A control system and method for a silent oxygen concentrator

Abstract

The present application relates to a control system and method for a silent oxygen concentrator. The method comprises: using a deep learning algorithm based on artificial intelligence to extract the time series offset variation law and dynamic multi-scale nonlinear fluctuation mode between the real-time oxygen flow data and the set target value, and using a classifier to construct a compressor controller to intelligently determine how the current compressor speed should be adjusted based on the time series offset variation law and dynamic multi-scale nonlinear fluctuation mode, thereby realizing automatic control of the silent oxygen concentrator.

Description

Technical Field

[0001] The present application relates to the technical field of silent oxygen concentrators, and in particular, to a control system and method for a silent oxygen concentrator. Background Art

[0002] With the growing demand for medical health in modern society, oxygen concentrators, as an important medical device, are becoming more and more valued for their performance and user experience. Oxygen concentrators are devices that use the principle of molecular sieve pressure swing adsorption to increase oxygen concentration by adsorbing nitrogen and other gas components. Their main function is to separate oxygen from the air and provide it to patients who need oxygen therapy.

[0003] Traditional oxygen concentrators are often accompanied by loud noises during operation, which not only affects the patient's treatment experience, but may also interfere with the surrounding environment. As a result, silent oxygen concentrators have emerged, which reduce noise while ensuring the efficiency and purity of oxygen supply, and have become a research hotspot in the field of medical equipment.

[0004] In the prior art, the control system of the oxygen concentrator generally adopts the PID control strategy. However, this control strategy still has some limitations. For example, there may be nonlinear and time-varying characteristics during the operation of the oxygen concentrator, which are difficult to solve by simple parameter adjustment for the traditional PID controller. In addition, the setting and adjustment of these parameters often require a lot of experiments and verifications, which may be very time-consuming and complicated. Therefore, an optimized silent oxygen concentrator control system and method are expected. Summary of the invention

[0005] This summary is provided to introduce concepts in a brief form that will be described in detail in the detailed description below. This summary is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0006] In a first aspect, the present application provides a control method for a silent oxygen concentrator, the method comprising:

[0007] Obtain the time series of real-time oxygen flow values ​​collected by the flow sensor;

[0008] Get the set oxygen flow target value;

[0009] Extracting multi-order offset correlation features between the time series of the real-time oxygen flow value and the set oxygen flow target value to obtain a sequence of local first-order time series correlation feature vectors of the target offset of the oxygen flow value and a sequence of local second-order time series correlation feature vectors of the target offset of the oxygen flow value;

[0010] Input each group of corresponding local first-order time series correlation feature vectors of oxygen flow value target offset and local second-order time series correlation feature vectors of oxygen flow value target offset in the sequence of local first-order time series correlation feature vectors of oxygen flow value target offset and the sequence of local second-order time series correlation feature vectors of oxygen flow value target offset into the feature vector dynamic interactive fusion module based on gated response respectively to obtain a sequence of local multi-order time series correlation fusion feature vectors of oxygen flow value target offset;

[0011] Based on the sequence of the local multi-order time-series correlation fusion feature vectors of the oxygen flow value target offset, a control instruction is determined, and the control instruction is used to indicate that the speed value of the compressor should increase, decrease or remain unchanged.

[0012] Optionally, extracting multi-order offset association features between the time series of the real-time oxygen flow value and the set oxygen flow target value to obtain a sequence of local first-order time series association feature vectors of the target offset of the oxygen flow value and a sequence of local second-order time series association feature vectors of the target offset of the oxygen flow value, including: respectively calculating the difference between each real-time oxygen flow value in the time series of the real-time oxygen flow value and the set oxygen flow target value to obtain a time series of target offsets of the oxygen flow value; inputting the time series of target offsets of the oxygen flow value into a first-order time series encoder to obtain a sequence of local first-order time series association feature vectors of the target offset of the oxygen flow value; inputting the sequence of local time series association feature vectors of the target offset of the oxygen flow value into a second-order time series encoder to obtain a sequence of local second-order time series association feature vectors of the target offset of the oxygen flow value.

[0013] Optionally, the time series of the target offset of the oxygen flow value is input into a first-order temporal encoder to obtain a sequence of local first-order temporal correlation feature vectors of the target offset of the oxygen flow value, including: inputting the time series of the target offset of the oxygen flow value into a first-order temporal encoder based on a 1D-CNN model to obtain a sequence of local first-order temporal correlation feature vectors of the target offset of the oxygen flow value.

[0014] Optionally, the sequence of local temporal association feature vectors of the target offset of oxygen flow value is input into a second-order temporal encoder to obtain a sequence of local secondary temporal association feature vectors of the target offset of oxygen flow value, including: inputting the sequence of local temporal association feature vectors of the target offset of oxygen flow value into a second-order temporal encoder based on a bidirectional long short-term memory neural network model to obtain a sequence of local secondary temporal association feature vectors of the target offset of oxygen flow value.

