Self-learning venting control method, device and medium for preventing surge of air blower

Abstract

This invention discloses a self-learning venting control method, device, and medium for preventing blower surge. The method includes: generating predictive data using collected data on blowers and venting valves from a wastewater treatment plant; calculating the probability of abnormal blower shutdown using a logistic regression model based on the predictive data; determining the venting state of the venting valve based on the calculated probability, performing corresponding processing, and updating the logistic regression model. This method enables adaptive self-learning for blower venting control, ensuring safe and efficient blower operation and preventing the problem of excessive pressure in aeration pipes leading to abnormal shutdowns caused by existing, more precise, and faster aeration control systems in wastewater treatment plants.

Description

Technical Field

[0001] This invention relates to the field of wastewater treatment, and in particular to a self-learning venting control method for preventing surge in blowers at wastewater treatment plants. Background Technology

[0002] As an essential component of activated sludge wastewater treatment plants, the aeration unit provides the necessary oxygen for the normal nitrogen and phosphorus removal by microorganisms within the plant. Simultaneously, the jetting airflow provides necessary mixing and effectively prevents sludge settling. The blower, as the gas source for the aeration unit, needs to constantly respond to changes in the influent flow and water quality, providing the necessary and appropriate air volume. Statistics show that the power consumption of aeration blowers in wastewater treatment plants accounts for 30% to 70% of the total energy consumption, highlighting the crucial importance of intelligent control in the aeration process. Furthermore, due to their high cost, blowers are among the most important electrical equipment in the fixed assets of wastewater treatment plants. Currently, with my country's economic development and the continuous implementation of dual-carbon policies, energy conservation, emission reduction, and low-carbon operation of wastewater treatment plants are increasingly valued by many projects. However, this also brings a strong demand for blowers that can quickly respond to precise aeration control commands from the wastewater treatment plant. Due to the frequent and precise aeration control commands, cases of blowers experiencing surge shutdowns due to excessive outlet pressure, or even damage after multiple abnormal shutdowns, are not uncommon. Therefore, the problem to be solved is how to provide a fully automatic venting control method that can adaptively learn and detect the operating status of the blower to prevent blower surge.

[0003] In view of this, the present invention is hereby proposed. Summary of the Invention

[0004] The purpose of this invention is to provide a self-learning venting control method to prevent blower surge, which can adaptively learn and automatically detect the operating status of blowers in sewage treatment plants, prevent blower surge from causing abnormal shutdowns, and thus solve the above-mentioned technical problems existing in the prior art.

[0005] The objective of this invention is achieved through the following technical solution:

[0006] A self-learning venting control method for preventing surge in a blower includes:

[0007] Step 1: Generate prediction data using collected data on wastewater treatment plant blowers and vent valves.

[0008] Step 2: Using the prediction data generated in Step 1, calculate the probability of the blower machine experiencing an abnormal shutdown using a logistic regression discriminant model;

[0009] Step 3: Determine if the probability obtained in Step 2 is greater than the first threshold. If yes, proceed to Step 4; otherwise, proceed to Step 7.

[0010] Step 4: Determine if there is a vent valve available for venting: if yes, proceed to step 5; otherwise, proceed to step 6.

[0011] Step 5: After switching the smallest numbered vent valve used for venting to the venting state, return to step 1;

[0012] Step 6: Prompt that manual intervention is needed to increase the number of vent valves, keep the position of each vent valve unchanged from the previous moment, and return to Step 1;

[0013] Step 7: Determine whether the probability obtained in Step 2 is less than the second threshold. If yes, proceed to Step 8; otherwise, proceed to Step 9.

[0014] Step 8: Determine if there is a vent valve that is currently venting. If yes, proceed to step 10; otherwise, proceed to step 9.

[0015] Step 9: Maintain the vent valve as is and proceed to step 11;

[0016] Step 10: Switch the largest numbered vent valve that is currently venting to the non-venting state;

[0017] Step 11: Determine if the blower has stopped. If yes, proceed to step 12; otherwise, return to step 1.

[0018] Step 12: Add the incorrectly judged samples to the training set;

[0019] Step 13: Retrain the logistic regression discriminant model using the training set updated in Step 12, and replace the old logistic regression discriminant model from Step 2 with the updated logistic regression discriminant model.

[0020] A processing apparatus, comprising:

[0021] At least one memory for storing one or more programs;

[0022] At least one processor is capable of executing one or more programs stored in the memory, such that when the processor executes one or more programs, the processor can implement the method of the present invention.

