iteratively trained using the training data, and the parameters and hyperparameters of the prediction model are continuously optimized until the preset stopping training condition is met, and the target prediction model is obtained, and the preset stopping training condition can be that the convergence degree of the target prediction model is greater than a preset convergence threshold, and the number of iterations can be not less than a preset number of rounds threshold.

[0033] In the embodiment, the iterative training of the initial prediction model by using the training data to obtain a target prediction model comprises: determining the initial prediction model as a current prediction model; training the current prediction model by using the training data to obtain a prediction result in a current round; wherein the prediction result comprises a wind speed prediction result and a differential pressure coefficient prediction result; determining a first mean square error of the wind speed prediction result and a second mean square error of the differential pressure coefficient prediction result by using a multi-task loss function, and obtaining a weighted mean square error of the first mean square error and the second mean square error; optimizing parameters of the current prediction model by using the weighted mean square error to obtain a next prediction model; determining the next prediction model as a new current prediction model, and jumping back to the step of training the current prediction model by using the training data to obtain a prediction result in a current round until a target prediction model meeting a preset training stop condition is obtained.

[0034] It should be noted that the prediction model in the embodiment is different from a conventional deep learning model in output, that is, the prediction model in the embodiment is a multi-output deep learning model, that is, not only a predicted wind speed, that is, a wind speed prediction result, but also a differential pressure coefficient prediction result is output. The initial prediction model is determined as the current prediction model, that is, the iterative training process is started, the training data is input into the current prediction model, the training in the current round is performed, the wind speed prediction result and the differential pressure coefficient prediction result in the current round are obtained, the model parameters are optimized by using the multi-task loss function and the prediction result, and the optimization process is: the first mean square error of the wind speed prediction result and the second mean square error of the differential pressure coefficient prediction result are determined by using the multi-task loss function, and the weighted mean square error of the first mean square error and the second mean square error is obtained, the parameters of the current prediction model are optimized by using the weighted mean square error to obtain a next prediction model; the next prediction model is determined as a new current prediction model, and the step of training the current prediction model by using the training data to obtain a prediction result in a current round is jumped back, that is, the training in the next round is started, until a target prediction model meeting a preset training stop condition is obtained.

[0035] In the embodiment, the optimization of the parameters of the current prediction model by using the weighted mean square error to obtain a next prediction model comprises: optimizing the parameters of the current prediction model by using the weighted mean square error and an Adam optimizer under a gradient clipping mechanism, and dynamically adjusting a learning rate of the current prediction model by using a scheduler to obtain a next prediction model.

[0036] Gradient clipping mechanism is applied in training to limit the gradient norm within 1.0 to prevent gradient explosion, and a multi-task loss function is used in the training process to combine the weighted mean square errors of wind speed prediction and differential pressure coefficient prediction, and an Adam optimizer is used with an initial learning rate of 0.001, and a ReduceLROnPlateau learning rate scheduler is used to automatically adjust the learning rate, that is, the weighted mean square error and the Adam optimizer are used to optimize the parameters of the current prediction model under the gradient clipping mechanism, and the scheduler is used to dynamically adjust the learning rate of the current prediction model to obtain the next prediction model.

[0037] In this embodiment, the training of the current prediction model using the training data to obtain the prediction result in the current round comprises: inputting the training data into the current prediction model; extracting a feature vector of the training data using the convolutional neural network through the current prediction model, and obtaining a feature sequence containing time sequence dependency based on the feature vector using the bidirectional long short-term memory network, obtaining a context vector from the feature sequence based on the attention mechanism, and generating a prediction result in the current round according to the shared feature representation extracted from the context vector.

[0038] The prediction model comprises a convolutional neural network, a bidirectional long short-term memory network and an attention mechanism, and further comprises an input projection layer and a shared feature layer.

[0039] The input projection layer expands the training data at a single time point into a sequence form, creating artificial time sequence length of 5 data, which is convenient for subsequent time sequence processing.

[0040] The feature extraction part is composed of three layers of convolutional neural network, each layer uses 32, 64 and 64 convolution kernels in turn, and the kernel size is 3, and is matched with batch normalization and ReLU activation function, and this structure design can capture the local correlation between the coaxial intelligent combiner parameters and extract key features, that is, the convolutional neural network extracts the feature vector of the training data.

[0041] Further, the feature vector needs to be dimensionally transformed to obtain a dimensionally transformed feature vector, and the bidirectional long short-term memory network obtains a feature sequence containing time sequence dependency based on the dimensionally transformed feature vector, the bidirectional long short-term memory network adopts a bidirectional structure, the hidden layer dimension is 64, and a dropout mechanism is set to prevent overfitting, and the bidirectional design enables the model to capture time dependency from both forward and backward directions, and models the dynamic characteristics of the coaxial intelligent combiner.

[0042] The attention mechanism is composed of two fully connected networks, with a Tanh activation function in the middle, and the output layer calculates the attention weight and then normalizes it by softmax to generate the context vector, that is, the attention mechanism obtains the context vector from the feature sequence, this mechanism identifies and focuses on the most critical time step in the sequence, enhancing the perception of the flow changes of the coaxial intelligent confluence device channel.

