Energy-saving solar exhaust fan and control method thereof

Abstract

The present application relates to the field of solar exhaust fans, and specifically to an energy-saving solar exhaust fan and a control method thereof. It uses the time series of the ambient temperature collected by the temperature sensor and the time series of the battery remaining power collected by the battery remaining power detector, and uses artificial intelligence technology to analyze the ambient temperature and the battery remaining power respectively to obtain the energy-saving comprehensive guide speed recommendation value, and generates a motor speed adjustment instruction based on the comparison between the energy-saving comprehensive guide speed recommendation value and the current speed value to adapt to the ventilation requirements under different environmental characteristics, and take into account the real-time power of the solar panel while meeting the ventilation requirements to extend the battery life. In this way, it can automatically adapt to the changes in the environmental state to make corresponding adjustments, avoiding the problem of shortening the battery life due to the long-term operation of the energy storage battery in a power-deficient state.

Description

Technical Field

[0001] The present application relates to the field of solar exhaust fans, and in particular to an energy-saving solar exhaust fan and a control method thereof. Background Art

[0002] Solar exhaust fan is a ventilation device that uses solar energy as a power source. It is widely used in residential, commercial buildings, agricultural greenhouses, outdoor facilities and other occasions to provide natural ventilation, improve air quality and reduce indoor temperature. For example, China's public patent CN113446250A discloses a solar exhaust fan, which effectively blocks rain from entering the room through the design of solar panels and rain shields, and the design of the mounting mechanism enables the solar exhaust fan to adapt to different roof inclinations to ensure that the air outlet is vertically facing upwards. The design of its adjustment mechanism also enables the solar panel to automatically adjust the angle according to the position of the sun to maximize the reception and utilization of solar energy.

[0003] However, in this prior art, it is found that it still uses the traditional exhaust fan control strategy, that is, the exhaust fan speed is controlled by manually controlling the switch and gear position. This control method is not only less intelligent, but also causes unnecessary energy waste. For example, if the indoor temperature is low but a higher-speed wind control is used, this will cause unnecessary energy waste. At the same time, the existing exhaust fan speed control strategy cannot adapt to the current battery power state change of the energy storage battery. Running under a long-term power shortage may cause battery loss, affecting the service life of the energy storage battery of the energy-saving solar exhaust fan.

[0004] Therefore, an energy-saving solar exhaust fan is desired. Summary of the invention

[0005] This application is made in view of the above problems. One object of this application is to provide an energy-saving solar exhaust fan and a control method thereof.

[0006] The embodiment of the present application provides an energy-saving solar exhaust fan, comprising: a housing, a motor installed in the housing, a fan blade installed at the power output end of the motor, a solar panel installed in the housing and exposed to the outside, an energy storage battery installed in the housing and electrically connected to the solar panel, and the energy storage battery is electrically connected to the motor; an air outlet is provided at the upper end of the housing, and it also includes:

[0007] a temperature sensor, wherein the temperature sensor is mounted on the housing;

[0008] A battery remaining power detector, which is used to detect the remaining power of the energy storage battery;

[0009] A motor central controller, the motor central controller is communicatively connected to the temperature sensor, the battery remaining power detector and the motor, wherein the motor central controller is used to control the rotation speed of the motor based on data collected by the temperature sensor and the battery remaining power detector.

[0010] For example, according to the energy-saving solar exhaust fan of an embodiment of the present application, the motor central controller includes:

[0011] An ambient temperature acquisition module, used to acquire a time series of ambient temperature acquired by the temperature sensor;

[0012] A battery remaining power acquisition module, used to acquire the time series of the battery remaining power acquired by the battery remaining power detector;

[0013] An ambient temperature local time series association module, used for inputting the time series of the ambient temperature into a first-order sequence encoder based on a 1D-CNN model to obtain a sequence of ambient temperature local time series association implicit feature vectors;

[0014] An ambient temperature feature enhancement module, used for inputting the sequence of the ambient temperature local temporal correlation implicit feature vectors into a reverse screening enhancement network based on the feature distribution of the anchor center to obtain a sequence of enhanced ambient temperature local temporal correlation implicit feature vectors;

[0015] An ambient temperature global temporal association module is used to input the sequence of the enhanced ambient temperature local temporal association implicit feature vectors into a second-order sequence encoder based on a bidirectional long short-term memory neural network model to obtain an ambient temperature global temporal association implicit feature vector;

[0016] A battery remaining power feature processing module, used for passing the time series of the battery remaining power through the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center, and the second-order sequence encoder based on the bidirectional long short-term memory neural network model to obtain a global time series associated implicit feature vector of the remaining power;

[0017] A speed recommendation module, used for inputting the remaining power global time series associated implicit feature vector and the ambient temperature global time series associated implicit feature vector into a speed recommender to obtain an ambient temperature guided speed recommendation value and a remaining power guided speed recommendation value;

[0018] The mediation instruction generation module is used to generate a motor speed mediation instruction based on the energy-saving comprehensive guided speed recommendation value and the current speed value of the motor.

[0019] For example, according to the energy-saving solar exhaust fan of an embodiment of the present application, the ambient temperature characteristic enhancement module includes:

[0020] An ambient temperature sequence class label calculation unit, used to calculate the mean vector of the sequence of the ambient temperature local time series associated implicit feature vectors to obtain the ambient temperature sequence distribution center vector;

[0021] An ambient temperature local time series implicit correlation coefficient calculation unit is used to calculate the correlation coefficient between each ambient temperature local time series implicit characteristic vector in the sequence of the ambient temperature local time series associated implicit characteristic vectors and the ambient temperature sequence distribution center vector to obtain a sequence of ambient temperature local time series implicit correlation coefficients;

[0022] An anti-correlation coefficient calculation unit, used for calculating the reciprocal of each ambient temperature local time series implicit correlation coefficient in the sequence of ambient temperature local time series implicit correlation coefficients to obtain a sequence of ambient temperature local time series anti-correlation coefficients;

[0023] A normalization processing unit, used for performing normalization processing on the sequence of the local time series anti-correlation coefficients of the ambient temperature using a Softmax function to obtain a sequence of the local time series anti-correlation energy coefficients of the ambient temperature;

[0024] A feature weighting unit, used to weight each ambient temperature local time series associated implicit feature vector in the sequence of ambient temperature local time series associated implicit feature vectors by using each ambient temperature local time series anti-correlation energy coefficient in the sequence of ambient temperature local time series anti-correlation energy coefficients as a weight to obtain a sequence of reversely inhibited ambient temperature local time series associated implicit feature vectors;

[0025] A differential calculation unit is used to calculate the positional difference of each corresponding sequence of the ambient temperature local timing association implicit feature vector and the reverse inhibition ambient temperature local timing association implicit feature vector between the sequence of the ambient temperature local timing association implicit feature vector and the sequence of the reverse inhibition ambient temperature local timing association implicit feature vector to obtain the sequence of the enhanced ambient temperature local timing association implicit feature vector.