[0023] A readable storage medium storing a computer program that, when executed by a processor, enables the implementation of the methods described in this invention.

[0024] Compared with the prior art, the self-learning venting control method for preventing surge in wastewater treatment plant blowers provided by this invention has the following advantages:

[0025] This method can adaptively learn and control the blower venting, combining the blower's status (total air volume, blower current, and blower pressure), the venting valve status of the aeration unit (opening degree of each aeration pipe), and the venting valve's changing status to prevent blower shutdowns caused by aeration control logic. This ensures the safe and efficient operation of the blower and effectively solves the problem of abnormal shutdowns caused by excessive pressure in the aeration pipeline due to frequent fluctuations in influent water volume and quality, as well as the need for more precise and rapid aeration control systems in wastewater treatment plants. Attached Figure Description

[0026] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is an overall flowchart of a self-learning venting control method for preventing surge in a wastewater treatment plant blower, provided in an embodiment of the present invention.

[0028] Figure 2 The automatic switching state curve of the vent valve numbered 8102 provided in the embodiment of the present invention according to the automatic vent control method of the present invention; wherein, a state value of 3 indicates closed, and a state value of 2 indicates venting.

[0029] Figure 3 The diagram shows a comparison of the blower pressure change curves when the self-learning venting control method provided in this embodiment of the invention is applied and when the venting control method is not applied; where the asterisk indicates the control method of this invention is not applied and the asterisk indicates the control method of this invention is applied. Detailed Implementation

[0030] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the specific content of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments, which do not constitute a limitation of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0031] First, the following explanations are provided for the terms that may be used in this article:

[0032] The term "and / or" means that either or both can be achieved simultaneously. For example, X and / or Y means that it includes both "X" or "Y" as well as the three cases of "X and Y".

[0033] The terms “including,” “comprising,” “containing,” “having,” or other similar semantic descriptions should be interpreted as non-exclusive inclusion. For example, “including a technical feature element (such as raw material, component, ingredient, carrier, dosage form, material, size, part, component, mechanism, device, step, process, method, reaction conditions, processing conditions, parameter, algorithm, signal, data, product or article of manufacture, etc.)” should be interpreted as including not only the expressly listed technical feature element, but also other technical feature elements that are not expressly listed and are well-known in the art.

[0034] The term "composed of" excludes any technical features not expressly listed. When used in a claim, it closes the claim to exclude all technical features other than those expressly listed, except for associated conventional impurities. If the term appears only in a clause of a claim, it limits the claim to the elements expressly listed in that clause; elements recited in other clauses are not excluded from the overall claim.

[0035] Unless otherwise explicitly specified or limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this document according to the specific circumstances.

[0036] The terms “center,” “longitudinal,” “lateral,” “length,” “width,” “thickness,” “upper,” “lower,” “front,” “back,” “left,” “right,” “vertical,” “horizontal,” “top,” “bottom,” “inner,” “outer,” “clockwise,” and “counterclockwise” indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience and simplification of description and do not imply that the device or component referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this document.

[0037] The self-learning venting control method for preventing surge in wastewater treatment plant blowers provided by this invention will be described in detail below. Contents not described in detail in the embodiments of this invention are prior art known to those skilled in the art. Where specific conditions are not specified in the embodiments of this invention, conventional conditions or conditions recommended by the manufacturer shall apply. Reagents or instruments used in the embodiments of this invention, unless otherwise specified, are all commercially available conventional products.

[0038] like Figure 1As shown, this embodiment of the invention provides a self-learning venting control method for preventing blower surge, comprising the following steps:

[0039] Step 1: Generate prediction data using collected data on wastewater treatment plant blowers and vent valves.

[0040] Step 2: Using the prediction data generated in Step 1, calculate the probability of the blower machine experiencing an abnormal shutdown using a logistic regression discriminant model;

[0041] Step 3: Determine if the probability obtained in Step 2 is greater than the first threshold. If yes, proceed to Step 4; otherwise, proceed to Step 7.

[0042] Step 4: Determine if there is a vent valve available for venting: if yes, proceed to step 5; otherwise, proceed to step 6.

[0043] Step 5: After switching the smallest numbered vent valve used for venting to the venting state, return to step 1;

[0044] Step 6: Prompt that manual intervention is needed to increase the number of vent valves, keep the position of each vent valve unchanged from the previous moment, and return to Step 1;

[0045] Step 7: Determine whether the probability obtained in Step 2 is less than the second threshold. If yes, proceed to Step 8; otherwise, proceed to Step 9.