[0043] The shared feature layer after the attention layer extracts shared feature representations from the context vector through a fully connected network, ReLU activation and dropout processing, and this multi-task learning design enables the wind speed and differential pressure coefficient prediction to share the underlying features while maintaining their respective prediction characteristics, that is, the prediction result in the current round is generated based on the shared feature representations extracted from the context vector.

[0044] The prediction model is finally divided into two parallel output branches: wind speed prediction and differential pressure coefficient prediction, and each branch is composed of two fully connected networks, and this dual-output structure enables the model to learn two tasks simultaneously, improving the control accuracy of the coaxial intelligent confluence device.

[0045] In this embodiment, the wind speed prediction based on the initial main channel opening degree and the current environmental parameters using the target prediction model to obtain the target main channel opening degree interval includes: cutting the initial main channel opening degree as a direction cutting point, and performing first search in the high opening degree direction and the low opening degree direction based on the first preset step and the preset opening degree constraint condition respectively to obtain each rough main channel opening degree; and using the target prediction model, each rough main channel opening degree and the current environmental parameters to perform wind speed prediction to obtain a target main channel opening degree interval corresponding to the prediction result and the expected wind speed.

[0046] In the process of obtaining the target main passage opening interval, the main passage is responsible for macro wind speed interval control, so first, a first search, namely a rough search, is performed: the initial main passage opening is taken as a direction cutting point, that is, the initial main passage opening is respectively the starting point of the high opening direction and the low opening direction. Specifically, a first preset step size and a preset opening constraint condition are set. Since this is a rough search, the first preset step size is relatively large and cannot be too small. The preset opening constraint condition can include the maximum and minimum values of the main passage opening. The first search is performed in the direction greater than the initial main passage opening with the initial main passage opening as the starting point, and the first search is performed in the direction less than the initial main passage opening with the initial main passage opening as the starting point, so as to obtain each rough main passage opening. Next, the target prediction model, each rough main passage opening, and the current environmental parameters are used to perform wind speed prediction to obtain a target main passage opening interval corresponding to the expected wind speed, that is, the current environmental parameters are collected by using a sensor. The current environmental parameters include temperature, pressure difference, and air density. It can be understood that the target prediction model learns the correlation between the environmental parameters, the main and auxiliary dual passage openings, and the wind speed. When the result parameters are determined, the target prediction model can also predict other unknown parameters. Therefore, each rough main passage opening and the current environmental parameters are input into the target prediction model. The target prediction model can determine the target main passage opening interval from each rough main passage opening according to the input data and the expected wind speed.

[0047] Step S13: performing a second search from each main and auxiliary dual passage opening combination under the target main passage opening interval by using the target prediction model to obtain a target main and auxiliary dual passage opening combination corresponding to the expected wind speed.

[0048] In this embodiment, the second search from each main and auxiliary dual passage opening combination under the target main passage opening interval by using the target prediction model to obtain a target main and auxiliary dual passage opening combination corresponding to the expected wind speed includes: obtaining each main and auxiliary dual passage opening combination under the target main passage opening interval based on a second preset step size; the second preset step size is smaller than the first preset step size; obtaining wind speed prediction results under each main and auxiliary dual passage opening combination by using the target prediction model and determining wind speed errors between each wind speed prediction result and the expected wind speed; and performing a second search in each main and auxiliary dual passage opening combination to obtain a target main and auxiliary dual passage opening combination with the minimum wind speed error.

[0049] The wind speed generation has a multi-solution problem, that is, the same wind speed can be generated by different main and auxiliary passage configuration combinations of the same coaxial intelligent current combiner. The main passage is responsible for macro wind speed interval control, and the auxiliary passage performs fine adjustment. The functional division fully utilizes the structural characteristics of the coaxial intelligent current combiner and effectively prevents the inconsistent or conflict problems in the wind speed increment process.