[0026] For example, according to the energy-saving solar exhaust fan of an embodiment of the present application, the ambient temperature local time series implicit correlation coefficient calculation unit is used to:

[0027] Calculate the point-by-point multiplication between the ambient temperature local time series associated implicit feature vector and the ambient temperature sequence distribution center vector to obtain the ambient temperature local time series associated implicit class support feature vector;

[0028] Calculate the sum of the eigenvalues ​​of each position in the ambient temperature local temporal association implicit class support eigenvector to obtain the ambient temperature local temporal association implicit class support sum;

[0029] Calculate the sum of squares of the eigenvalues ​​at each position in the ambient temperature local time series associated implicit eigenvector and the sum of squares of the eigenvalues ​​at each position in the ambient temperature sequence distribution center vector to obtain the ambient temperature local time series associated implicit square value and the ambient temperature sequence distribution center square value;

[0030] Calculate the square root of the sum of the implicit square value of the local temporal correlation of the ambient temperature and the central square value of the ambient temperature sequence distribution to obtain an ambient temperature correlation value;

[0031] The ambient temperature local timing association implicit class support sum value is divided by the ambient temperature association value to obtain the ambient temperature local timing implicit association coefficient.

[0032] For example, according to the energy-saving solar exhaust fan of an embodiment of the present application, the normalization processing unit is used to:

[0033] Using each local temporal anti-correlation coefficient of the sequence of local temporal anti-correlation coefficients of ambient temperature as an exponent of a natural constant to calculate an exponential function value based on the natural constant according to the position to obtain a sequence of local temporal anti-correlation exponential coefficients of ambient temperature;

[0034] Calculating the sum of each ambient temperature local time series anti-correlation index coefficient in the sequence of ambient temperature local time series anti-correlation index coefficients to obtain a total sum of the ambient temperature local time series anti-correlation coefficients;

[0035] Calculate the sum of the total sum of the ambient temperature local temporal anti-correlation coefficients and a constant one to obtain an ambient temperature local temporal anti-correlation modulation coefficient;

[0036] Each ambient temperature local time series anti-correlation index coefficient in the sequence of ambient temperature local time series anti-correlation index coefficients is divided by the ambient temperature local time series anti-correlation modulation coefficient to obtain the sequence of ambient temperature local time series anti-correlation energy coefficients.

[0037] For example, according to the energy-saving solar exhaust fan of an embodiment of the present application, the speed recommendation module is used to: input the remaining power global timing associated implicit feature vector and the ambient temperature global timing associated implicit feature vector into the decoder-based speed recommender respectively to obtain the ambient temperature guided speed recommendation value and the remaining power guided speed recommendation value.

[0038] For example, according to the energy-saving solar exhaust fan of an embodiment of the present application, it also includes a training module for training the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center, the second-order sequence encoder based on the bidirectional long short-term memory neural network model, and the decoder-based speed recommender.

[0039] For example, according to the energy-saving solar exhaust fan of an embodiment of the present application, the training module includes:

[0040] A training data acquisition unit, used to acquire training data, wherein the training data includes a time series of a training environment temperature acquired by the temperature sensor and a time series of a training battery remaining capacity acquired by the battery remaining capacity detector;

[0041] A training environment temperature local time series association unit, used for inputting the time series of the training environment temperature into the first-order sequence encoder based on the 1D-CNN model to obtain a sequence of implicit feature vectors of the training environment temperature local time series association;

[0042] A training environment temperature feature enhancement unit, used for inputting the sequence of the training environment temperature local time series associated implicit feature vectors into the anchor center-based feature distribution reverse screening enhancement network to obtain a sequence of training enhanced environment temperature local time series associated implicit feature vectors;

[0043] A training ambient temperature global temporal association unit is used to input the sequence of the training enhanced ambient temperature local temporal association implicit feature vectors into the second-order sequence encoder based on the bidirectional long short-term memory neural network model to obtain the training ambient temperature global temporal association implicit feature vector;

[0044] A training battery remaining power feature processing unit, used for passing the time series of the training battery remaining power through the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center, and the second-order sequence encoder based on the bidirectional long short-term memory neural network model to obtain a global time series associated implicit feature vector of the training remaining power;

[0045] A decoding loss function value calculation unit, used for inputting the training remaining power global time series associated implicit feature vector and the training ambient temperature global time series associated implicit feature vector into the decoder-based speed recommender to obtain an ambient temperature decoding loss function value and a remaining power decoding loss function value;

[0046] A predetermined loss function value calculation unit, used to calculate the predetermined loss function value of the training ambient temperature global time series associated implicit feature vector to obtain the ambient temperature global time series associated implicit loss function value;

[0047] A model training unit is used to use the weighted sum of the ambient temperature decoding loss function value, the remaining power decoding loss function value and the ambient temperature global time series associated implicit loss function value as the loss function value and train the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center, the second-order sequence encoder based on the bidirectional long short-term memory neural network model and the decoder-based speed recommender through gradient descent back propagation.

[0048] An embodiment of the present application further provides an energy-saving solar exhaust fan control method, wherein the energy-saving solar exhaust fan control method is used to control any of the energy-saving solar exhaust fans described above.

[0049] For example, according to an embodiment of the present application, the energy-saving solar exhaust fan control method includes:

[0050] Get the time series of ambient temperature collected by the temperature sensor;

[0051] Obtaining a time series of the remaining battery power collected by the remaining battery power detector;

[0052] Inputting the time series of the ambient temperature into a first-order sequence encoder based on a 1D-CNN model to obtain a sequence of local temporal correlation implicit feature vectors of the ambient temperature;

[0053] Inputting the sequence of the local temporal correlation implicit feature vectors of the ambient temperature into a reverse screening reinforcement network based on the feature distribution of the anchor center to obtain a sequence of the local temporal correlation implicit feature vectors of the enhanced ambient temperature;

[0054] Inputting the sequence of the enhanced ambient temperature local temporal correlation implicit feature vectors into a second-order sequence encoder based on a bidirectional long short-term memory neural network model to obtain the ambient temperature global temporal correlation implicit feature vector;

[0055] The time series of the remaining battery power is passed through the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center, and the second-order sequence encoder based on the bidirectional long short-term memory neural network model to obtain a global time series associated implicit feature vector of the remaining power;

[0056] Inputting the remaining power global time series associated implicit feature vector and the ambient temperature global time series associated implicit feature vector into a speed recommender to obtain an ambient temperature guided speed recommendation value and a remaining power guided speed recommendation value;

[0057] Based on the energy-saving comprehensive guide speed recommendation value and the current speed value of the motor, a motor speed adjustment instruction is generated.