[0015] Optionally, each group of corresponding oxygen flow value target offset local first-order time series correlation feature vectors and oxygen flow value target offset local second-order time series correlation feature vectors in the sequence of the oxygen flow value target offset local first-order time series correlation feature vectors and the sequence of the oxygen flow value target offset local second-order time series correlation feature vectors are respectively input into the feature vector dynamic interaction fusion module based on gated response to obtain a sequence of oxygen flow value target offset local multi-order time series correlation fusion feature vectors, including: inputting the oxygen flow value target offset local first-order time series correlation feature vectors and the oxygen flow value target offset local second-order time series correlation feature vectors into the gated response-based feature vector dynamic interaction fusion module to obtain a sequence of oxygen flow value target offset local multi-order time series correlation fusion feature vectors. The time series correlation feature vector is input into the feature union module to obtain the local multi-order time series correlation joint feature vector of the oxygen flow value target offset; the local multi-order time series correlation joint feature vector of the oxygen flow value target offset is input into the gated response function to obtain the response gate of the oxygen flow offset information fusion; the response gate of the oxygen flow offset information fusion is used as a weight to calculate the position-weighted sum of the local first-order time series correlation feature vector of the oxygen flow value target offset and the local second-order time series correlation feature vector of the oxygen flow value target offset to obtain the local multi-order time series correlation fusion feature vector of the oxygen flow value target offset.

[0016] Optionally, the local first-order time series correlation feature vector of the oxygen flow value target offset and the local second-order time series correlation feature vector of the oxygen flow value target offset are input into a feature combination module to obtain a local multi-order time series correlation joint feature vector of the oxygen flow value target offset, including: cascading the local first-order time series correlation feature vector of the oxygen flow value target offset and the local second-order time series correlation feature vector of the oxygen flow value target offset to obtain the local multi-order time series correlation joint feature vector of the oxygen flow value target offset.

[0017] Optionally, the local multi-order temporal correlation joint eigenvector of the target offset of the oxygen flow value is input into a gated response function to obtain a response gate for oxygen flow offset information fusion, including: calculating the sum of the matrix product between the weight parameter vector and the local multi-order temporal correlation joint eigenvector of the target offset of the oxygen flow value and the bias term, and performing nonlinear activation on the summed value through a sigmoid function to obtain the response gate for oxygen flow offset information fusion.

[0018] Optionally, the response gate of the oxygen flow offset information fusion is used as a weight to calculate the position-weighted sum of the local first-order time series association feature vector of the oxygen flow value target offset and the local second-order time series association feature vector of the oxygen flow value target offset to obtain the local multi-order time series association fusion feature vector of the oxygen flow value target offset, including: using the response gate of the oxygen flow offset information fusion as the weight of the local first-order time series association feature vector of the oxygen flow value target offset, using the value obtained by subtracting one from the response gate of the oxygen flow offset information fusion as the weight of the local second-order time series association feature vector of the oxygen flow value target offset, and calculating the position-weighted sum of the local first-order time series association feature vector of the oxygen flow value target offset and the local second-order time series association feature vector of the oxygen flow value target offset to obtain the local multi-order time series association fusion feature vector of the oxygen flow value target offset.

[0019] Optionally, based on the sequence of local multi-order timing-related fusion feature vectors of the oxygen flow value target offset, a control instruction is determined, and the control instruction is used to indicate that the speed value of the compressor should increase, decrease or remain unchanged, including: inputting the sequence of local multi-order timing-related fusion feature vectors of the oxygen flow value target offset into a classifier-based compressor controller to obtain the control instruction.

[0020] In a second aspect, the present application provides a control system for a silent oxygen concentrator, the system comprising:

[0021] The oxygen real-time flow value acquisition module is used to acquire the time series of the oxygen real-time flow value collected by the flow sensor;

[0022] An oxygen flow target value acquisition module is used to acquire a set oxygen flow target value;

[0023] A multi-order offset correlation feature extraction module is used to extract the multi-order offset correlation features between the time series of the real-time oxygen flow value and the set oxygen flow target value to obtain a sequence of local first-order time series correlation feature vectors of the target offset of the oxygen flow value and a sequence of local second-order time series correlation feature vectors of the target offset of the oxygen flow value;

[0024] A dynamic interactive fusion module is used to input each group of corresponding oxygen flow value target offset local first-order time series correlation feature vectors and oxygen flow value target offset local second-order time series correlation feature vectors in the sequence of the oxygen flow value target offset local first-order time series correlation feature vectors and the sequence of the oxygen flow value target offset local second-order time series correlation feature vectors into a feature vector dynamic interactive fusion module based on gated response to obtain a sequence of oxygen flow value target offset local multi-order time series correlation fusion feature vectors;

[0025] The control instruction determination module is used to determine the control instruction based on the sequence of the local multi-order time-series correlation fusion feature vectors of the oxygen flow value target offset, and the control instruction is used to indicate that the speed value of the compressor should increase, decrease or remain unchanged.

[0026] By adopting the above technical solution, the timing offset change law and dynamic multi-scale nonlinear fluctuation pattern between the real-time oxygen flow data and the set target value are extracted by using a deep learning algorithm based on artificial intelligence, and a classifier is used to build a compressor controller to intelligently determine how the current compressor speed should be adjusted based on the timing offset change law and the dynamic multi-scale nonlinear fluctuation pattern, thereby realizing automatic control of the silent oxygen concentrator.

[0027] Other features and advantages of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The above and other features, advantages and aspects of the embodiments of the present application will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale. In the drawings:

[0029] Figure 1 The figure is a flow chart of a control method of a silent oxygen concentrator according to an exemplary embodiment.

[0030] Figure 2 It is a schematic diagram of the structure of a small molecular sieve oxygen generator according to an exemplary embodiment.

[0031] Figure 3 It is a block diagram of a control system of a silent oxygen concentrator according to an exemplary embodiment.

[0032] Figure 4 It is a block diagram of an electronic device according to an exemplary embodiment.

[0033] Figure 5 It is an application scenario diagram of a control method for a silent oxygen concentrator according to an exemplary embodiment. DETAILED DESCRIPTION

[0034] The embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be construed as being limited to the embodiments described herein. Instead, these embodiments are provided to provide a more thorough and complete understanding of the present application. It should be understood that the drawings and embodiments of the present application are only for exemplary purposes and are not intended to limit the scope of protection of the present application.

