Blower aeration control method, device, equipment, medium and product

Abstract

The invention discloses an air blower aeration control method, device and equipment, a medium and a product. The method comprises the following steps: acquiring the current operating parameters of the biochemical pool and the current opening degree of an aeration regulating valve; inputting the current operation parameters and the current opening degree into a pre-trained blast opening degree and determination model to obtain a to-be-processed air volume opening degree sum; the blast opening and the determination model are obtained by training the dissolved oxygen concentration corresponding to each state node in the state transition track of the biochemical pool; the dissolved oxygen concentration is preset and / or determined based on the dissolved oxygen concentration corresponding to the previous state node and the opening degree of the air volume to be used; the to-be-processed air volume opening degree sum is processed according to the target optimization function, and the target air volume opening degree of each air blower is obtained; and controlling each air blower to convey air to the biochemical pool according to the target air volume opening. According to the technical scheme provided by the embodiment of the invention, the accuracy of aeration control of the air blower can be improved, and the energy consumption is reduced, so that the sewage treatment effect is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of sewage treatment, and in particular to a blower aeration control method, device, equipment, medium and product. Background Art

[0002] With the development of urbanization and industrialization, sewage treatment issues have received more and more attention. In the sewage treatment process, it is usually necessary to control the blower to provide gas for the biochemical pool to increase the dissolved oxygen content of the water in the biochemical pool, so that microorganisms can better perform biodegradation in an aerobic environment and convert organic matter into inorganic matter, thereby achieving the purpose of purifying water quality.

[0003] At present, the aeration control method of blowers is usually to rely on manual control of the aeration flow of the blowers according to the actual sewage treatment needs to provide the corresponding gas for the biochemical pool. However, when it comes to aeration control of multiple blowers, this method not only has the problem of poor aeration control accuracy, but also leads to high energy consumption, resulting in poor sewage treatment effect. Summary of the invention

[0004] The present invention provides a blower aeration control method, device, equipment, medium and product to improve the accuracy of blower aeration control, reduce energy consumption, and thus improve sewage treatment effect.

[0005] According to one aspect of the present invention, a blower aeration control method is provided, which is applied to a sewage treatment system, wherein the sewage treatment system includes a biochemical pool and a plurality of blowers connected by an aeration pipe; an aeration regulating valve is provided on the aeration pipe; the method comprises:

[0006] Obtaining the current operating parameters of the biochemical pool and the current opening of the aeration regulating valve; wherein the current operating parameters at least include dissolved oxygen concentration, influent parameters and sludge concentration;

[0007] The current operating parameters and the current opening are input into a pre-trained blast opening and determination model to obtain the opening and of the air volume to be processed; wherein the blast opening and determination model is trained based on the dissolved oxygen concentration corresponding to each state node in the state transition trajectory of the biochemical pool; the dissolved oxygen concentration corresponding to the state node is preset and / or determined based on the dissolved oxygen concentration corresponding to the previous state node and the opening and of the air volume to be used;

[0008] Processing the air volume opening to be processed and the air volume opening to be processed according to the target optimization function to obtain the target air volume opening of each of the blowers;

[0009] Each of the blowers is controlled to deliver air to the biochemical pool according to the corresponding target air volume opening.

[0010] According to another aspect of the present invention, a blower aeration control device is provided, which is configured in a sewage treatment system, wherein the sewage treatment system includes a biochemical pool and a plurality of blowers connected by an aeration pipe; an aeration regulating valve is provided on the aeration pipe; the device includes:

[0011] A data acquisition module, used to acquire the current operating parameters of the biochemical pool and the current opening of the aeration regulating valve; wherein the current operating parameters at least include dissolved oxygen concentration, influent parameters and sludge concentration;

[0012] An air volume opening and determination module is used to input the current operating parameters and the current opening into a pre-trained air blast opening and determination model to obtain the air volume opening and to be processed; wherein the air blast opening and determination model is obtained by training based on the dissolved oxygen concentration corresponding to each state node in the state transition trajectory of the biochemical pool; the dissolved oxygen concentration corresponding to the state node is preset and / or determined based on the dissolved oxygen concentration corresponding to the previous state node and the air volume opening and to be used;

[0013] A target air volume opening determination module, used for processing the air volume opening to be processed and the target air volume opening of each blower according to a target optimization function;

[0014] The control module is used to control each of the blowers to deliver air volume to the biochemical pool according to the corresponding target air volume opening.

[0015] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0016] at least one processor; and a memory in communication with the at least one processor; wherein,

[0017] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the blower aeration control method described in any embodiment of the present invention.

[0018] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the blower aeration control method described in any embodiment of the present invention when executed.

[0019] According to another aspect of the present invention, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the blower aeration control method as described in any embodiment of the present invention.