[0058] According to the energy-saving solar exhaust fan and its control method of the embodiment of the present application, it can automatically adapt to the changes in environmental conditions to make corresponding adjustments, avoiding the problem of shortening the battery life due to the energy storage battery running for a long time in a low-power state. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments of the present application are briefly introduced below. Obviously, the drawings described below only relate to some embodiments of the present application, and are not intended to limit the present application.

[0060] Figure 1 A schematic diagram of the structure of an energy-saving solar exhaust fan in an embodiment of the present application is shown;

[0061] Figure 2 A schematic diagram of the structure of a motor central controller of an energy-saving solar exhaust fan in an embodiment of the present application is shown;

[0062] Figure 3 A schematic diagram showing the structure of an ambient temperature characteristic enhancement module for an energy-saving solar exhaust fan in an embodiment of the present application is shown; and

[0063] Figure 4 A flow chart of an energy-saving solar exhaust fan control method in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0064] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.

[0065] The terms used in this specification are those common terms currently widely used in the art in consideration of the functions of the present application, but these terms may vary according to the intentions of those skilled in the art, precedents, or new technologies in the art. In addition, specific terms may be selected, and in this case, their detailed meanings will be described in the detailed description of the present application. Therefore, the terms used in the specification should not be understood as simple names, but rather as a general description based on the meaning of the terms and the present application.

[0066] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0067] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. At the same time, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0068] In addition, the application architecture diagram in the embodiment of the present application is intended to more clearly illustrate the technical solution in the embodiment of the present application, and does not constitute a limitation on the technical solution provided in the embodiment of the present application. Of course, for other application architectures and business applications, the technical solution provided in the embodiment of the present application is also applicable to similar problems.

[0069] The following is a non-restrictive description of the energy-saving solar exhaust fan provided according to at least one embodiment of the present application through several examples or embodiments. As described below, different features in these specific examples or embodiments can be combined with each other without conflicting with each other to obtain new examples or embodiments, and these new examples or embodiments also fall within the scope of protection of the present application.

[0070] This application proposes an energy-saving solar exhaust fan, such as Figure 1 As shown, it includes: a housing 10, a motor 11 installed in the housing 10, a fan blade 12 installed at the power output end of the motor 11, a solar panel 13 installed in the housing 10 and exposed to the outside, and an energy storage battery 14 installed in the housing 10 and electrically connected to the solar panel 13, wherein the energy storage battery 14 is electrically connected to the motor 11; an air outlet 15 is provided at the upper end of the housing 10. Further, it also includes: a temperature sensor 20, wherein the temperature sensor 20 is installed in the housing 10; a battery remaining power detector 30, wherein the battery remaining power detector 30 is used to detect the remaining power of the energy storage battery 14; and a motor central controller 100, wherein the motor central controller 100 is communicatively connected to the temperature sensor 20, the battery remaining power detector 30 and the motor 11, wherein the motor central controller 100 is used to control the speed of the motor 11 based on the data collected by the temperature sensor 20 and the battery remaining power detector 30.

[0071] Correspondingly, in the above motor central controller 100, the time series of the ambient temperature collected by the temperature sensor 20 and the time series of the battery remaining power collected by the battery remaining power detector 30 are used, and the artificial intelligence technology is used to perform feature encoding and decoding on the ambient temperature and the battery remaining power respectively to obtain the ambient temperature guided speed recommendation value and the remaining power guided speed recommendation value, wherein the ambient temperature guided speed recommendation value and the remaining power guided speed recommendation value respectively represent the speed recommendation value guided by the ambient temperature time series feature and the speed recommendation value guided by the remaining power of the solar exhaust fan. Then, the average of the ambient temperature guided speed recommendation value and the remaining power guided speed recommendation value is used as the energy-saving comprehensive guided speed recommendation value, and the motor speed adjustment instruction is generated based on the comparison between the energy-saving comprehensive guided speed recommendation value and the current speed value to adapt to the ventilation requirements under different environmental characteristics, and the real-time power of the solar panel 13 is taken into account while meeting the ventilation requirements to extend the service life of the energy storage battery 14. This solar exhaust fan control method is more intelligent and can automatically adapt to changes in environmental conditions to make corresponding adjustments, thereby avoiding the problem of the energy storage battery 14 running for a long time in a low-power state and shortening the battery life.

[0072] Based on this, Figure 2As shown, the motor central controller 100 includes: an ambient temperature acquisition module 110, which is used to obtain the time series of the ambient temperature collected by the temperature sensor; a battery remaining power acquisition module 120, which is used to obtain the time series of the battery remaining power collected by the battery remaining power detector; an ambient temperature local time series association module 130, which is used to input the time series of the ambient temperature into a first-order sequence encoder based on a 1D-CNN model to obtain a sequence of ambient temperature local time series association implicit feature vectors; an ambient temperature feature enhancement module 140, which is used to input the sequence of ambient temperature local time series association implicit feature vectors into a reverse screening enhancement network based on the feature distribution of the anchor center to obtain a sequence of enhanced ambient temperature local time series association implicit feature vectors; an ambient temperature global time series association module 150, which is used to input the sequence of enhanced ambient temperature local time series association implicit feature vectors into a first-order sequence encoder based on a 1D-CNN model to obtain a sequence of enhanced ambient temperature local time series association implicit feature vectors. A second-order sequence encoder based on a bidirectional long short-term memory neural network model is used to obtain an ambient temperature global time series associated implicit feature vector; a battery remaining power feature processing module 160 is used to pass the time series of the battery remaining power through the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center and the second-order sequence encoder based on the bidirectional long short-term memory neural network model to obtain a remaining power global time series associated implicit feature vector; a speed recommendation module 170 is used to input the remaining power global time series associated implicit feature vector and the ambient temperature global time series associated implicit feature vector into a speed recommender respectively to obtain an ambient temperature guided speed recommendation value and a remaining power guided speed recommendation value; a mediation instruction generation module 180 is used to generate a motor speed mediation instruction based on the energy-saving comprehensive guided speed recommendation value and the current speed value of the motor.