[0035] It should be understood that the various steps described in the method implementation of the present application can be performed in different orders and / or performed in parallel. In addition, the method implementation may include additional steps and / or omit the steps shown. The scope of the present application is not limited in this respect.

[0036] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment". The relevant definitions of other terms will be given in the following description.

[0037] It should be noted that the concepts such as "first" and "second" mentioned in this application are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0038] It should be noted that the modifications of "one" and "plurality" mentioned in the present application are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0039] The names of the messages or information exchanged between multiple devices in the embodiments of the present application are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0040] The specific implementation of the present application is described in detail below with reference to the accompanying drawings.

[0041] Figure 1 is a flow chart of a control method of a silent oxygen concentrator according to an exemplary embodiment. Figure 1 As shown, the method includes:

[0042] Step S101, obtaining a time series of real-time oxygen flow values ​​collected by a flow sensor;

[0043] Step S102, obtaining and setting a target value of oxygen flow rate;

[0044] Step S103, extracting multi-order offset correlation features between the time series of the real-time oxygen flow value and the set oxygen flow target value to obtain a sequence of local first-order time series correlation feature vectors of the target offset of the oxygen flow value and a sequence of local second-order time series correlation feature vectors of the target offset of the oxygen flow value;

[0045] Step S104, inputting each group of corresponding local first-order time series correlation feature vectors of oxygen flow value target offset and local second-order time series correlation feature vectors of oxygen flow value target offset in the sequence of local first-order time series correlation feature vectors of oxygen flow value target offset and the sequence of local second-order time series correlation feature vectors of oxygen flow value target offset into a feature vector dynamic interactive fusion module based on gated response to obtain a sequence of local multi-order time series correlation fusion feature vectors of oxygen flow value target offset;

[0046] Step S105: determining a control instruction based on the sequence of the local multi-order time-series correlation fusion feature vectors of the oxygen flow value target offset, wherein the control instruction is used to indicate that the speed value of the compressor should increase, decrease, or remain unchanged.

[0047] In response to the above technical problems, the technical concept of the present application is: to use a deep learning algorithm based on artificial intelligence to extract the timing offset change law and dynamic multi-scale nonlinear fluctuation pattern between the real-time oxygen flow data and the set target value, and use a classifier to build a compressor controller to intelligently determine how the current compressor speed should be adjusted based on the timing offset change law and dynamic multi-scale nonlinear fluctuation pattern, thereby realizing automatic control of the silent oxygen concentrator.

[0048] Based on this, in the technical solution of the present application, first, the time series of the real-time oxygen flow value collected by the flow sensor is obtained, and the oxygen flow target value is set; and the difference between each real-time oxygen flow value in the time series of the real-time oxygen flow value and the set oxygen flow target value is calculated to obtain the time series of the target offset of the oxygen flow value. Among them, the set oxygen flow target value provides the benchmark control condition of the silent oxygen concentrator during operation, which ensures that the control system understands the oxygen flow level expected to be achieved in actual operation, and compares it with each real-time oxygen flow value in the time series of the real-time oxygen flow value, thereby reflecting the real gap between the actual operation and the expected oxygen flow level, providing an important decision-making basis for subsequent automated control.

[0049] In one embodiment of the present application, extracting multi-order offset association features between the time series of the real-time oxygen flow value and the set oxygen flow target value to obtain a sequence of local first-order time series association feature vectors of the target offset of the oxygen flow value and a sequence of local second-order time series association feature vectors of the target offset of the oxygen flow value includes: respectively calculating the difference between each real-time oxygen flow value in the time series of the real-time oxygen flow value and the set oxygen flow target value to obtain a time series of target offsets of the oxygen flow value; inputting the time series of target offsets of the oxygen flow value into a first-order time series encoder to obtain a sequence of local first-order time series association feature vectors of the target offset of the oxygen flow value; inputting the sequence of local time series association feature vectors of the target offset of the oxygen flow value into a second-order time series encoder to obtain a sequence of local second-order time series association feature vectors of the target offset of the oxygen flow value.

[0050] Considering that the silent oxygen concentrator is affected by many factors during operation, such as changes in environmental conditions such as temperature, humidity, and pressure, these factors will cause dynamic fluctuations in the gas flow value. And there are also nonlinear behaviors in the working process of the oxygen concentrator, such as the adsorption and desorption process of the molecular sieve. This complex time-series nonlinear dynamic fluctuation characteristic is difficult to fit with a traditional linear model. Therefore, in the technical solution of the present application, it is expected to use a multi-scale analysis method to extract the implicit time-series nonlinear fluctuation characteristic pattern contained in the time series of the target offset of the oxygen flow value. Specifically, the time series of the target offset of the oxygen flow value is input into a first-order time series encoder based on the 1D-CNN model to obtain a sequence of local first-order time series correlation feature vectors of the target offset of the oxygen flow value; at the same time, the sequence of local time series correlation feature vectors of the target offset of the oxygen flow value is input into a second-order time series encoder based on a bidirectional long short-term memory neural network model to obtain a sequence of local secondary time series correlation feature vectors of the target offset of the oxygen flow value. That is, the first-order temporal encoder is constructed using the 1D-CNN model to capture the characteristic pattern of oxygen flow offset in the local neighborhood, and the second-order temporal encoder is constructed using the bidirectional long short-term memory neural network model to capture the broader contextual temporal dependency in the time series of the target offset of the oxygen flow value.