[0020] The technical solution of the embodiment of the present invention obtains the current operating parameters of the biochemical pool and the current opening of the aeration regulating valve; inputs the current operating parameters and the current opening into the pre-trained blower opening and determination model to obtain the to-be-processed air volume opening and; the blower opening and determination model is obtained by training based on the dissolved oxygen concentration corresponding to each state node in the state transition trajectory of the biochemical pool; the dissolved oxygen concentration corresponding to the state node is preset and / or determined based on the dissolved oxygen concentration corresponding to the previous state node and the to-be-used air volume opening and; the to-be-processed air volume opening and and are processed according to the target optimization function to obtain the target air volume opening of each blower; and each blower is controlled to deliver air volume to the biochemical pool according to the corresponding target air volume opening, thereby solving the problem of the prior art relying on In order to solve the problem of poor aeration control accuracy, high energy consumption and poor sewage treatment effect, the aeration flow rate of the blower is controlled manually according to the actual sewage treatment needs. This paper realizes that the current operating parameters of the biochemical pool and the current opening of the aeration regulating valve are processed by using the blower opening and determination model to obtain the corresponding openings and of the to-be-treated air volumes of all blowers. Then, the openings and of the to-be-treated air volumes are processed according to the target optimization function, and the energy consumption of the target air volume opening of each blower is controlled to minimize the energy consumption. By adjusting the opening of each blower according to the target air volume opening, the flow rate of the transported gas can be accurately controlled, which improves the accuracy of the blower aeration control while ensuring that the biochemical pool obtains the required amount of gas, thereby improving the sewage treatment effect.

[0021] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0023] Figure 1 is a flow chart of a blower aeration control method provided according to Embodiment 1 of the present invention;

[0024] Figure 2 is a flow chart of a blower aeration control method provided according to Embodiment 2 of the present invention;

[0025] Figure 3 is a flow chart of a blower aeration control method provided according to Embodiment 3 of the present invention;

[0026] Figure 4 is a schematic diagram of the structure of a dissolved oxygen concentration prediction model provided according to Embodiment 3 of the present invention;

[0027] Figure 5 is a structural schematic diagram of a blower aeration control device provided according to Embodiment 4 of the present invention;

[0028] Figure 6 It is a schematic diagram of the structure of an electronic device for implementing the blower aeration control method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0029] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0031] Embodiment 1

[0032] Figure 1 This is a flow chart of a blower aeration control method provided according to the first embodiment of the present invention. This embodiment can be applied to the situation of regulating the aeration air volume opening of each blower in the sewage treatment system. The method can be executed by a blower aeration control device. The blower aeration control device can be implemented in the form of hardware and / or software. The blower aeration control device can be configured in the sewage treatment system. The sewage treatment system includes a biochemical pool and multiple blowers connected by an aeration pipe; an aeration regulating valve is provided on the aeration pipe. Figure 1 As shown, the method includes:

[0033] S110, obtaining the current operating parameters of the biochemical pool and the current opening of the aeration regulating valve.

[0034] The operating parameters include at least dissolved oxygen concentration, inlet parameters and sludge concentration. The inlet parameters include but are not limited to inlet temperature, inlet chemical oxygen demand, inlet ammonia nitrogen, inlet total phosphorus and inlet flow rate, etc. The current opening can be used for the air flow rate of the blower during the aeration process of the biochemical pool.

[0035] In this embodiment, multiple blowers can deliver air to the biochemical pool through an aeration pipe, and transmit the air to the biochemical pool to increase the dissolved oxygen content of the water in the biochemical pool to achieve the purpose of purifying the water quality. The valve opening of the aeration regulating valve on the aeration pipe affects the flow rate of aeration to the biochemical pool. In order to accurately predict the air volume opening and of all blowers, the current operating parameters of the biochemical pool and the current opening of the aeration regulating valve can be obtained in real time or periodically to predict the air volume opening and of all blowers based on these data.

[0036] S120, inputting the current operating parameters and the current opening into the pre-trained blower opening and determination model to obtain the air volume opening and to be processed.

[0037] Among them, the sum of the air volume openings to be processed is the sum of the air volume openings for all blowers. The air volume opening and determination model is obtained by training based on the dissolved oxygen concentration corresponding to each state node in the state transition trajectory of the biochemical pool. The dissolved oxygen concentration corresponding to the state node is pre-set and / or determined based on the dissolved oxygen concentration corresponding to the previous state node and the sum of the air volume openings to be used based on the dissolved oxygen concentration prediction model. The dissolved oxygen concentration corresponding to the state node is determined. The dissolved oxygen concentration prediction model is obtained based on the historical operating parameters of the biochemical pool, the historical opening of the aeration control valve, and the historical air volume openings and training of multiple blowers.

[0038] In this embodiment, the current operating parameters and the current opening can be used as inputs of the air blast opening and determination model to obtain the opening and distribution data. The maximum value of the opening and distribution data can be selected as the air volume opening and the air volume opening; or the opening and distribution data can be randomly sampled to obtain the air volume opening and the air volume opening and the air volume opening.

[0039] S130, processing the air volume opening to be processed according to the target optimization function to obtain the target air volume opening of each blower.

[0040] Among them, the target air volume opening can be used to control the air supply volume of the corresponding blower.