[0073] Specifically, in the technical solution of the present application, first, the time series of the ambient temperature collected by the temperature sensor is obtained, and the time series of the remaining battery power collected by the battery remaining power detector is obtained. Next, considering that the time series of the ambient temperature contains local temporal feature information about the temperature environment, and the 1D-CNN model can effectively capture the local patterns and features in the time series data, and extract the data time series features at the local time scale, which helps to better understand and represent the time series information of the ambient temperature. Based on this, in the technical solution of the present application, the time series of the ambient temperature is input into a first-order sequence encoder based on the 1D-CNN model to capture and mine the local time series patterns in the time series of the ambient temperature, thereby obtaining a sequence of implicit feature vectors associated with the local time series of the ambient temperature.

[0074] Next, considering that different ambient temperature local time series associated implicit features in the sequence of the ambient temperature local time series associated implicit feature vectors express ambient temperature information to different degrees, there are ambient temperature local time series associated features that are more critical to motor speed control, but there are also features that have lower correlation with the overall temperature correlation. Therefore, in order to filter out key and important ambient temperature local time series associated feature information from the sequence of the ambient temperature local time series associated implicit feature vectors and reduce the influence of irrelevant features, so as to strengthen the ambient temperature local time series associated implicit features, in the technical solution of the present application, the sequence of the ambient temperature local time series associated implicit feature vectors is input into a reverse screening and strengthening network based on the feature distribution of the anchor center to obtain a sequence of strengthened ambient temperature local time series associated implicit feature vectors. Specifically, firstly, the mean vector of the sequence of the ambient temperature local time series associated implicit feature vectors is calculated to provide the central trend of the feature distribution of the entire ambient temperature sequence, understand the general pattern and behavior of the ambient temperature data, and obtain the central vector of the ambient temperature sequence distribution. Next, the correlation between the center vector of the ambient temperature sequence distribution and the implicit feature vector associated with each local time series of the ambient temperature is calculated to help screen out the local feature information that is strongly associated with the overall distribution characteristics of the ambient temperature to obtain a sequence of implicit correlation coefficients. Then, the inverse of each implicit correlation coefficient is calculated to obtain a sequence of anti-correlation coefficients, so that the implicit feature vectors with higher correlation are given a greater weight, and the feature vectors with lower correlation are given a smaller weight, in preparation for subsequent reverse screening. Then, the Softmax function is used to normalize each anti-correlation coefficient obtained, so that the sum of the weights of different features is 1, so as to better understand the relative importance of each feature in the model, thereby enhancing the stability of the model. Finally, each normalized anti-correlation energy coefficient is used as the weight value to weight each ambient temperature local time series associated implicit feature vector, and the difference between the sequence of the ambient temperature local time series associated implicit feature vector and the sequence of the weighted feature vector is calculated, so as to reversely screen out the ambient temperature local time series features that contribute more to the performance of the model, while reducing the interference of noise with lower correlation, and obtain the sequence of enhanced ambient temperature local time series associated implicit feature vectors after feature enhancement.

[0075] Accordingly, if Figure 3As shown, the ambient temperature feature enhancement module 140 includes: an ambient temperature sequence class label calculation unit 141, which is used to calculate the mean vector of the sequence of the ambient temperature local time series associated implicit feature vectors to obtain the ambient temperature sequence distribution center vector; an ambient temperature local time series implicit correlation coefficient calculation unit 142, which is used to calculate the correlation coefficient between each ambient temperature local time series associated implicit feature vector in the sequence of the ambient temperature local time series associated implicit feature vectors and the ambient temperature sequence distribution center vector to obtain the sequence of ambient temperature local time series implicit correlation coefficients; an anti-correlation coefficient calculation unit 143, which is used to calculate the inverse of each ambient temperature local time series implicit correlation coefficient in the sequence of the ambient temperature local time series implicit correlation coefficients to obtain the sequence of ambient temperature local time series anti-correlation coefficients; a normalization processing unit 144, which is used to use the Softmax function to perform anti-correlation on the ambient temperature local time series. The sequence of coefficients is normalized to obtain a sequence of ambient temperature local time series anti-correlation energy coefficients; a feature weighting unit 145 is used to weight each ambient temperature local time series associated implicit feature vector in the sequence of ambient temperature local time series associated implicit feature vectors using each ambient temperature local time series anti-correlation energy coefficient in the sequence of ambient temperature local time series anti-correlation energy coefficients as a weight to obtain a sequence of reverse inhibition ambient temperature local time series associated implicit feature vectors; a difference calculation unit 146 is used to calculate the positional difference of each corresponding ambient temperature local time series associated implicit feature vector and the reverse inhibition ambient temperature local time series associated implicit feature vector between the sequence of ambient temperature local time series associated implicit feature vectors and the sequence of reverse inhibition ambient temperature local time series associated implicit feature vectors to obtain a sequence of enhanced ambient temperature local time series associated implicit feature vectors.

[0076] Among them, the ambient temperature local time series implicit correlation coefficient calculation unit 142 is used to: calculate the point-by-point multiplication between the ambient temperature local time series association implicit feature vector and the ambient temperature sequence distribution center vector to obtain the ambient temperature local time series association implicit class support feature vector; calculate the sum of the feature values ​​of each position in the ambient temperature local time series association implicit class support feature vector to obtain the ambient temperature local time series association implicit class support sum value; calculate the square sum of the feature values ​​of each position in the ambient temperature local time series association implicit feature vector and the square sum of the feature values ​​of each position in the ambient temperature sequence distribution center vector to obtain the ambient temperature local time series association implicit square value and the ambient temperature sequence distribution center square value; calculate the square root of the sum of the ambient temperature local time series association implicit square value and the ambient temperature sequence distribution center square value to obtain the ambient temperature correlation value; divide the ambient temperature local time series association implicit class support sum value by the ambient temperature correlation value to obtain the ambient temperature local time series implicit correlation coefficient.

[0077] Among them, the normalization processing unit 144 is used to: use each ambient temperature local timing anti-correlation coefficient of the sequence of ambient temperature local timing anti-correlation coefficients as the exponent of the natural constant to calculate the exponential function value with the natural constant as the base according to the position to obtain the sequence of ambient temperature local timing anti-correlation exponential coefficients; calculate the sum of each ambient temperature local timing anti-correlation exponential coefficient in the sequence of ambient temperature local timing anti-correlation exponential coefficients to obtain the total sum of ambient temperature local timing anti-correlation coefficients; calculate the sum of the total sum of ambient temperature local timing anti-correlation coefficients and the constant one to obtain the ambient temperature local timing anti-correlation modulation coefficient; divide each ambient temperature local timing anti-correlation exponential coefficient in the sequence of ambient temperature local timing anti-correlation exponential coefficients by the ambient temperature local timing anti-correlation modulation coefficient to obtain the sequence of ambient temperature local timing anti-correlation energy coefficients.