[0050] For example Figure 3 A specific target main-auxiliary dual-channel opening combination acquisition flowchart is shown. A target prediction model is used to perform a second search, i.e., an accurate search, from each main-auxiliary dual-channel opening combination under a target main channel opening interval, so as to obtain a target main-auxiliary dual-channel opening combination corresponding to an expected wind speed. Specifically, because this is an accurate search, a second preset step size smaller than the first preset step size is set, and each main-auxiliary dual-channel opening combination under the target main channel opening interval is obtained based on the second preset step size. It can be understood that the main-auxiliary dual-channel opening combination is the main-auxiliary dual-channel opening. The target prediction model is used to obtain a wind speed prediction result under the current environmental parameter and each main-auxiliary dual-channel opening combination, and to determine a wind speed error between each wind speed prediction result and the expected wind speed. A second search is performed in each main-auxiliary dual-channel opening combination to obtain a target main-auxiliary dual-channel opening combination with the smallest wind speed error. For example, the current main-auxiliary dual-channel opening combination is determined under the target main channel opening interval, then the target prediction model is used to obtain a wind speed prediction result R1 under the current environmental parameter and the main-auxiliary dual-channel opening combination C1, the main-auxiliary dual-channel opening combination C2 is obtained based on the second preset step size and the main-auxiliary dual-channel opening combination C1, the target prediction model is used to obtain a wind speed prediction result R2 under the current environmental parameter and the main-auxiliary dual-channel opening combination C2, the wind speed prediction result R1 and the wind speed prediction result R2 are compared with the expected wind speed respectively, the wind speed error E1 and the wind speed error E2 are obtained, the main-auxiliary dual-channel opening combination with the smaller wind speed error is retained, and the main-auxiliary dual-channel opening combination C3 is obtained based on the second preset step size and the main-auxiliary dual-channel opening combination C2. According to the above steps, the wind speed error E3 between the wind speed prediction result R3 of the main-auxiliary dual-channel opening combination C3 and the expected wind speed is determined. In this way, the target main-auxiliary dual-channel opening combination with the smallest wind speed error is searched in the target main channel opening interval.

[0051] The core architecture of the intelligent control system based on the coaxial intelligent current collector consists of a channel calibration module and a reverse channel optimizer. These two parts realize the reverse deduction process from the target wind speed to the optimal channel configuration through the algorithm architecture. The channel calibration module establishes the mapping relationship between the wind speed and the main channel opening of the coaxial intelligent current collector. By storing nine key calibration points and applying the extrapolation algorithm, it provides a reliable initial main channel estimate for any target wind speed. The reverse channel optimizer adopts a multi-stage search strategy, including bidirectional search and fine tuning, systematically exploring the main-auxiliary channel space of the coaxial intelligent current collector and locating the optimal configuration. The control system also provides practical functions such as wind speed-channel configuration table generation, interpolation calculation, and adjustment process simulation, while supporting search process visualization, providing a comprehensive solution for wind tunnel operation, making the complex wind speed control process efficient, accurate, and visualized.

[0052] The wind speed generation has a multi-solution problem, that is, the same wind speed can be generated by different main and auxiliary channel configuration combinations of the coaxial intelligent collector. Lack of calibration will lead to a too large search space of the control algorithm and significantly reduce the efficiency. The calibration makes full use of the structural characteristics of the coaxial intelligent collector by explicitly defining the function division of the main channel responsible for the macro wind speed interval control and the auxiliary channel performing fine adjustment, effectively preventing the inconsistent or conflicting problems of channel configuration in the wind speed increasing process. More importantly, these calibration data provide key prior knowledge for the reverse control algorithm, enabling the system to quickly locate a reasonable initial search interval, significantly improving the control accuracy, stability and response speed of the coaxial intelligent collector, and providing a basic guarantee for high-quality wind tunnel experiments.

[0053] The reverse channel optimizer adopts a hierarchical structure design and a progressive search strategy, forming a complete control solution based on the coaxial intelligent collector. In the initialization stage, the module integrates the prediction model and the channel calibration two core functional components, and sets the channel physical limit and wind speed tolerance and other key parameters, laying the foundation for the subsequent search process. The core is to find the optimal channel configuration, which realizes the accurate mapping from the target wind speed to the optimal channel parameters through three organically connected search stages: first, based on the calibration data to obtain the initial main channel estimate, then perform bidirectional coarse search to determine the approximate range, and finally perform fine search near the best point to further optimize the results. This multi-stage search strategy of rough positioning-fine tuning not only balances the search efficiency and accuracy, but also highly matches the main and auxiliary dual-channel physical structure of the coaxial intelligent collector, effectively avoiding the trap of local optimal solution, and ensuring that the system can find the best channel control parameter combination under various working conditions.

[0054] Step S14: Adjusting the current main and auxiliary channel regulating valves based on the target main and auxiliary channel opening degree combination to control the wind speed of the collector corresponding to the expected wind speed.

[0055] For example Figure 4 As shown in a specific collector wind speed control schematic diagram, after determining the target main and auxiliary channel opening degree combination, the target main and auxiliary channel opening degree combination can be used to control the ground jet test device to adjust the current main and auxiliary channel regulating valves, so as to control the wind speed of the collector corresponding to the expected wind speed.

[0056] Further, the collector wind speed under different environmental parameters and main and auxiliary channel openings can also be visualized, that is, the relationship change curve between environmental parameters, main and auxiliary channel openings and collector wind speed is constructed, so that the user can more intuitively understand the collector wind speed.