[0046] Step 8: Determine if there is a vent valve that is currently venting. If yes, proceed to step 10; otherwise, proceed to step 9.

[0047] Step 9: Maintain the vent valve as is and proceed to step 11;

[0048] Step 10: Switch the largest numbered vent valve that is currently venting to the non-venting state;

[0049] Step 11: Determine if the blower has stopped. If yes, proceed to step 12; otherwise, return to step 1.

[0050] Step 12: Add the incorrectly judged samples to the training set;

[0051] Step 13: Retrain the logistic regression discriminant model using the training set updated in Step 12, and replace the old logistic regression discriminant model from Step 2 with the updated logistic regression discriminant model.

[0052] Specifically, the vent valve in the above method refers to the aeration vent valve.

[0053] Preferably, in step 1 of the above method, the predictive data is generated using the collected data related to the wastewater treatment plant blower and the vent valve in the following manner:

[0054] The data collected from the wastewater treatment plant's blower and vent valves include: total air volume, current, pressure, opening degree of each vent valve, opening degree of the vent valve at the previous moment, and whether the blower has stopped.

[0055] The relative position P of the vent valve is determined by mapping the relative position of the vent valve to a unit interval using the following formula (4). i for:

[0056]

[0057] In formula (4), a i i = 0, 1, 2 are the parameters to be optimized; The relative position of the vent valve is estimated using formula (3) by a linear combination of the relative opening degree and the relative air volume of the vent valve. Formula (3) is as follows:

[0058]

[0059] In formula (3), rv i The relative opening of the vent valve is calculated using the following formula (1):

[0060]

[0061] In formula (1), v ij Let i represent the opening degree of the i-th vent valve at time j, where i = 1, 2, ..., n, n is the number of vent valves, and j = 1, 2, ..., m, m is the number of training samples;

[0062] ra i The relative air volume of the vent valve is calculated using the following formula (2):

[0063]

[0064] In formula (2), a ij Let i represent the gas volume of the i-th vent valve at time j, where i = 1, 2, ..., n, n is the number of vent valves, and j = 1, 2, ..., m, m is the number of training samples;

[0065] The sum of the three vectors—the feature combination vector z, the position encoding vector p, and the feature encoding vectors of different categories s—is used to form x = z + p + s. The correlation between different features σ(x) is captured by the following formula (5):

[0066]

[0067] In formula (5), the superscript T indicates matrix transpose of parameter x;

[0068] All feature combination vectors z are: z = [b1, b2, b3, v1, v2, ..., v n ,dv1,dv2,…,dv n ], where b1, b2, b3 represent the total air volume of the blower, the blower pressure, and the blower current, respectively; v1, v2, ..., v n These represent the opening degrees of each vent valve; dv1, dv2, ..., dv n These represent the differences between the opening degree of each vent valve at the current moment and the corresponding vent valve opening degree at the previous moment;

[0069] The positional encoding vector p is: p = [0, 0, 0, p1, p2, ..., p n ,p1,p2,…,p n ], where 0 represents the blower position code, p i This indicates the position code of the vent valve, i = 1, 2…n;

[0070] The feature encoding vectors for different categories are: s=[0,0,0,1,1,…,1,-1,-1,…,-1], where 0 represents the blower-related feature encoding; 1 represents the vent valve opening feature encoding; and -1 represents the vent valve opening change feature encoding.

[0071] Preferably, step 1 of the above method further includes:

[0072] The correlation between different features σ(x) is linearly transformed using the following formula (6) to expand the feature fission:

[0073] f(σ(x))=linearσ(x)) (6);

[0074] Specifically, the input for this step is σ(x), and the output is f(σ(x)), with the output data dimension being [1.5·d]. x [], [] represents the floor function, whose value is the smallest integer not exceeding 0, and d x =2n+3.

[0075] Preferably, in step 2 of the above method, the probability of the blower machine experiencing an abnormal shutdown is calculated using the logistic regression discriminant model of the following formula (7). Formula (7) is: the prediction data generated in step 1.

[0076]

[0077] Where f(σ(x)) is the prediction data generated in step 1, and w and b are the parameters to be optimized;

[0078] The logistic regression discriminant model uses the cross-entropy of formula (8) as the objective function and is optimized using the stochastic gradient descent method:

[0079]

[0080] In formula (8), i = 1, 2, ..., m, where m is the number of training samples; y i This represents the actual probability that the blower will experience an abnormal shutdown due to the i-th vent valve. This represents the predicted probability that the blower will experience an abnormal shutdown due to the i-th vent valve.