[0051] In one embodiment of the present application, the time series of the target offset of the oxygen flow value is input into a first-order temporal encoder to obtain a sequence of local first-order temporal correlation feature vectors of the target offset of the oxygen flow value, including: inputting the time series of the target offset of the oxygen flow value into a first-order temporal encoder based on a 1D-CNN model to obtain a sequence of local first-order temporal correlation feature vectors of the target offset of the oxygen flow value.

[0052] Furthermore, in one embodiment of the present application, the sequence of local temporal association feature vectors of the target offset of oxygen flow value is input into a second-order temporal encoder to obtain a sequence of local secondary temporal association feature vectors of the target offset of oxygen flow value, including: inputting the sequence of local temporal association feature vectors of the target offset of oxygen flow value into a second-order temporal encoder based on a bidirectional long short-term memory neural network model to obtain a sequence of local secondary temporal association feature vectors of the target offset of oxygen flow value.

[0053] It should be understood that the sequence of the local first-order time-series correlation feature vectors of the target offset of oxygen flow value and the sequence of the local second-order time-series correlation feature vectors of the target offset of oxygen flow value respectively express the local trend of the oxygen flow offset characteristic and the change pattern of the two-way mid-range distance dependence, and the fusion of the two can provide more scales and richer feature representations for the automatic control of the silent oxygen concentrator. Specifically, in the technical solution of the present application, each group of corresponding local first-order time-series correlation feature vectors of the target offset of oxygen flow value and local second-order time-series correlation feature vectors of the target offset of oxygen flow value in the sequence of the local first-order time-series correlation feature vectors of the target offset of oxygen flow value and the sequence of the local second-order time-series correlation feature vectors of the target offset of oxygen flow value are respectively input into the feature vector dynamic interactive fusion module based on gated response to obtain a sequence of local multi-order time-series correlation fusion feature vectors of the target offset of oxygen flow value.

[0054] In one embodiment of the present application, each group of corresponding oxygen flow value target offset local first-order time series correlation feature vectors and oxygen flow value target offset local second-order time series correlation feature vectors in the sequence of the oxygen flow value target offset local first-order time series correlation feature vectors and the sequence of the oxygen flow value target offset local second-order time series correlation feature vectors are respectively input into the feature vector dynamic interaction fusion module based on gated response to obtain a sequence of oxygen flow value target offset local multi-order time series correlation fusion feature vectors, including: the oxygen flow value target offset local first-order time series correlation feature vector and the oxygen flow value target offset local second-order time series correlation feature vector are respectively input into the feature vector dynamic interaction fusion module based on gated response to obtain a sequence of oxygen flow value target offset local multi-order time series correlation fusion feature vectors. The local secondary time series correlation feature vector is input into the feature combination module to obtain the local multi-order time series correlation joint feature vector of the oxygen flow value target offset; the local multi-order time series correlation joint feature vector of the oxygen flow value target offset is input into the gated response function to obtain the response gate of the oxygen flow offset information fusion; the response gate of the oxygen flow offset information fusion is used as a weight to calculate the position-weighted sum of the local first-order time series correlation feature vector of the oxygen flow value target offset and the local secondary time series correlation feature vector of the oxygen flow value target offset to obtain the local multi-order time series correlation fusion feature vector of the oxygen flow value target offset.

[0055] Further, in one embodiment of the present application, the local first-order time series correlation feature vector of the oxygen flow value target offset and the local second-order time series correlation feature vector of the oxygen flow value target offset are input into a feature combination module to obtain a local multi-order time series correlation joint feature vector of the oxygen flow value target offset, including: cascading the local first-order time series correlation feature vector of the oxygen flow value target offset and the local second-order time series correlation feature vector of the oxygen flow value target offset to obtain the local multi-order time series correlation joint feature vector of the oxygen flow value target offset.

[0056] Specifically, each group of corresponding oxygen flow value target offset local first-order time series correlation feature vectors and oxygen flow value target offset local second-order time series correlation feature vectors in the sequence of the oxygen flow value target offset local first-order time series correlation feature vectors and the sequence of the oxygen flow value target offset local second-order time series correlation feature vectors are respectively input into the feature vector dynamic interaction fusion module based on gated response to obtain a sequence of oxygen flow value target offset local multi-order time series correlation fusion feature vectors; wherein, the fusion formula is:

[0057]

[0058] Wherein, v1 and v2 are the local first-order temporal correlation feature vector of the target offset of the oxygen flow value and the local second-order temporal correlation feature vector of the target offset of the oxygen flow value, respectively, W and b are the weight parameter vector and the bias term, respectively, σ is the sigmoid function, is the response gate for oxygen flow offset information fusion, * represents dot product, Indicates adding by position, v f It is the local multi-order time series correlation fusion feature vector of the oxygen flow value target offset.

[0059] In particular, the dynamic interactive fusion module of the feature vector based on the gated response effectively integrates the sequence information of the local first-order time series correlation feature vector of the target offset of the oxygen flow value and the sequence information of the local second-order time series correlation feature vector of the target offset of the oxygen flow value through a dynamic interaction and weighting mechanism, so as to optimize the interaction process and comprehensive feature expression of the two. More specifically, the dynamic interactive fusion module of the feature vector based on the gated response firstly integrates each group of corresponding local first-order time series correlation feature vectors of the target offset of the oxygen flow value and the local second-order time series correlation feature vectors of the target offset of the oxygen flow value through a feature union module, and then inputs the preliminarily integrated features into the gated response function to analyze the dynamic characteristics of the input features, thereby obtaining the response gate of the oxygen flow offset information fusion. Here, the introduction of the gated response function enables the model to dynamically adjust the weight of the feature in the fusion process according to its importance, realizes feature selection and weighting, thereby highlighting key features and suppressing noise features, and enhancing the model's ability to capture useful information.