[0057] The application has the following beneficial effects: the current collector has a main-aid dual-channel structure and comprises a main-aid dual-channel regulating valve and sensors for collecting various environmental parameters; the method comprises: determining an initial main-channel opening degree of the current collector according to an expected wind speed of the current collector and a target mapping relationship; wherein the target mapping relationship is a mapping relationship between various historical main-channel opening degrees of the current collector and wind speeds of the current collector; performing wind speed prediction based on the initial main-channel opening degree, various rough main-channel opening degrees obtained through first searching, and current environmental parameters by using a target prediction model to obtain a target main-channel opening degree interval; performing second searching from various main-aid dual-channel opening degree combinations under the target main-channel opening degree interval by using the target prediction model to obtain a target main-aid dual-channel opening degree combination corresponding to the expected wind speed; and controlling a ground jet test device to adjust the main-aid dual-channel regulating valve based on the target main-aid dual-channel opening degree combination to control the wind speed of the current collector to correspond to the expected wind speed; wherein the target prediction model is a deep learning model that fuses a convolutional neural network, a bidirectional long short-term memory network, and an attention mechanism, and the training data of the target prediction model is historical main-aid dual-channel opening degree combinations and historical flow field dynamic data of the current collector. As can be seen, the current collector adopts a main-aid dual-channel structure and a main-aid dual-channel regulating valve, can control the macroscopic wind speed interval through the main channel and finely adjust through the auxiliary channel, and can combine the sensors for collecting environmental parameters to sense data in real time, so as to more accurately control the wind speed compared with a single channel or simple mechanical control. The application determines the initial main-channel opening degree according to the expected wind speed and the target mapping relationship between the historical main-channel opening degree and the wind speed, which is more efficient than the existing method that relies on empirical calibration points. The target prediction model is used to predict the wind speed based on the initial main-channel opening degree, the rough main-channel opening degree obtained through rough searching, and the environmental parameters to determine the main-channel opening degree interval, and the main-aid dual-channel opening degree combination corresponding to the expected wind speed is selected from the interval to control the regulating valve. The model has stronger nonlinear feature extraction and key parameter focusing capabilities due to the training data comprising the historical main-aid dual-channel opening degree combinations and the flow field dynamic data, and the fusion of the convolutional neural network, the bidirectional long short-term memory network, and the attention mechanism. The model also has adaptability to environmental parameter changes, can more accurately predict the wind speed compared with traditional models, realizes the corresponding control of the current collector wind speed and the expected wind speed, and improves the control precision, efficiency, and flexibility.

[0058] The learning and convergence characteristics of the multi-output CNN-BiLSTM-Attention model for the coaxial intelligent current collector are analyzed, and these loss curve graphs include Figure 5 , Figure 6 and Figure 7Each group shows the performance changes of both the training set and the validation set simultaneously. The curves exhibit typical deep learning model training characteristics: a steep decline phase in the first 5 epochs, with the loss value rapidly dropping from the initial 0.35 to about 0.1, indicating that the model quickly captures the main patterns in the coaxial intelligent manifold primary and secondary channel configuration data; followed by a stable decline period from the 5th to the 15th epoch, where the learning rate slows down but still continues to optimize, showing that the model is refining its understanding of the complex nonlinear response relationship of the coaxial intelligent manifold; finally, it enters a stable period between the 15th and 32nd epochs, with the loss curve fluctuation decreasing and converging, and the training loss maintaining at about 0.07. It is worth noting that the validation loss (orange line) is always significantly lower than the training loss (blue line) throughout the training process, and is close to zero (about 0.003), which indicates that the model has strong generalization ability and can accurately predict the flow field parameters of the coaxial intelligent manifold configuration that has not been seen, without overfitting. The loss contributions of the two sub-tasks (wind speed prediction and pressure coefficient prediction) are roughly equivalent, indicating that the multi-task learning framework reasonably balances the dual control objectives of the coaxial intelligent manifold. The training process triggers the early stopping mechanism at the 32nd epoch, and the final model achieves near-perfect performance indicators on the independent test set (wind speed R²=0.9997, pressure coefficient R²=0.9992). These results fully demonstrate that the deep learning architecture can accurately model the complex relationship between the primary and secondary channel opening, temperature, pressure, and wind speed, pressure coefficient in the coaxial intelligent manifold, and reveal the internal working mechanism of the coaxial intelligent manifold as a structure-control collaborative unit. The model provides a reliable algorithm foundation for building a high-precision, low-turbulence, real-time adaptive wind tunnel control system, significantly reducing the cost of traditional wind tunnel calibration, and improving the adaptability of the coaxial intelligent manifold to environmental variable changes, providing strong intelligent support for the stable operation of the device in complex aerodynamic environments.