[0078] In a specific example, the ambient temperature feature enhancement module 140 is used to: input the sequence of the ambient temperature local temporal correlation implicit feature vectors into the feature distribution reverse screening enhancement network based on the anchor center, and process it with the following enhancement formula to obtain the sequence of the enhanced ambient temperature local temporal correlation implicit feature vectors; wherein the enhancement formula is:

[0079] X={x (1) ,x (2) ,…,x (N)}

[0080]

[0081] Z={z (1) ,z (2) ,…,z (N)}, z (i) =x (i) -y (i)

[0082] Wherein, X is the sequence of the implicit feature vectors associated with the local time series of the ambient temperature, N is the number of feature vectors in the sequence of the implicit feature vectors associated with the local time series of the ambient temperature, and x (i) is the i-th ambient temperature local time series associated implicit feature vector in the sequence of the ambient temperature local time series associated implicit feature vectors, V r is the center vector of the ambient temperature sequence distribution, ⊙ is the point multiplication by position, L is the number of eigenvalues ​​in each of the ambient temperature local time series associated implicit eigenvectors, is the eigenvalue of the jth position in each of the local temporal correlation implicit eigenvectors of the ambient temperature, is the eigenvalue of the jth position in the center vector of the ambient temperature sequence distribution, r(i) is the i-th ambient temperature local time series implicit correlation coefficient in the sequence of the ambient temperature local time series implicit correlation coefficient, ρ (i) is the i-th ambient temperature local time series anti-correlation coefficient in the sequence of the ambient temperature local time series anti-correlation coefficients, w (i) is the i-th ambient temperature local time series anti-correlation energy coefficient in the sequence of the ambient temperature local time series anti-correlation energy coefficients, Y is the sequence of the reverse suppression ambient temperature local time series correlation implicit feature vectors, y (i) is the i-th reverse inhibition ambient temperature local time series association implicit feature vector in the sequence of the reverse inhibition ambient temperature local time series association implicit feature vectors, Z is the sequence of the enhanced ambient temperature local time series association implicit feature vectors, z (i) It is the i-th enhanced ambient temperature local time series associated implicit feature vector in the sequence of the enhanced ambient temperature local time series associated implicit feature vectors.

[0083] Then, considering that there are different local temporal interdependencies of ambient temperature between different time scales in the sequence of the enhanced ambient temperature local temporal association implicit feature vector, the bidirectional long short-term memory neural network model can capture the long-term dependencies between different time scales in the time series data, thereby providing an understanding of the overall pattern of different local temporal series of ambient temperature changes. Based on this, in the technical solution of the present application, the sequence of the enhanced ambient temperature local temporal association implicit feature vector is input into a second-order sequence encoder based on the bidirectional long short-term memory neural network model to capture and mine the ambient temperature temporal interdependencies under different time spans, thereby obtaining the ambient temperature global temporal association implicit feature vector.

[0084] Similarly, considering that the time series of the remaining battery power contains the time series feature information of the remaining battery power in the local time, and the time series criticality and importance in different local times are different, the time series correlation effects between different time scales are different. Therefore, in order to better capture and extract the local time series features implied by the time series of the remaining battery power, strengthen and mine the important feature information between the local time series correlations of the remaining power and the mutual dependence between different local times, in the technical solution of the present application, the time series of the remaining battery power is passed through the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center, and the second-order sequence encoder based on the bidirectional long short-term memory neural network model to obtain the implicit feature vector of the global time series correlation of the remaining power.

[0085] Next, it is considered that both the global timing-related implicit feature vector of the remaining power and the global timing-related implicit feature vector of the ambient temperature have an influence on the regulation of the motor speed. Therefore, in order to better recommend the speed value based on the timing correlation relationship between the two in the full time domain, in the technical solution of the present application, the global timing-related implicit feature vector of the remaining power and the global timing-related implicit feature vector of the ambient temperature are respectively input into the decoder-based speed recommender to obtain the ambient temperature-guided speed recommendation value and the remaining power-guided speed recommendation value. In other words, the decoder-based speed recommender can better understand the timing-related implicit relationship between each global timing-related implicit feature vector and generate a more accurate speed recommendation value.

[0086] Correspondingly, the speed recommendation module 170 is used to: input the remaining power global time series associated implicit feature vector and the ambient temperature global time series associated implicit feature vector into the decoder-based speed recommender to obtain the ambient temperature guided speed recommendation value and the remaining power guided speed recommendation value.

[0087] Furthermore, the average of the ambient temperature guided speed recommendation value and the remaining power guided speed recommendation value is calculated as the energy-saving comprehensive guided speed recommendation value. In this way, these two factors can be comprehensively considered, and a balance point can be found based on these two factors to make the recommended speed more comprehensive and balanced, thereby obtaining a speed that can adapt to the current ambient temperature without excessively consuming the battery.

[0088] Then, based on the energy-saving comprehensive guidance speed recommendation value and the current speed value of the motor, a motor speed adjustment instruction is generated. That is, the motor speed adjustment instruction is automatically generated according to the average value of the drilling speed recommendation guided by the ambient temperature and the remaining battery power and the current speed value of the motor. In this way, the motor central controller can automatically adjust the motor speed according to the real-time ambient temperature and the remaining battery power to adapt to different ventilation requirements, so as to minimize energy consumption while meeting the ventilation effect, thereby extending the battery life.

[0089] Furthermore, in the technical solution of the present application, the energy-saving solar exhaust fan also includes a training module for training the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center, the second-order sequence encoder based on the bidirectional long short-term memory neural network model, and the decoder-based speed recommender.

[0090] Among them, the training module includes: a training data acquisition unit, which is used to obtain training data, and the training data includes a time series of training environment temperature collected by the temperature sensor and a time series of training battery remaining power collected by the battery remaining power detector; a training environment temperature local time series association unit, which is used to input the time series of the training environment temperature into the first-order sequence encoder based on the 1D-CNN model to obtain a sequence of training environment temperature local time series association implicit feature vectors; a training environment temperature feature enhancement unit, which is used to input the sequence of training environment temperature local time series association implicit feature vectors into the feature distribution reverse screening enhancement network based on the anchor center to obtain a sequence of training enhanced environment temperature local time series association implicit feature vectors; a training environment temperature global time series association unit, which is used to input the sequence of training enhanced environment temperature local time series association implicit feature vectors into the second-order sequence encoder based on the bidirectional long short-term memory neural network model to obtain a training environment temperature global time series association implicit feature vector; a training battery remaining power feature processing unit, which is used to encode the time series of the training battery remaining power through the first-order sequence encoder based on the 1D-CNN model The invention relates to a first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center, and the second-order sequence encoder based on the bidirectional long short-term memory neural network model to obtain the training remaining power global time series associated implicit feature vector; a decoding loss function value calculation unit, which is used to input the training remaining power global time series associated implicit feature vector and the training ambient temperature global time series associated implicit feature vector into the speed recommender based on the decoder to obtain the ambient temperature decoding loss function value and the remaining power decoding loss function value; a predetermined loss function value calculation unit, which is used to calculate the predetermined loss function value of the training ambient temperature global time series associated implicit feature vector to obtain the ambient temperature global time series associated implicit loss function value; a model training unit, which is used to use the weighted sum of the ambient temperature decoding loss function value, the remaining power decoding loss function value and the ambient temperature global time series associated implicit loss function value as the loss function value and train the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center, the second-order sequence encoder based on the bidirectional long short-term memory neural network model, and the speed recommender based on the decoder through back propagation of gradient descent.