[0060] In one embodiment of the present application, the local multi-order temporal correlation joint feature vector of the target offset of the oxygen flow value is input into a gated response function to obtain a response gate for oxygen flow offset information fusion, including: calculating the sum of the matrix product between the weight parameter vector and the local multi-order temporal correlation joint feature vector of the target offset of the oxygen flow value and the bias term, and performing nonlinear activation on the summed value through a sigmoid function to obtain the response gate for oxygen flow offset information fusion.

[0061] Finally, the response gate of the oxygen flow offset information fusion is used to construct weights, and the oxygen flow value target offset local first-order time series correlation feature vector and the oxygen flow value target offset local second-order time series correlation feature vector are weighted and summed by position to obtain the oxygen flow value target offset local multi-order time series correlation fusion feature vector. This step reflects the dynamic interaction and fusion of feature vectors. Through the weighted sum operation, the information complementarity and enhancement between feature vectors are achieved, and the expression ability of features is improved. Among them, the weighted sum by position not only considers the relative importance between feature vectors, but also considers the contribution of each element within the feature vector, so that the fused feature vector can more comprehensively reflect the characteristics of the original data.

[0062] In one embodiment of the present application, the response gate of the oxygen flow offset information fusion is used as a weight to calculate the position-weighted sum of the local first-order time series association feature vector of the oxygen flow value target offset and the local second-order time series association feature vector of the oxygen flow value target offset to obtain the local multi-order time series association fusion feature vector of the oxygen flow value target offset, including: using the response gate of the oxygen flow offset information fusion as the weight of the local first-order time series association feature vector of the oxygen flow value target offset, subtracting the value of the response gate of the oxygen flow offset information fusion from one as the weight of the local second-order time series association feature vector of the oxygen flow value target offset, and calculating the position-weighted sum of the local first-order time series association feature vector of the oxygen flow value target offset and the local second-order time series association feature vector of the oxygen flow value target offset to obtain the local multi-order time series association fusion feature vector of the oxygen flow value target offset.

[0063] Subsequently, the sequence of the oxygen flow value target offset local multi-order time series correlation fusion feature vectors is input into the classifier-based compressor controller to obtain a control instruction, which is used to indicate that the speed value of the compressor should increase, decrease or remain unchanged.

[0064] In one embodiment of the present application, a control instruction is determined based on a sequence of local multi-order timing-related fusion feature vectors of the oxygen flow value target offset, and the control instruction is used to indicate that the speed value of the compressor should increase, decrease, or remain unchanged, including: inputting the sequence of local multi-order timing-related fusion feature vectors of the oxygen flow value target offset into a classifier-based compressor controller to obtain the control instruction.

[0065] In a preferred embodiment, inputting the sequence of the oxygen flow value target offset local multi-order time series correlation fusion feature vectors into a classifier-based compressor controller to obtain a control instruction includes:

[0066] Probabilistically quantizing each eigenvalue of the sequence of the target offset local multi-order time series associated fusion feature vectors of the oxygen flow value based on a probability activation function, such as a sigmoid function, to obtain a probabilistic sequence of the target offset local multi-order time series associated fusion feature vectors of the oxygen flow value;

[0067] The sequence of the local multi-order time series correlation fusion feature vector of the target deviation of the oxygen flow value is obtained and inputted into the compressor controller based on the classifier to obtain the increase probability value, decrease probability value and unchanged probability value corresponding to the speed value of the compressor that should increase, decrease or remain unchanged;

[0068] After deducting one third of each of the increase probability value, the decrease probability value and the unchanged probability value, the square root of the sum of their squares is calculated, and the square root is added to one third to obtain the control probability value;

[0069] Determine a control cognitive symbol value based on a comparison between each eigenvalue of the sequence of the probabilistic oxygen flow value target offset local multi-order time series associated fusion eigenvector and the control probability value, wherein the control cognitive symbol value is equal to one, zero and negative one in response to the eigenvalue of the sequence of the probabilistic oxygen flow value target offset local multi-order time series associated fusion eigenvector being greater than, equal to and less than the control probability value, respectively;

[0070] Calculate the mean of all eigenvalues ​​of the sequence of the probabilistic oxygen flow value target offset local multi-order time series associated fusion eigenvector to obtain the overall control phase shift value;

[0071] After multiplying each eigenvalue of the sequence of the probabilistic oxygen flow value target offset local multi-order time series associated fusion eigenvector by the control cognitive symbol value and the control overall phase shift value, a weighted difference calculation is performed, and the absolute value is taken to obtain the optimized eigenvalue of the sequence of the probabilistic oxygen flow value target offset local multi-order time series associated fusion eigenvector;

[0072] The sequence of the optimized oxygen flow value target offset local multi-order time series correlation fusion feature vectors composed of the optimized feature values ​​is input into the classifier-based compressor controller to obtain a control instruction.

[0073] Specifically, the optimization steps are expressed as:

[0074]

[0075] Among them, f i and f i ′ are the i-th eigenvalues ​​in the sequence of the probabilistic oxygen flow value target offset local multi-order time series associated fusion feature vector and the sequence of the optimized oxygen flow value target offset local multi-order time series associated fusion feature vector, p1, p2 and p3 represent the increase probability value, decrease probability value and unchanged probability value, respectively, p is the control probability value, sgn(v i , p) is the control recognition symbol value, is the overall phase shift value of the control, α and β are the first weighted hyperparameter and the second weighted hyperparameter respectively, and |·| represents the absolute value.