[0081] Preferably, in the above method, in step 3, the first threshold is 0.5;

[0082] In step 7, the second threshold is 0.5.

[0083] This invention also provides a processing device, comprising:

[0084] At least one memory for storing one or more programs;

[0085] At least one processor is capable of executing one or more programs stored in the memory, such that when the processor executes one or more programs, the processor can implement the methods described above.

[0086] The present invention further provides a readable storage medium storing a computer program that can implement the above-described method when executed by a processor.

[0087] In summary, this invention is an adaptive and self-learning venting control method that combines the blower's status (total air volume, blower current, and blower pressure), the venting valve status of the aeration unit (opening degree of each aeration pipe), and changes in the venting valve status to prevent blower shutdowns caused by aeration control logic. This ensures the safe and efficient operation of the blower and effectively solves the problems of frequent fluctuations in influent flow and quality, as well as abnormal shutdowns caused by more precise and rapid aeration control systems in wastewater treatment plants. It also effectively addresses the shortcomings of existing control methods, such as difficulty in making appropriate judgments for situations exceeding manually set rules, limitations of human experience and rules in covering all situations, and the inability to update rules promptly to address existing problems and ensure that the rules cover the latest issues. To address these problems, this invention adopts the following technical solution:

[0088] To more clearly demonstrate the technical solution and its effects provided by the present invention, the following describes in detail the self-learning venting control method for preventing blower surge provided by the embodiments of the present invention using specific examples.

[0089] Example 1

[0090] The following is combined with Figures 1-3 This invention provides a detailed description of the self-learning venting control method for preventing blower surge, and illustrates its effectiveness with a specific embodiment. For ease of calculation, historical data is used to manually label data on abnormal blower shutdowns induced by water quality and quantity fluctuations, aeration control logic, etc., as well as data on normal operation. Data on abnormal shutdowns is marked as 1, and data on normal operation is marked as 0. The total air volume, current, pressure, opening degree of each aeration venting valve, and the opening degree of the aeration venting valve at the previous moment are collected during blower operation. Furthermore, due to the variation in pipe diameter of the aeration pipeline with distance from the blower, and the different pressure effects of venting valves with the same opening degree depending on their location, features describing the venting valve positions are added.

[0091] (1) Let b be the relevant variables of the blower in the sewage treatment plant, and let b1, b2 and b3 be the total air volume, blower pressure and blower current respectively.

[0092] Let v be the relevant variable of the vent valve in the wastewater treatment plant, n be the number of vent valves, and v1, v2, v3, ..., v be the opening degree of each vent valve. n Simultaneously, the difference between the opening degree of each vent valve at the current time and the opening degree of the vent valve at the previous time is recorded as dv1, dv2, ..., dv n ;

[0093] All features are combined into a combination vector z = [b1, b2, b3, v1, v2, ..., v n ,dv1,dv2,…,dv n If ], then the dimension of the combined vector z is 2n+3;

[0094] Let the air volume corresponding to each vent valve be a1, a2, ..., a n Since the airflow of the vent valve is related to the vent valve opening degree, the total airflow of the blower, the vent valve position, and the opening degrees of other vent valves, the relative opening degrees rv1, rv2, ..., rv of the vent valve are used here to represent the influence of the vent valve position on the airflow. n The relative air volume ra1, ra2, ..., ra of the vent valve n The relative position of the vent valve is estimated, and the relative opening degree of the vent valve is calculated using the following formula (1):

[0095]

[0096] In formula (1), v ij Let i represent the opening degree of the i-th vent valve at time j, where i = 1, 2, ..., n, n represents the number of vent valves, and j = 1, 2, ..., m, m represents the number of training samples;

[0097]

[0098] In formula (2), a ij Let i represent the gas volume of the i-th vent valve at time j, where i = 1, 2, ..., n, n represents the number of vent valves, and j = 1, 2, ..., m, m represents the number of training samples.

[0099] The relative position of the vent valve is estimated using a linear combination of the relative opening degree and the relative air volume of the vent valve:

[0100]

[0101] For ease of calculation, the relative position of the vent valve is... Mapped to a unit interval, the relative position of the vent valve of the i-th vent valve is:

[0102]

[0103] Among them, a i i = 0, 1, 2 are parameters to be determined.