[0091] Here, the sequence of the local temporal association implicit feature vectors of the training environment temperature expresses the one-dimensional local temporal association features of the training environment temperature in the local time domain under the local time domain. In this way, after further reverse screening and reinforcement based on the feature distribution of the anchor center and local time domain association encoding of the second-order sequence encoder based on the bidirectional long short-term memory neural network model, the global temporal association implicit feature vector of the training environment temperature will have feature aggregation decoding regression representation imbalance caused by the difference in feature reinforcement weight distribution caused by the unbalanced temporal feature distribution of each local time domain as a node, that is, the global temporal association implicit feature vector of the training environment temperature has a more significant local feature manifold distribution variability as a whole feature distribution, which makes it difficult to decode regression recognition and affects the training efficiency of decoding training.

[0092] Therefore, the applicant of the present application further introduces a predetermined loss function value other than the decoding loss function value during the model training process, that is, training the model through gradient back propagation based on the loss function value includes the following steps: calculating a first ambient temperature global temporal association implicit weight matrix and a second ambient temperature global temporal association implicit weight matrix based on the training ambient temperature global temporal association implicit feature vector, wherein the eigenvalues ​​at the (i, j)th positions of the first and second ambient temperature global temporal association implicit weight matrices are respectively half of the mean and absolute value of the difference of the i-th eigenvalue and the j-th eigenvalue of the training ambient temperature global temporal association implicit feature vector; performing query matrix multiplication of the training ambient temperature global temporal association implicit feature vector with the first and second ambient temperature global temporal association implicit weight matrices to obtain the first ambient temperature global temporal association implicit feature vector; The intermediate vector and the second ambient temperature global timing association implicit intermediate vector; calculate the vector inner product of the first ambient temperature global timing association implicit intermediate vector and the second ambient temperature global timing association implicit intermediate vector to obtain the first ambient temperature global timing association implicit loss term; matrix multiply the first ambient temperature global timing association implicit weight matrix and the second ambient temperature global timing association implicit weight matrix, and calculate the F norm of the result matrix to obtain the second ambient temperature global timing association implicit loss term; subtract the product of the predetermined weight hyperparameter and the second ambient temperature global timing association implicit loss term from the first ambient temperature global timing association implicit loss term to obtain the ambient temperature global timing association implicit loss function value; based on the weighted sum of the ambient temperature global timing association implicit loss function value and the decoding loss function value, optimize the model parameters through gradient back propagation.

[0093] The calculation process of the ambient temperature global time series associated implicit loss function value is specifically expressed by the following loss calculation formula:

[0094]

[0095]

[0096] Wherein, V is the global temporal correlation implicit feature vector of the training environment temperature, M μ and M σ are the first ambient temperature global temporal correlation implicit weight matrix and the second ambient temperature global temporal correlation implicit weight matrix, M μ (i,j) and M σ (i, j) are the eigenvalues ​​of the (i, j)th position of the first ambient temperature global temporal association implicit weight matrix and the second ambient temperature global temporal association implicit weight matrix, respectively. i and v j are respectively the i-th eigenvalue and the j-th eigenvalue of the global time series associated implicit eigenvector of the training environment temperature, is matrix multiplication, ||·|| F is the Frobenius norm of the matrix, α is a predefined weight hyperparameter, is the implicit loss function value associated with the global time series of ambient temperature.

[0097] That is, the global temporal association implicit loss function value of the ambient temperature is represented by the structured feature representation of the short-range and long-range cross-scale detail links of the training global temporal association implicit feature vector to perform the query composition of the detail inner product space in the training global temporal association implicit feature vector, so as to approximate the low-rank independent observable composition of the link detail composition provided by the structured detail interaction of the training global temporal association implicit feature vector of the ambient temperature. In this way, by training with the global temporal association implicit loss function value of the ambient temperature, the distributed detail group of the training global temporal association implicit feature vector of the ambient temperature can be used to decompose the detail group on the basis of detail complexity, so as to promote the decoding regression decomposition recognition of the complex feature manifold structure of the training global temporal association implicit feature vector of the ambient temperature, and improve the decoding training efficiency. In this way, the motor central controller can automatically adjust the motor speed according to the real-time ambient temperature and the remaining battery power to adapt to different ventilation requirements, so as to minimize energy consumption while meeting the ventilation effect, thereby extending the battery life.

[0098] Based on the above embodiments, see Figure 4 As shown, it is a flow chart of an energy-saving solar exhaust fan control method in an embodiment of the present application. The energy-saving solar exhaust fan control method is used to control the energy-saving solar exhaust fan as described above. Figure 4As shown, according to the energy-saving solar exhaust fan control method of the embodiment of the present application, the steps include: S110, obtaining the time series of the ambient temperature collected by the temperature sensor; S120, obtaining the time series of the battery remaining power collected by the battery remaining power detector; S130, inputting the time series of the ambient temperature into a first-order sequence encoder based on the 1D-CNN model to obtain a sequence of local temporal association implicit feature vectors of the ambient temperature; S140, inputting the sequence of local temporal association implicit feature vectors of the ambient temperature into a reverse screening reinforcement network based on the feature distribution of the anchor center to obtain a sequence of enhanced local temporal association implicit feature vectors of the ambient temperature; S150, inputting the sequence of enhanced local temporal association implicit feature vectors of the ambient temperature into a network based on the bidirectional long short-term memory. The second-order sequence encoder of the neural network model is used to obtain the global time series associated implicit feature vector of the ambient temperature; S160, the time series of the remaining battery power is passed through the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center and the second-order sequence encoder based on the bidirectional long short-term memory neural network model to obtain the global time series associated implicit feature vector of the remaining power; S170, the global time series associated implicit feature vector of the remaining power and the global time series associated implicit feature vector of the ambient temperature are respectively input into the speed recommender to obtain the ambient temperature guided speed recommendation value and the remaining power guided speed recommendation value; S180, based on the energy-saving comprehensive guided speed recommendation value and the current speed value of the motor, a motor speed adjustment instruction is generated.