[0076] Here, the sequence of the local first-order time series association feature vectors of the target offset of oxygen flow value and the sequence of the local second-order time series association feature vectors of the target offset of oxygen flow value respectively express the first-order local time series association features and the second-order bidirectional short-range-long-range time series association features of the target offset of oxygen flow value. Thus, after each group of corresponding local first-order time series association feature vectors of the target offset of oxygen flow value and local second-order time series association feature vectors of the target offset of oxygen flow value in the sequence of the local first-order time series association feature vectors of the target offset of oxygen flow value and the sequence of the local second-order time series association feature vectors of the target offset of oxygen flow value are respectively input into the dynamic interactive fusion module of feature vectors based on gated response, the sequence of the local multi-order time series association fusion feature vectors of the target offset of oxygen flow value obtained will also have the lack of prior-posteriori class probability causal association relative to the source time series distribution due to the dynamic interactive fusion difference of different-order time series distributions under the local time series, thereby affecting the accuracy of the classification result.

[0077] Therefore, in the above-mentioned preferred embodiment, the applicant of the present application obtains the control cognitive phase transformation response of the sequence of local multi-order time-series associated fusion feature vectors of the target offset of oxygen flow value by comparing the probabilistic amplitude of the feature value of the sequence of local multi-order time-series associated fusion feature vectors of the target offset of oxygen flow value with the control probability representation, and performs a feature distribution sequence invariance transformation based on the expansion of the control label probability differential distribution on the phase shift response of the feature value of the sequence of local multi-order time-series associated fusion feature vectors of the target offset of oxygen flow value relative to the control probability representation of the feature as a whole, so as to realize the causal constraint of the posterior class probability of the sequence of local multi-order time-series associated fusion feature vectors of the target offset of oxygen flow value on its prior feature distribution representation, so as to improve the accuracy of the control instructions obtained by the compressor controller based on the classifier when the sequence of local multi-order time-series associated fusion feature vectors of the target offset of oxygen flow value is input.

[0078] In summary, the above scheme is adopted to extract the timing offset change law and dynamic multi-scale nonlinear fluctuation pattern between the real-time oxygen flow data and the set target value by using a deep learning algorithm based on artificial intelligence, and a classifier is used to construct a compressor controller to intelligently determine how the current compressor speed should be adjusted based on the timing offset change law and dynamic multi-scale nonlinear fluctuation pattern, thereby realizing automatic control of the silent oxygen concentrator.

[0079] In one embodiment of the present application, a small molecular sieve oxygen generator is provided, such as Figure 2As shown, the oxygen generator is generally composed of an oxygen generator host, a flow meter, a humidifier and an oxygen concentration status indicator. Among them, 1. The air compressor provides the gas pressure required for adsorption and the raw material (air) for separating oxygen; 2. The air pretreatment system mainly includes gas cooling, water removal, filtration and other functions; 3. The control valve controls the compressed air treated by the air pretreatment system to enter the molecular sieve adsorption tower for periodic adsorption and desorption; 4. The molecular sieve adsorption tower is a closed container tightly filled with molecular sieves. The molecular sieve is used to separate the oxygen in the air by using its selective adsorption characteristics for gas under pressurized conditions; 5. The control and alarm system performs automatic control and fault alarm according to the pre-set working procedures; 6. The output gas processing system mainly refers to the part that collects, filters, regulates pressure, humidifies and processes the oxygen produced by the oxygen generator.

[0080] An oxygen generator is a device that uses the principle of molecular sieve pressure swing adsorption to increase the oxygen concentration by adsorbing nitrogen and other gas components.

[0081] When the equipment is working, compressed air is injected into a closed adsorption tower equipped with a molecular sieve, causing the pressure in the adsorption tower to increase. The molecular sieve absorbs a large amount of nitrogen in the compressed air as the ambient pressure increases, while the oxygen in the compressed air still exists in the form of gas and is collected through a certain pipeline. This process is usually called the "adsorption" process. When the molecular sieve in the container adsorbs nitrogen to the critical state of adsorption saturation, the adsorption tower is blown and decompressed. As the ambient pressure decreases, the molecular sieve's ability to adsorb nitrogen decreases, and the nitrogen is released back to the gas phase from the molecular sieve and discharged as waste gas. This process is usually called "desorption". In order to ensure the continuous and stable output of oxygen, the oxygen generator often uses two (or more) molecular sieve adsorption towers, which are controlled by a rotary separation valve so that one adsorption tower is in the adsorption process while the other adsorption tower is in the desorption process. The two work alternately to complete the continuous oxygen production process.

[0082] Figure 3 FIG. 1 is a block diagram of a control system of a silent oxygen concentrator according to an exemplary embodiment. Figure 3 As shown, the control system 200 includes:

[0083] The oxygen real-time flow value acquisition module 201 is used to acquire the time series of the oxygen real-time flow value collected by the flow sensor;

[0084] The oxygen flow target value acquisition module 202 is used to acquire and set the oxygen flow target value;

[0085] A multi-order offset correlation feature extraction module 203 is used to extract the multi-order offset correlation features between the time series of the real-time oxygen flow value and the set oxygen flow target value to obtain a sequence of local first-order time series correlation feature vectors of the target offset of the oxygen flow value and a sequence of local second-order time series correlation feature vectors of the target offset of the oxygen flow value;

[0086] The dynamic interactive fusion module 204 is used to input each group of corresponding oxygen flow value target offset local first-order time series correlation feature vectors and oxygen flow value target offset local second-order time series correlation feature vectors in the sequence of the oxygen flow value target offset local first-order time series correlation feature vectors and the sequence of the oxygen flow value target offset local second-order time series correlation feature vectors into the feature vector dynamic interactive fusion module based on gated response to obtain a sequence of oxygen flow value target offset local multi-order time series correlation fusion feature vectors;

[0087] The control instruction determination module 205 is used to determine the control instruction based on the sequence of the local multi-order time-series correlation fusion feature vectors of the oxygen flow value target offset, and the control instruction is used to indicate that the speed value of the compressor should increase, decrease or remain unchanged.