[0059] For example Figure 8 a specific 7°C wind speed curve graph, Figure 9 a specific 9°C wind speed curve graph, Figure 10 a specific 11°C wind speed curve graph, Figure 11 a specific 7°C pressure coefficient curve graph, Figure 12 a specific 9°C pressure coefficient curve graph, Figure 13The figure shows a specific pressure differential coefficient curve under 11°C conditions, illustrating the influence of the main and auxiliary channel openings of a coaxial smart combiner on wind speed and pressure differential coefficient under three different temperature conditions (7°C, 9°C, and 11°C). This set of three-dimensional response surfaces demonstrates the systematic impact of temperature changes on the control characteristics of the coaxial smart combiner, revealing its internal fluid dynamics mechanism and structure-control coupling relationship. The figure shows nonlinear characteristics of wind speed and pressure differential coefficient changes with the main and auxiliary channel openings: wind speed exhibits an "S"-shaped growth curve with increasing channel opening, increasing slowly in the low opening range (0-30%), rapidly in the medium opening range (30-70%), and tending to level off in the high opening range (70-100%); the pressure differential coefficient shows a characteristic of first rising slowly and then increasing rapidly, with more significant changes after the main channel opening exceeds 40%. The auxiliary channel has a relatively small impact on the two parameters, but its adjustment effect becomes more obvious in the high-opening region of the main channel, reflecting the interactive coupling between the main and auxiliary channels of the coaxial intelligent combiner, which is consistent with its "coarse adjustment + fine adjustment" structure-control design. As the temperature increases (from 7°C to 11°C), the three-dimensional surfaces of wind speed and pressure differential coefficient exhibit systematic changes: the surface shifts upwards overall, indicating that under the same channel configuration, higher temperatures produce greater wind speed and pressure differential coefficient output; the shape and curvature characteristics of the surface also change, at 11°C ( Figure 10 , Figure 13 The curved surface exhibits a smoother transition characteristic in the high channel opening region, while at 7°C ( Figure 8 , Figure 11 The corresponding region exhibits a steeper rate of change. This temperature dependence is consistent with gas dynamics theory, namely, that increased temperature leads to decreased gas density and increased kinematic viscosity, affecting Reynolds number and pressure distribution, thereby altering the flow field characteristics inside the coaxial smart combiner. Temperature changes not only cause surface displacement but also alter surface morphology, indicating that the effect of temperature on the coaxial smart combiner is nonlinear. This multi-parameter coupling characteristic makes it difficult for traditional control methods based on physical models or lookup tables to accurately describe and predict the system behavior of the coaxial smart combiner. The multi-output CNN-BiLSTM-Attention deep learning architecture, by fusing feature extraction from convolutional neural networks, temporal modeling from bidirectional LSTM, and key information focusing through an attention mechanism, achieves modeling of the complex nonlinear system of the coaxial smart combiner (wind speed R² = 0.9997, pressure differential R² = 0.9992). Visualization results verify the model's ability to capture temperature effects, providing operators with a reference for the dynamic behavior of the coaxial smart combiner under different environmental conditions. This method has application value in improving wind tunnel testing accuracy, reducing calibration costs, and realizing adaptive control of coaxial intelligent combiners, and provides a reference for the collaborative design of fluid dynamics testing devices.

[0060] For example Figure 14 A specific search path diagram of one of the channel parameters is shown, showing the search trajectory of the main channel and auxiliary channel opening, marked with blue dots, green triangles and purple squares for the three stages of down search, up search and fine search respectively. The search starts from the initial estimated point of the main channel 39% and the auxiliary channel 5%, and finally converges to the optimal configuration point (main channel 40%, auxiliary channel 2%) through the coarse-fine search strategy, marked with a red star. The three gray guide lines show the algorithm's gradual approach to the optimal solution by exploring the parameter space. Figure 15 A specific search process wind speed change diagram is shown, which traces the dynamic changes of wind speed during the entire search process, starting from the initial point (blue dot for down search) and the up search starting point (green triangle), and gradually approaching the target value 110 m / s (red dotted line) through 18 steps of fine search (purple square), and finally reaching the optimal solution of 110.21 m / s (red star). This process shows the adaptive adjustment of the algorithm in the wind speed space, avoiding local optimal solution. Figure 16 A specific search error change diagram is shown, which records the trend of search error with steps. The error gradually decreases from the initial 0.26 m / s, with fluctuations in the search process (rising and then falling between steps 6-9), and finally converges to the minimum error of 0.21 m / s (red star). This non-monotonic error curve reflects the algorithm's balance strategy between exploration and exploitation. Figure 17 A specific wind speed heat map is shown, which visualizes the influence of the coaxial intelligent current collector main and auxiliary channel opening combination on wind speed in the form of a heat map. The color changes from dark blue to yellow, corresponding to the change of wind speed from 110.2 m / s to 110.344 m / s. Figure 14 and Figure 17 The optimal region centered on the main channel 40% and auxiliary channel 2% is shown (red star), surrounded by a wind speed gradient distribution. The heat map reveals the sensitivity characteristics of wind speed to channel parameters - the sensitivity of wind speed to main channel opening is higher than that to auxiliary channel, consistent with the physical characteristics of coaxial intelligent current collector. This method combines channel calibration data and the predictive ability of the CNN-BiLSTM-Attention model to achieve control of the coaxial intelligent current collector nonlinear system. In the test case of target wind speed 110 m / s, the error is controlled within 0.21 m / s (relative error about 0.19%) through 18 steps of iteration, which is better than traditional control methods. This visual search path verifies the effectiveness of the algorithm for parameter optimization of the coaxial intelligent current collector, providing a reference for test procedures and control strategies.