[0099] Here, those skilled in the art will appreciate that the specific operations of each step in the above energy-saving solar exhaust fan control method have been described in detail above. Figure 2 to Figure 3 The energy-saving solar exhaust fan has been described in detail, and therefore, its repeated description will be omitted.

[0100] Based on the above embodiments, another exemplary embodiment of an electronic device is also provided in the embodiments of the present application. In some possible implementations, the electronic device in the embodiments of the present application may include a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor can implement the steps of the energy-saving solar exhaust fan control method in the above embodiments when executing the program.

[0101] The embodiment of the present application also provides a computer-readable storage medium, on which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, the energy-saving solar exhaust fan control method according to the embodiment of the present application described with reference to the above figures can be executed. The computer-readable storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. The volatile memory may, for example, include a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may, for example, include a read-only memory (ROM), a hard disk, a flash memory, etc.

[0102] The embodiment of the present application also provides a computer program product or a computer program, which includes computer executable instructions, and the computer executable instructions are stored in a computer readable storage medium. The processor of the computer device reads the computer executable instructions from the computer readable storage medium, and the processor executes the computer executable instructions, so that the computer device executes the energy-saving solar exhaust fan control method according to the embodiment of the present application.

[0103] Those skilled in the art will appreciate that the contents disclosed in this application may be subject to various variations and improvements. For example, the various devices or components described above may be implemented by hardware, or by software, firmware, or a combination of some or all of the three.

[0104] In addition, although the present application makes various references to certain units in the system according to embodiments of the present application, any number of different units can be used and run on the client and / or server. The units are only illustrative, and different aspects of the system and method can use different units.

[0105] Those skilled in the art will appreciate that all or part of the steps in the above method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk or an optical disk. Optionally, all or part of the steps in the above embodiment can also be implemented using one or more integrated circuits. Accordingly, each module / unit in the above embodiment can be implemented in the form of hardware or in the form of software function modules. The present application is not limited to any particular form of combination of hardware and software.

[0106] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art to which this application belongs. It should also be understood that terms such as those defined in common dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an idealized or extremely formal sense, unless explicitly defined as such herein.

[0107] The above is an explanation of the present application and should not be considered as limiting thereof. Although several exemplary embodiments of the present application have been described, those skilled in the art will readily appreciate that many modifications may be made to the exemplary embodiments without departing from the novel teachings and advantages of the present application.

Claims

1. An energy-saving solar exhaust fan, comprising: A housing, a motor installed in the housing, a fan blade installed at the power output end of the motor, a solar panel installed in the housing and exposed to the outside, an energy storage battery installed in the housing and electrically connected to the solar panel, the energy storage battery being electrically connected to the motor; an air outlet is provided at the upper end of the housing, and the housing further comprises: a temperature sensor, wherein the temperature sensor is mounted on the housing; A battery remaining power detector, which is used to detect the remaining power of the energy storage battery; a motor central controller, the motor central controller being communicatively connected to the temperature sensor, the battery remaining capacity detector and the motor, wherein the motor central controller is used to control the rotation speed of the motor based on data collected by the temperature sensor and the battery remaining capacity detector; The motor central controller comprises: An ambient temperature acquisition module, used to acquire a time series of ambient temperature acquired by the temperature sensor; A battery remaining power acquisition module, used to acquire the time series of the battery remaining power acquired by the battery remaining power detector; An ambient temperature local time series association module, used for inputting the time series of the ambient temperature into a first-order sequence encoder based on a 1D-CNN model to obtain a sequence of ambient temperature local time series association implicit feature vectors; An ambient temperature feature enhancement module, used for inputting the sequence of the ambient temperature local temporal correlation implicit feature vectors into a reverse screening enhancement network based on the feature distribution of the anchor center to obtain a sequence of enhanced ambient temperature local temporal correlation implicit feature vectors; An ambient temperature global temporal association module is used to input the sequence of the enhanced ambient temperature local temporal association implicit feature vectors into a second-order sequence encoder based on a bidirectional long short-term memory neural network model to obtain an ambient temperature global temporal association implicit feature vector; A battery remaining power feature processing module, used for passing the time series of the battery remaining power through the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center, and the second-order sequence encoder based on the bidirectional long short-term memory neural network model to obtain a global time series associated implicit feature vector of the remaining power; A speed recommendation module, used for inputting the remaining power global time series associated implicit feature vector and the ambient temperature global time series associated implicit feature vector into a speed recommender respectively to obtain an ambient temperature guided speed recommendation value and a remaining power guided speed recommendation value, and calculating an average of the ambient temperature guided speed recommendation value and the remaining power guided speed recommendation value as an energy-saving comprehensive guided speed recommendation value; A mediation instruction generation module, used for generating a motor speed mediation instruction based on the energy-saving comprehensive guided speed recommendation value and the current speed value of the motor; The ambient temperature feature enhancement module includes: An ambient temperature sequence class label calculation unit, used to calculate the mean vector of the sequence of the ambient temperature local time series associated implicit feature vectors to obtain the ambient temperature sequence distribution center vector; An ambient temperature local time series implicit correlation coefficient calculation unit is used to calculate the correlation coefficient between each ambient temperature local time series implicit characteristic vector in the sequence of the ambient temperature local time series associated implicit characteristic vectors and the ambient temperature sequence distribution center vector to obtain a sequence of ambient temperature local time series implicit correlation coefficients; An anti-correlation coefficient calculation unit, used for calculating the reciprocal of each ambient temperature local time series implicit correlation coefficient in the sequence of ambient temperature local time series implicit correlation coefficients to obtain a sequence of ambient temperature local time series anti-correlation coefficients; A normalization processing unit, used for performing normalization processing on the sequence of the local time series anti-correlation coefficients of the ambient temperature using a Softmax function to obtain a sequence of the local time series anti-correlation energy coefficients of the ambient temperature; A feature weighting unit, used to weight each ambient temperature local time series associated implicit feature vector in the sequence of ambient temperature local time series associated implicit feature vectors by using each ambient temperature local time series anti-correlation energy coefficient in the sequence of ambient temperature local time series anti-correlation energy coefficients as a weight to obtain a sequence of reversely inhibited ambient temperature local time series associated implicit feature vectors; A differential calculation unit is used to calculate the positional difference of each corresponding sequence of the ambient temperature local timing association implicit feature vector and the reverse inhibition ambient temperature local timing association implicit feature vector between the sequence of the ambient temperature local timing association implicit feature vector and the sequence of the reverse inhibition ambient temperature local timing association implicit feature vector to obtain the sequence of the enhanced ambient temperature local timing association implicit feature vector.