[0088] Reference below Figure 4 , which shows a schematic diagram of the structure of an electronic device 600 suitable for implementing the embodiment of the present application. The terminal device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 4 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0089] like Figure 4 As shown, the electronic device 600 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 are also stored. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0090] Typically, the following devices may be connected to the I / O interface 605: an input device 606 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 608 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 609. The communication device 609 may allow the electronic device 600 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 4 The electronic device 600 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0091] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 609, or installed from the storage device 608, or installed from the ROM 602. When the computer program is executed by the processing device 601, the above-mentioned functions defined in the method of the embodiment of the present application are executed.

[0092] It should be noted that the computer-readable medium mentioned above in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0093] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0094] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0095] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0096] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0097] The modules involved in the embodiments described in this application can be implemented by software or hardware. The name of the module does not limit the module itself in some cases. For example, the test parameter acquisition module can also be described as a "module for acquiring device test parameters corresponding to the target device".

[0098] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0099] In the context of the present application, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0100] Figure 5 is an application scenario diagram of a control method for a silent oxygen concentrator according to an exemplary embodiment. Figure 5 As shown, in this application scenario, first, a time series of real-time oxygen flow values ​​collected by a flow sensor is obtained (for example, Figure 5 C1 shown in FIG); obtaining a set oxygen flow target value (for example, Figure 5 Then, the obtained oxygen real-time flow value and oxygen flow target value are input to a server (for example, Figure 5 In the S) shown in FIG. 1 , the server is capable of processing the real-time oxygen flow value and the target oxygen flow value based on the control algorithm of the silent oxygen concentrator to determine a control instruction, wherein the control instruction is used to indicate whether the speed value of the compressor should increase, decrease, or remain unchanged.

[0101] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.

[0102] In addition, although each operation is described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or to be performed in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the application. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0103] Although the subject matter has been described in language specific to structural features and / or method logic actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims. Regarding the device in the above embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.

Claims

1. A control method for a silent oxygen concentrator, characterized in that: include: Obtain the time series of real-time oxygen flow values ​​collected by the flow sensor; Get the set oxygen flow target value; Extracting multi-order offset correlation features between the time series of the real-time oxygen flow value and the set oxygen flow target value to obtain a sequence of local first-order time series correlation feature vectors of the target offset of the oxygen flow value and a sequence of local second-order time series correlation feature vectors of the target offset of the oxygen flow value; Input each group of corresponding local first-order time series correlation feature vectors of oxygen flow value target offset and local second-order time series correlation feature vectors of oxygen flow value target offset in the sequence of local first-order time series correlation feature vectors of oxygen flow value target offset and the sequence of local second-order time series correlation feature vectors of oxygen flow value target offset into the feature vector dynamic interactive fusion module based on gated response respectively to obtain a sequence of local multi-order time series correlation fusion feature vectors of oxygen flow value target offset; Determining a control instruction based on a sequence of local multi-order time-series correlation fusion feature vectors of the oxygen flow value target offset, wherein the control instruction is used to indicate that the speed value of the compressor should increase, decrease, or remain unchanged; Among them, extracting the multi-order offset correlation features between the time series of the real-time oxygen flow value and the set oxygen flow target value to obtain a sequence of local first-order time series correlation feature vectors of the target offset of the oxygen flow value and a sequence of local second-order time series correlation feature vectors of the target offset of the oxygen flow value, includes: Calculating the difference between each oxygen real-time flow value in the time series of the oxygen real-time flow value and the set oxygen flow target value to obtain a time series of oxygen flow value target offset; Inputting the time series of the oxygen flow value target offset into a first-order temporal encoder to obtain a sequence of local first-order temporal correlation feature vectors of the oxygen flow value target offset; The sequence of local temporal correlation feature vectors of target offset of oxygen flow value is input into a second-order temporal encoder to obtain the sequence of local secondary temporal correlation feature vectors of target offset of oxygen flow value.

2. The control method of the silent oxygen concentrator according to claim 1, characterized in that: Inputting the time series of the oxygen flow value target offset into a first-order temporal encoder to obtain a sequence of local first-order temporal correlation feature vectors of the oxygen flow value target offset, including: The time series of the oxygen flow value target offset is input into a first-order temporal encoder based on a 1D-CNN model to obtain a sequence of local first-order temporal correlation feature vectors of the oxygen flow value target offset.

3. The control method of the silent oxygen concentrator according to claim 2, characterized in that: Inputting the sequence of the local temporal correlation feature vectors of the target offset of the oxygen flow value into a second-order temporal encoder to obtain the sequence of the local secondary temporal correlation feature vectors of the target offset of the oxygen flow value, including: The sequence of the local temporal correlation feature vectors of the target offset of the oxygen flow value is input into a second-order temporal encoder based on a bidirectional long short-term memory neural network model to obtain a sequence of the local secondary temporal correlation feature vectors of the target offset of the oxygen flow value.