[0104] (2) For the blower feature, the blower position code is set to 0 because the distance of the blower from itself is 0, and the blower itself does not need a position code. Therefore, the position code vector is: p = [0, 0, 0, p1, p2, ..., p n ,p1,p2,…,p n If ], then the dimension of the position encoding vector p is 2n+3;

[0105] In order for subsequent calculations to be able to capture the differences between features of different categories, it is necessary to encode the features of different categories.

[0106] Specifically, the blower-related features are encoded as 0, the vent valve opening feature is encoded as 1, and the vent valve opening change feature is encoded as -1. Therefore, the feature encoding vectors for different categories are: s=[0,0,0,1,1,…,1,-1,-1,…,-1], and the dimension of the feature encoding vectors s for different categories is 2n+3;

[0107] Summing the three feature vectors z, p, and s, we get x = z + p + s. To capture the correlation between different features, we use the following formula:

[0108]

[0109] To separate the broad meaning represented by a single numerical value, a feature fission layer is added here, that is, the features are expanded to a larger dimension through linear transformation:

[0110] f(σ(x))=linear(σ(x)) (6);

[0111] Specifically, in this step, σ(x) is input into the above formula (6), and the output f(σ(x)) is the prediction data, with the output data dimension being [1.5·d]. x The [] rounding operation returns the smallest integer value that does not exceed the original value, and dx = 2n + 3;

[0112] The logistic regression model using the following formula (7) is used as a discriminant to judge the data used for prediction:

[0113]

[0114] Using the cross-entropy of formula (8) as the objective function, the stochastic gradient descent method is used for optimization:

[0115]

[0116] In formula (8), i = 1, 2, ..., m, where m is the number of training samples.

[0117] Since the prediction problem in this embodiment is essentially a binary classification problem, the calculated... This indicates the probability of the blower experiencing an abnormal shutdown. Therefore, when When the value is greater than 0.5, if any vent valve is closed, the vent valves will be opened from the smallest closed valve in ascending order of their numbers, thus putting them into the venting state; when When the value is less than η (η<0.5), if any vent valve is in the open state, then close the largest numbered vent valve in the open state according to the numbering from largest to smallest, so that it exits the venting state:

[0118]

[0119] Once the automatic venting control method is put into operation, if the blower experiences an abnormal shutdown due to the control logic of the aeration unit, i.e. the automatic venting program fails to perform the correct venting action in a timely manner, the automatic venting program enters a self-learning state: it adds the data that failed to correctly determine whether the blower should enter the automatic venting state to the training set and repeats each control step until the parameter update is completed; otherwise, the self-learning program remains in a silent state.

[0120] from Figure 2The diagram shown illustrates the automatic switching status curve of the vent valve numbered 8102 according to the automatic venting logic. Figure 2 In the diagram, a status value of 3 indicates closed, and a status value of 2 indicates venting. This demonstrates that the control method of this invention can automatically determine when to open and close the venting valve based on current data. Simultaneously, from... Figure 3 The outlet pressure fluctuation curves before and after deploying the control method of this invention show that before deployment (before the asterisk), the fan experienced surge (a sharp drop in air pressure to 0), but after deployment (after the asterisk), the fan outlet pressure became relatively stable and surge did not occur, verifying the effectiveness of this solution. This is because before deploying this solution, manual intervention of the fan was mainly relied upon, making it difficult to guarantee real-time performance and accuracy. However, the control method of this invention does not rely on manual intervention and has higher real-time performance.

[0121] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0122] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims. The information disclosed in the background section is intended only to enhance the understanding of the overall background technology of the present invention and should not be construed as an admission or implication in any way that such information constitutes prior art known to those skilled in the art.