[0061] Referring to Figure 18As shown, the embodiment of the present application discloses a wind speed control device of a flow concentrator, the flow concentrator has a main-aid dual-channel structure and contains a main-aid dual-channel regulating valve and sensors for collecting various environmental parameters; the device comprises: An initial opening degree determination module 11 is configured to determine an initial main channel opening degree of the flow concentrator according to an expected wind speed of the flow concentrator and a target mapping relationship; wherein the target mapping relationship is a mapping relationship between various historical main channel opening degrees of the flow concentrator and wind speeds of the flow concentrator; An opening degree interval acquisition module 12 is configured to perform wind speed prediction based on the initial main channel opening degree, various rough main channel opening degrees obtained by first search using a target prediction model, and current environmental parameters, so as to obtain a target main channel opening degree interval; An opening degree combination screening module 13 is configured to perform second search from various main-aid dual-channel opening degree combinations under the target main channel opening degree interval using the target prediction model, so as to obtain a target main-aid dual-channel opening degree combination corresponding to the expected wind speed; A wind speed control module 14 is configured to control a ground jet test device to adjust the main-aid dual-channel regulating valve based on the target main-aid dual-channel opening degree combination, so as to control the wind speed of the flow concentrator to correspond to the expected wind speed. Wherein, the target prediction model is a deep learning model fusing a convolutional neural network, a bidirectional long short-term memory network and an attention mechanism, and training data of the target prediction model is historical main-aid dual-channel opening degree combinations and historical flow field dynamic data of the flow concentrator.

[0062] The application has the following beneficial effects: the current collector has a main-aid dual-channel structure and comprises a main-aid dual-channel regulating valve and sensors for collecting various environmental parameters; the method comprises: determining an initial main-channel opening degree of the current collector according to an expected wind speed of the current collector and a target mapping relationship; wherein the target mapping relationship is a mapping relationship between various historical main-channel opening degrees of the current collector and wind speeds of the current collector; performing wind speed prediction based on the initial main-channel opening degree, various rough main-channel opening degrees obtained through first searching, and current environmental parameters by using a target prediction model to obtain a target main-channel opening degree interval; performing second searching from various main-aid dual-channel opening degree combinations under the target main-channel opening degree interval by using the target prediction model to obtain a target main-aid dual-channel opening degree combination corresponding to the expected wind speed; and controlling a ground jet test device to adjust the main-aid dual-channel regulating valve based on the target main-aid dual-channel opening degree combination to control the wind speed of the current collector to correspond to the expected wind speed; wherein the target prediction model is a deep learning model that fuses a convolutional neural network, a bidirectional long short-term memory network, and an attention mechanism, and training data of the target prediction model is historical main-aid dual-channel opening degree combinations and historical flow field dynamic data of the current collector. As can be seen, the current collector adopts a main-aid dual-channel structure and a main-aid dual-channel regulating valve, can control the macroscopic wind speed interval through the main channel and finely adjust through the auxiliary channel, and can more accurately control the wind speed by combining the real-time sensing data of the sensors collecting environmental parameters compared with single-channel or simple mechanical control. The application determines the initial main-channel opening degree according to the expected wind speed and the target mapping relationship between the historical main-channel opening degrees and the wind speed, which is more efficient than the existing method that relies on empirical calibration points. The target prediction model is used to predict the wind speed based on the initial main-channel opening degree, the rough main-channel opening degrees obtained through rough searching, and the environmental parameters to determine the main-channel opening degree interval, and the main-aid dual-channel opening degree combination corresponding to the expected wind speed is selected from the interval to control the regulating valve. The model has stronger nonlinear feature extraction and key parameter focusing capabilities due to the training data comprising historical main-aid dual-channel opening degree combinations and flow field dynamic data and the fusion of the convolutional neural network, the bidirectional long short-term memory network, and the attention mechanism, can adapt to changes in environmental parameters, can more accurately predict the wind speed compared with traditional models, realizes the corresponding control of the wind speed of the current collector and the expected wind speed, and improves the control precision, efficiency, and flexibility.

[0063] Further, the embodiment of the application also provides an electronic device. Figure 19 The electronic device 20 is shown in the structure diagram according to an exemplary embodiment, and the content in the figure cannot be considered as any limitation on the use range of the application.

[0064] Figure 19A structural schematic diagram of an electronic device is provided in the embodiments of the present application. Specifically, it can include at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 is configured to store a computer program, which is loaded and executed by the processor 21 to implement the related steps in the busbar wind speed control method performed by the electronic device disclosed in any of the preceding embodiments.

[0065] In the embodiments, the power supply 23 is configured to provide working voltage for each hardware device on the electronic device; the communication interface 24 is capable of creating a data transmission channel between the electronic device and external devices, and the communication protocol followed by the communication interface 24 is any communication protocol applicable to the technical solutions of the present application, which is not limited specifically herein; the input / output interface 25 is configured to obtain external input data or output data to the outside world, and the specific interface type can be selected according to the specific application needs, which is not limited specifically herein.