2. The energy-saving solar exhaust fan according to claim 1, characterized in that: The ambient temperature local time series implicit correlation coefficient calculation unit is used to: Calculate the point-by-point multiplication between the ambient temperature local time series associated implicit feature vector and the ambient temperature sequence distribution center vector to obtain the ambient temperature local time series associated implicit class support feature vector; Calculate the sum of the eigenvalues ​​of each position in the ambient temperature local temporal association implicit class support eigenvector to obtain the ambient temperature local temporal association implicit class support sum; Calculate the sum of squares of the eigenvalues ​​at each position in the ambient temperature local time series associated implicit eigenvector and the sum of squares of the eigenvalues ​​at each position in the ambient temperature sequence distribution center vector to obtain the ambient temperature local time series associated implicit square value and the ambient temperature sequence distribution center square value; Calculate the square root of the sum of the implicit square value of the local temporal correlation of the ambient temperature and the central square value of the ambient temperature sequence distribution to obtain an ambient temperature correlation value; The ambient temperature local timing association implicit class support sum value is divided by the ambient temperature association value to obtain the ambient temperature local timing implicit association coefficient.

3. The energy-saving solar exhaust fan according to claim 2, characterized in that: The normalization processing unit is used for: Using each local temporal anti-correlation coefficient of the sequence of local temporal anti-correlation coefficients of ambient temperature as an exponent of a natural constant to calculate an exponential function value based on the natural constant according to the position to obtain a sequence of local temporal anti-correlation exponential coefficients of ambient temperature; Calculating the sum of each ambient temperature local time series anti-correlation index coefficient in the sequence of ambient temperature local time series anti-correlation index coefficients to obtain a total sum of the ambient temperature local time series anti-correlation coefficients; Calculate the sum of the total sum of the ambient temperature local temporal anti-correlation coefficients and a constant one to obtain an ambient temperature local temporal anti-correlation modulation coefficient; Each ambient temperature local time series anti-correlation index coefficient in the sequence of ambient temperature local time series anti-correlation index coefficients is divided by the ambient temperature local time series anti-correlation modulation coefficient to obtain the sequence of ambient temperature local time series anti-correlation energy coefficients.

4. The energy-saving solar exhaust fan according to claim 3, characterized in that: The speed recommendation module is used to: input the remaining power global time series associated implicit feature vector and the ambient temperature global time series associated implicit feature vector into a decoder-based speed recommender to obtain the ambient temperature guided speed recommendation value and the remaining power guided speed recommendation value.

5. The energy-saving solar exhaust fan according to claim 4, characterized in that: It also includes a training module for training the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center, the second-order sequence encoder based on the bidirectional long short-term memory neural network model, and the speed recommender based on the decoder.

6. The energy-saving solar exhaust fan according to claim 5, characterized in that: The training module comprises: A training data acquisition unit, used to acquire training data, wherein the training data includes a time series of a training environment temperature acquired by the temperature sensor and a time series of a training battery remaining capacity acquired by the battery remaining capacity detector; A training environment temperature local time series association unit, used for inputting the time series of the training environment temperature into the first-order sequence encoder based on the 1D-CNN model to obtain a sequence of implicit feature vectors of the training environment temperature local time series association; A training environment temperature feature enhancement unit, used for inputting the sequence of the training environment temperature local time series associated implicit feature vectors into the anchor center-based feature distribution reverse screening enhancement network to obtain a sequence of training enhanced environment temperature local time series associated implicit feature vectors; A training ambient temperature global temporal association unit is used to input the sequence of the training enhanced ambient temperature local temporal association implicit feature vectors into the second-order sequence encoder based on the bidirectional long short-term memory neural network model to obtain the training ambient temperature global temporal association implicit feature vector; A training battery remaining power feature processing unit, used for passing the time series of the training battery remaining power through the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center, and the second-order sequence encoder based on the bidirectional long short-term memory neural network model to obtain a global time series associated implicit feature vector of the training remaining power; A decoding loss function value calculation unit, used for inputting the training remaining power global time series associated implicit feature vector and the training ambient temperature global time series associated implicit feature vector into the decoder-based speed recommender to obtain an ambient temperature decoding loss function value and a remaining power decoding loss function value; A predetermined loss function value calculation unit, used to calculate the predetermined loss function value of the training ambient temperature global time series associated implicit feature vector to obtain the ambient temperature global time series associated implicit loss function value; A model training unit is used to use the weighted sum of the ambient temperature decoding loss function value, the remaining power decoding loss function value and the ambient temperature global time series associated implicit loss function value as the loss function value and train the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center, the second-order sequence encoder based on the bidirectional long short-term memory neural network model and the decoder-based speed recommender through gradient descent back propagation.

7. An energy-saving solar exhaust fan control method, characterized in that: The energy-saving solar exhaust fan control method is used to control the energy-saving solar exhaust fan as described in any one of claims 1-6.

8. The energy-saving solar exhaust fan control method according to claim 7, characterized in that: include: Get the time series of ambient temperature collected by the temperature sensor; Obtaining a time series of the remaining battery power collected by the remaining battery power detector; Inputting the time series of the ambient temperature into a first-order sequence encoder based on a 1D-CNN model to obtain a sequence of local temporal correlation implicit feature vectors of the ambient temperature; Inputting the sequence of the local temporal correlation implicit feature vectors of the ambient temperature into a reverse screening reinforcement network based on the feature distribution of the anchor center to obtain a sequence of the local temporal correlation implicit feature vectors of the enhanced ambient temperature; Inputting the sequence of the enhanced ambient temperature local temporal correlation implicit feature vectors into a second-order sequence encoder based on a bidirectional long short-term memory neural network model to obtain the ambient temperature global temporal correlation implicit feature vector; The time series of the remaining battery power is passed through the first-order sequence encoder based on the 1D-CNN model, the feature distribution reverse screening reinforcement network based on the anchor center, and the second-order sequence encoder based on the bidirectional long short-term memory neural network model to obtain a global time series associated implicit feature vector of the remaining power; Inputting the remaining power global time series associated implicit feature vector and the ambient temperature global time series associated implicit feature vector into a speed recommender to obtain an ambient temperature guided speed recommendation value and a remaining power guided speed recommendation value; Based on the energy-saving comprehensive guide speed recommendation value and the current speed value of the motor, a motor speed adjustment instruction is generated.

Images