Claims

1. A self-learning venting control method for preventing surge in a blower, characterized in that, include: Step 1: Generate prediction data using collected data on wastewater treatment plant blowers and vent valves. In step 1, predictive data is generated using the collected data on the wastewater treatment plant's blowers and vent valves, in the following manner: The relative position P of the vent valve is determined by mapping the relative position of the vent valve to a unit interval using the following formula (4). i for: （4）； In formula (4), a i i=0,1,2 are the parameters to be optimized; The relative position of the vent valve is estimated using formula (3) by a linear combination of the relative opening degree and the relative air volume of the vent valve. Formula (3) is as follows: （3）； In formula (3), rv i The relative opening of the vent valve is calculated using the following formula (1): （1）； In formula (1), v ij Let i represent the opening degree of the i-th vent valve at time j, where i = 1, 2, ..., n, n is the number of vent valves, and j = 1, 2, ..., m, m is the number of training samples. ra i The relative air volume of the vent valve is calculated using the following formula (2): （2）； In formula (2), a ij Let i represent the gas volume of the i-th vent valve at time j, where i = 1, 2, ..., n, n is the number of vent valves, and j = 1, 2, ..., m, m is the number of training samples. The sum of the three vectors—the feature combination vector z, the position encoding vector p, and the feature encoding vectors of different categories s—is used to form x = z + p + s. The correlation between different features is captured by the following formula (5). Formula (5) is: （5）； In formula (5), the superscript T indicates matrix transpose of parameter x; All feature combination vectors z are: z = [b1, b2, b3, v1, v2, ..., v n ,dv1,dv2,…,dv n ], where b1, b2, b3 represent the total air volume of the blower, the blower pressure, and the blower current, respectively; v1, v2, ..., v n These represent the opening degrees of each vent valve; dv1, dv2, ..., dv n These represent the differences between the opening degree of each vent valve at the current moment and the corresponding vent valve opening degree at the previous moment; The position encoding vector p is: p=[0,0,0,p1,p2,…,p n ,p1,p2,…,p n ], where 0 represents the blower position code, p i This indicates the position code of the vent valve, i=1,2…n; The feature encoding vectors for different categories are: s=[0,0,0,1,1,…,1,-1,-1,…,-1], where 0 represents the blower-related feature encoding; 1 represents the vent valve opening feature encoding; and -1 represents the vent valve opening change feature encoding. Step 2: Using the prediction data generated in Step 1, calculate the probability of the blower machine experiencing an abnormal shutdown using a logistic regression discriminant model; Step 3: Determine if the probability obtained in Step 2 is greater than the first threshold. If yes, proceed to Step 4; otherwise, proceed to Step 7. Step 4: Determine if there is a vent valve available for venting: if yes, proceed to step 5; otherwise, proceed to step 6. Step 5: After switching the smallest numbered vent valve used for venting to the venting state, return to step 1; Step 6: Prompt that manual intervention is needed to increase the number of vent valves, keep the position of each vent valve unchanged from the previous moment, and return to Step 1; Step 7: Determine whether the probability obtained in Step 2 is less than the second threshold. If yes, proceed to Step 8; otherwise, proceed to Step 9. Step 8: Determine if there is a vent valve that is currently venting. If yes, proceed to step 10; otherwise, proceed to step 9. Step 9: Maintain the vent valve as is and proceed to step 11; Step 10: Switch the largest numbered vent valve that is currently venting to the non-venting state; Step 11: Determine if the blower has stopped. If yes, proceed to step 12; otherwise, return to step 1. Step 12: Add the incorrectly judged samples to the training set; Step 13: Retrain the logistic regression discriminant model using the training set updated in Step 12, and replace the old logistic regression discriminant model from Step 2 with the updated logistic regression discriminant model.

2. The self-learning venting control method for preventing blower surge according to claim 1, characterized in that, Step 1 further includes: The correlation between different features obtained The characteristic fission expansion is achieved through linear transformation using the following formula (6): Formula (6) is: （6）； Specifically, the input for this step is... The output is The output data dimension is [] represents the floor function, whose value is the smallest integer not exceeding 0, and d x =2n+3.

3. The self-learning venting control method for preventing surge in a blower according to any one of claims 1-2, characterized in that, In step 2, the probability of the blower machine experiencing an abnormal shutdown is calculated using the logistic regression discriminant model of the following formula (7). Formula (7) is: the prediction data generated in step 1. （7）； in, The prediction data generated in step 1; w and b are the parameters to be optimized, respectively; The logistic regression discriminant model uses the cross-entropy of formula (8) as the objective function and is optimized using the stochastic gradient descent method: （8）； In formula (8), i = 1, 2, ..., m, where m is the number of training samples; This represents the actual probability that the blower will experience an abnormal shutdown due to the i-th vent valve. This represents the predicted probability that the blower will experience an abnormal shutdown due to the i-th vent valve.

4. The self-learning venting control method for preventing blower surge according to any one of claims 1-2, characterized in that, In step 3, the first threshold is 0.5; In step 7, the second threshold is , The following formula (9) is used to calculate the result: （9）； In formula (9), This represents the predicted probability that the blower will experience an abnormal shutdown due to the i-th vent valve.

5. A processing device, characterized in that, include: At least one memory for storing one or more programs; At least one processor is capable of executing one or more programs stored in the memory, such that when the one or more programs are executed by the processor, the processor can perform the method according to any one of claims 1-4.

6. A readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it can implement the method described in any one of claims 1-4.

Images