[0066] The processor 21 can include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 21 can be implemented in at least one of a hardware form of a DSP (Digital Signal Processing), an FPGA (Field-Programmable Gate Array), and a PLA (Programmable Logic Array). The processor 21 can also include a main processor and a coprocessor. The main processor is a processor for processing data in a wake-up state, also known as a CPU (Central Processing Unit). The coprocessor is a low-power processor for processing data in a standby state. In some embodiments, the processor 21 can be integrated with a GPU (Graphics Processing Unit) that is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 21 can also include an AI (Artificial Intelligence) processor for processing machine learning-related computing operations.

[0067] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, a random access memory, a magnetic disk, or an optical disk, etc. The resources stored thereon include an operating system 221, a computer program 222, and data 223, etc. The storage mode can be temporary storage or permanent storage.

[0068] The operating system 221 is used to manage and control each hardware device on the electronic device and the computer program 222, so as to realize the operation and processing of the processor 21 on the mass data 223 in the memory 22, and can be Windows, Unix, Linux, etc. In addition to the computer program capable of completing the wind speed control method of the busbar disclosed by the electronic device executed by the electronic device, the computer program 222 can further include a computer program capable of completing other specific work. The data 223 can include the data transmitted by the external device received by the electronic device, and can also include the data collected by the self input and output interface 25, etc.

[0069] Further, the application also discloses a computer readable storage medium for storing a computer program; wherein the computer program is executed by the processor to realize the wind speed control method of the busbar disclosed above. For the specific steps of the method, please refer to the corresponding content disclosed in the foregoing embodiments, which will not be repeated here.

[0070] In the specification, each embodiment is described in a progressive manner, and each embodiment focuses on the difference from other embodiments. For the same or similar parts between each embodiment, please refer to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant part is described in the method part.

[0071] Those skilled in the art will further appreciate that the units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or any combination thereof. To clearly illustrate this interchangeability of hardware and software, various examples have been described herein in terms of their functionality, which has been described generally and symbolically in flow charts. Having thus described the functionality of the examples in terms of a process, it is appreciated that this functionality can be implemented by one or more types of electrical circuits or computer software, which are collectively referred to herein as a "circuit" that can carry out a variety of operations described herein. The circuit can include a variety of different types of general purpose or special purpose circuits, or combinations thereof. In addition, it is further noted that the embodiments disclosed herein can be modified to comprise more or less steps or operations than those disclosed herein, and such modifications are contemplated and considered within the scope of embodiments of the present application. The steps or operations of the methods or algorithms described in connection with the embodiments disclosed herein can be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module can reside in Random Access Memory (RAM), flash memory, Read-only memory (ROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor such that the processor can read information from, and write information to, the storage medium. In the alternative, hard disk can be used as a non-transitory storage medium to store software modules.

[0072] Finally, it should be noted that the terms "first", "second", and the like, herein do not denote any order, quantity, combination, or importance, but rather are used to distinguish one element from another, and are more especially used for the purpose of illustration and not of limitation. Also, the terms "comprises", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without limitation, an element preceded by "comprises a" does not, without more constraints, foreclose the existence of additional identical elements in the process, method, article, or apparatus.

[0073] The wind speed control method, device, equipment and medium of the busbar provided by the application are described in detail above, the principle and implementation mode of the application are described by applying specific examples in this paper, and the above example description is only used to help understand the method of the application and its core idea; meanwhile, for those skilled in the art, according to the idea of the application, the specific implementation mode and application range will be changed, and the above description should not be understood as a limitation on the application.

Claims

1. A wind speed control method of a concentrator, characterized by, The collector has a main-aid dual-channel structure and comprises a main-aid dual-channel regulating valve and sensors for collecting various environmental parameters; the method comprises: determining an initial main-channel opening degree of the collector according to an expected wind speed of the collector and a target mapping relationship; wherein the target mapping relationship is a mapping relationship between various historical main-channel opening degrees of the collector and wind speeds of the collector; performing wind speed prediction based on the initial main-channel opening degree, various rough main-channel opening degrees obtained through first searching, and current environmental parameters, to obtain a target main-channel opening degree interval; performing second searching from various main-aid dual-channel opening degree combinations under the target main-channel opening degree interval by using the target prediction model, to obtain a target main-aid dual-channel opening degree combination corresponding to the expected wind speed; controlling a ground jet test device to adjust the main-aid dual-channel regulating valve based on the target main-aid dual-channel opening degree combination, so as to control the wind speed of the collector to correspond to the expected wind speed. The target prediction model is a deep learning model that fuses a convolutional neural network, a bidirectional long short-term memory network, and an attention mechanism, and training data of the target prediction model is historical main-aid dual-channel opening degree combinations and historical flow field dynamic data of the collector.