4. The control method of the silent oxygen concentrator according to claim 3, characterized in that: Inputting each group of corresponding oxygen flow value target offset local first-order time series correlation feature vectors and oxygen flow value target offset local second-order time series correlation feature vectors in the sequence of the oxygen flow value target offset local first-order time series correlation feature vectors and the sequence of the oxygen flow value target offset local second-order time series correlation feature vectors into the feature vector dynamic interactive fusion module based on gated response respectively to obtain a sequence of oxygen flow value target offset local multi-order time series correlation fusion feature vectors, including: Inputting the oxygen flow value target offset local first-order time series correlation feature vector and the oxygen flow value target offset local second-order time series correlation feature vector into a feature combination module to obtain the oxygen flow value target offset local multi-order time series correlation combination feature vector; Inputting the local multi-order time series associated joint feature vector of the oxygen flow value target offset into a gated response function to obtain a response gate for oxygen flow offset information fusion; The response gate of the oxygen flow offset information fusion is used as a weight to calculate the position-weighted sum of the local first-order temporal association feature vector of the oxygen flow value target offset and the local second-order temporal association feature vector of the oxygen flow value target offset to obtain the local multi-order temporal association fusion feature vector of the oxygen flow value target offset.

5. The control method of the silent oxygen concentrator according to claim 4, characterized in that: Inputting the oxygen flow value target offset local first-order time series correlation feature vector and the oxygen flow value target offset local second-order time series correlation feature vector into a feature combination module to obtain the oxygen flow value target offset local multi-order time series correlation combination feature vector, including: The oxygen flow value target offset local first-order time series correlation feature vector and the oxygen flow value target offset local second-order time series correlation feature vector are cascaded to obtain the oxygen flow value target offset local multi-order time series correlation joint feature vector.

6. The control method of the silent oxygen concentrator according to claim 5, characterized in that: Inputting the local multi-order time series associated joint feature vector of the oxygen flow value target offset into the gated response function to obtain a response gate for oxygen flow offset information fusion, including: The matrix product between the weight parameter vector and the local multi-order temporal correlation joint eigenvector of the target offset of the oxygen flow value and the sum of the bias term are calculated, and the summed value is nonlinearly activated by a sigmoid function to obtain a response gate for the oxygen flow offset information fusion.

7. The control method of the silent oxygen concentrator according to claim 6, characterized in that: The response gate of the oxygen flow offset information fusion is used as a weight to calculate the position-weighted sum of the local first-order temporal association feature vector of the oxygen flow value target offset and the local second-order temporal association feature vector of the oxygen flow value target offset to obtain the local multi-order temporal association fusion feature vector of the oxygen flow value target offset, including: The response gate of the oxygen flow offset information fusion is used as the weight of the local first-order time series association feature vector of the oxygen flow value target offset, the value obtained by subtracting the response gate of the oxygen flow offset information fusion from one is used as the weight of the local secondary time series association feature vector of the oxygen flow value target offset, and the position-weighted sum of the local first-order time series association feature vector of the oxygen flow value target offset and the local secondary time series association feature vector of the oxygen flow value target offset is calculated to obtain the local multi-order time series association fusion feature vector of the oxygen flow value target offset.

8. The control method of the silent oxygen concentrator according to claim 7, characterized in that: Based on the sequence of the local multi-order time series correlation fusion feature vectors of the oxygen flow value target offset, a control instruction is determined, where the control instruction is used to indicate that the speed value of the compressor should increase, decrease, or remain unchanged, including: The sequence of the oxygen flow value target offset local multi-order time series correlation fusion feature vectors is input into a classifier-based compressor controller to obtain the control instruction.

9. A control system for a silent oxygen concentrator, characterized in that: include: The oxygen real-time flow value acquisition module is used to acquire the time series of the oxygen real-time flow value collected by the flow sensor; An oxygen flow target value acquisition module is used to acquire a set oxygen flow target value; A multi-order offset correlation feature extraction module is used to extract the multi-order offset correlation features between the time series of the real-time oxygen flow value and the set oxygen flow target value to obtain a sequence of local first-order time series correlation feature vectors of the target offset of the oxygen flow value and a sequence of local second-order time series correlation feature vectors of the target offset of the oxygen flow value; A dynamic interactive fusion module is used to input each group of corresponding oxygen flow value target offset local first-order time series correlation feature vectors and oxygen flow value target offset local second-order time series correlation feature vectors in the sequence of the oxygen flow value target offset local first-order time series correlation feature vectors and the sequence of the oxygen flow value target offset local second-order time series correlation feature vectors into a feature vector dynamic interactive fusion module based on gated response to obtain a sequence of oxygen flow value target offset local multi-order time series correlation fusion feature vectors; A control instruction determination module, used to determine a control instruction based on a sequence of local multi-order time series correlation fusion feature vectors of the oxygen flow value target offset, wherein the control instruction is used to indicate that the speed value of the compressor should increase, decrease or remain unchanged; Among them, extracting the multi-order offset correlation features between the time series of the real-time oxygen flow value and the set oxygen flow target value to obtain a sequence of local first-order time series correlation feature vectors of the target offset of the oxygen flow value and a sequence of local second-order time series correlation feature vectors of the target offset of the oxygen flow value, includes: Calculating the difference between each oxygen real-time flow value in the time series of the oxygen real-time flow value and the set oxygen flow target value to obtain a time series of oxygen flow value target offset; Inputting the time series of the oxygen flow value target offset into a first-order temporal encoder to obtain a sequence of local first-order temporal correlation feature vectors of the oxygen flow value target offset; The sequence of local temporal correlation feature vectors of target offset of oxygen flow value is input into a second-order temporal encoder to obtain the sequence of local secondary temporal correlation feature vectors of target offset of oxygen flow value.

Images