2. The wind speed control method of the concentrator according to claim 1, wherein, The target prediction model is obtained by: determining historical main-aid dual-channel opening degree combinations and historical flow field dynamic data of the collector as training data; wherein the historical flow field dynamic data comprises historical environmental parameters and historical wind speeds, and the environmental parameters comprise temperature, air pressure, and air density of the collector; establishing an initial prediction model comprising a convolutional neural network, a bidirectional long short-term memory network, and an attention mechanism, and iteratively training the initial prediction model by using the training data, to obtain the target prediction model.

3. The wind speed control method of the concentrator according to claim 2, wherein, The target prediction model is obtained by iteratively training the initial prediction model by using the training data, comprising: determining the initial prediction model as a current prediction model; training the current prediction model by using the training data, to obtain a prediction result in a current round; wherein the prediction result comprises a wind speed prediction result and a differential pressure coefficient prediction result; determining a first mean square error of the wind speed prediction result and a second mean square error of the differential pressure coefficient prediction result by using a multi-task loss function, and obtaining a weighted mean square error of the first mean square error and the second mean square error; optimizing parameters of the current prediction model by using the weighted mean square error, to obtain a next prediction model; determining the next prediction model as a new current prediction model, and returning to the step of training the current prediction model by using the training data, to obtain a prediction result in a current round, until a target prediction model that meets a preset training stop condition is obtained.

4. The wind speed control method of the concentrator according to claim 3, wherein, The target prediction model is obtained by iteratively training the initial prediction model by using the training data, comprising: The parameters of the current prediction model are optimized by using the weighted mean square error and the Adam optimizer under a gradient clipping mechanism, and a learning rate of the current prediction model is dynamically adjusted by using a scheduler to obtain a next prediction model.

5. The wind speed control method of the concentrator according to claim 3, wherein, The training of the current prediction model by using the training data to obtain a prediction result in a current round comprises: inputting the training data into the current prediction model; extracting a feature vector of the training data by using the convolutional neural network and obtaining a feature sequence containing a time sequence dependency relationship based on the feature vector by using the bidirectional long short-term memory network, obtaining a context vector from the feature sequence based on the attention mechanism, and generating a prediction result in a current round according to a shared feature representation extracted from the context vector by using the current prediction model.

6. The wind speed control method of the concentrator according to claim 1, wherein, The wind speed prediction by using a target prediction model, each rough main channel opening obtained by performing a first search based on the initial main channel opening, and a current environmental parameter to obtain a target main channel opening interval comprises: cutting the initial main channel opening as a direction cutting point, and performing a first search in a high opening direction and a low opening direction based on a first preset step length and a preset opening constraint condition to obtain each rough main channel opening; The wind speed prediction by using a target prediction model, each rough main channel opening, and a current environmental parameter to obtain a target main channel opening interval corresponding to a prediction result and the expected wind speed comprises:

7. The wind speed control method of a concentrator according to claim 6, wherein The second search from each main auxiliary double-channel opening combination under the target main channel opening interval by using the target prediction model to obtain a target main auxiliary double-channel opening combination corresponding to the expected wind speed comprises: obtaining each main auxiliary double-channel opening combination under the target main channel opening interval based on a second preset step length; wherein the second preset step length is smaller than the first preset step length; obtaining wind speed prediction results under each main auxiliary double-channel opening combination by using the target prediction model, and determining wind speed errors between each wind speed prediction result and the expected wind speed; performing a second search in each main auxiliary double-channel opening combination to obtain a target main auxiliary double-channel opening combination with the minimum wind speed error.

8. A wind speed control device for a concentrator, characterized by, The current collector has a main auxiliary double-channel structure and comprises a main auxiliary double-channel regulating valve and a sensor for collecting each environmental parameter; the device comprises: An initial opening determination module is configured to determine an initial main channel opening of the current collector according to an expected wind speed of the current collector and a target mapping relationship; wherein the target mapping relationship is a mapping relationship between each historical main channel opening of the current collector and a wind speed of the current collector. An opening interval obtaining module is configured to perform wind speed prediction by using a target prediction model, each rough main channel opening obtained by performing a first search based on the initial main channel opening, and a current environmental parameter to obtain a target main channel opening interval. An opening combination screening module is configured to perform a second search from each main auxiliary double-channel opening combination under the target main channel opening interval by using the target prediction model to obtain a target main auxiliary double-channel opening combination corresponding to the expected wind speed. A wind speed control module is configured to control the ground jet test device to adjust the main and auxiliary dual-channel regulating valve based on the target main and auxiliary dual-channel opening combination, so that the wind speed of the current flow collector corresponds to the expected wind speed. The target prediction model is a deep learning model integrating a convolutional neural network, a bidirectional long short-term memory network, and an attention mechanism, and the training data of the target prediction model is historical main and auxiliary dual-channel opening combinations and historical flow field dynamic data of the flow collector.

9. An electronic device, comprising: The method comprises the following steps: a memory for storing a computer program; a processor for executing the computer program to implement the steps of the wind speed control method of the flow collector according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, a memory for storing a computer program; wherein the computer program, when executed by a processor, implements the steps of the wind speed control method of the flow collector according to any one of claims 1 to 7.

Images