[0041] In this embodiment, the sum of the air volume openings to be processed can be processed according to the target optimization function, with the goal of minimizing the energy consumption corresponding to the air volume openings of different blowers, and the sum of the air volume openings of each blower is consistent with the sum of the air volume openings to be processed, so as to obtain the target air volume opening of each blower.

[0042] In this embodiment, the objective optimization function includes an energy consumption optimization function and at least one constraint condition; the air volume opening and are processed according to the objective optimization function to obtain the target air volume opening of each blower, including: processing the air volume opening and to be processed according to at least one constraint condition to obtain an air volume opening set; determining the energy consumption corresponding to the air volume opening set according to the energy consumption optimization function, and determining the target air volume opening of each blower based on the air volume opening set when the target is reached, with the goal of minimizing the energy consumption.

[0043] The air volume opening set includes the air volume opening of each blower. The constraint condition may refer to constraint information for determining the air volume opening of each blower. The constraint condition at least includes that the sum of the air volume openings of each blower in the air volume opening set is consistent with the air volume opening sum.

[0044] Specifically, the sum of the air volume openings to be processed can be analyzed based on at least one constraint condition, and the sum of the air volume openings to be processed can be divided into each blower to obtain the air volume opening of each blower as an air volume opening set. If each blower operates at the air volume opening in the air volume opening set, the energy consumed may not be the minimum. The air volume opening of each blower in the air volume opening set can be used as an input parameter of the energy consumption optimization function to obtain the energy consumption corresponding to the air volume opening set. The air volume opening of each blower in the air volume opening set can be adjusted with the goal of minimizing energy consumption. At the same time, in order to ensure the operating stability of the blower aeration system, the air volume opening can be constrained using constraints, and finally a set of air volume openings is obtained to minimize the overall energy consumption. When the energy consumption corresponding to the air volume opening set is the minimum, it is considered that the target is achieved. Furthermore, the air volume opening of each blower in the air volume opening set when the target is achieved is determined as the target air volume opening for each blower.

[0045] For example, given the air volume opening to be processed and F, the target air volume opening of each blower can be solved based on the energy consumption optimization function and at least one constraint condition to minimize the energy consumption. The expression of the energy consumption optimization function can be expressed as:

[0046] min Power = w1x1 + w2x2 + ... + w n x n +0·(x n+1 +x n+2 +…+x 2n );

[0047] Constraints include at least one of the following:

[0048]

[0049] Among them, min means minimization, Power means energy consumption, and x n Indicates the air volume opening of the nth blower; w n represents the weight or deviation corresponding to the nth blower. F represents the sum of the air volume openings to be processed. f and g represent the maximum and minimum values ​​of the air volume openings of the blowers. For example, when the dth blower is turned on, its air volume opening f≤x d ≤g. It should be noted that x n+1 、x n+2 , …, x 2n It is a parameter set only to constrain the air volume opening of n blowers and does not participate in the sum of the air volume opening of the blowers.

[0050] The technical solution provided in this embodiment minimizes the energy consumption of each blower at the target air volume opening by constraining the sum of the target air volume openings of each blower to be consistent with the sum of the air volume openings to be processed, thereby achieving the effect of improving the aeration accuracy of the blower while reducing energy consumption.

[0051] In this embodiment, before processing the air volume opening to be processed according to the target optimization function to obtain the target air volume opening of each blower, multiple second samples can be obtained, and for each second sample, the air volume opening group in the second sample is input into the optimization function to be trained to obtain the predicted energy consumption; based on the third loss function, the energy consumption label in the second sample and the predicted energy consumption, the third loss value is determined, and the parameters of the optimization function to be trained are corrected based on the third loss value; with the convergence of the third loss function as the goal, the target optimization function is obtained.

[0052] Among them, the second sample includes an air volume opening group and an energy consumption label; the air volume opening group includes the air volume openings of multiple blowers. The energy consumption label can be used to characterize the sum of energy consumption corresponding to the air volume openings of multiple blowers in the air volume opening group, that is, the sum of energy consumed when multiple blowers operate at corresponding air volume openings. The optimization function to be trained may refer to an untrained energy consumption optimization function. The third loss value refers to the error between the theoretical energy consumption label and the predicted energy consumption.

[0053] In this embodiment, in order to make the obtained target optimization function have a higher accuracy, the second samples can be obtained as many and rich as possible. Each air volume opening in the air volume opening group in the second sample can be used as an input parameter of the optimization function to be trained to obtain the predicted energy consumption. Further, according to the third loss function, the energy consumption label and the predicted energy consumption in the second sample can be processed for loss, and the loss value obtained can be used as the third loss value. Furthermore, the parameters in the optimization function to be trained (such as the weight and deviation corresponding to the blower) can be adjusted by the third loss value. The training error of the third loss function, that is, the loss parameter, can be used as a condition for detecting whether the third loss function has reached convergence at present, such as whether the training error is less than the preset error or whether the error change trend tends to be stable, or whether the current number of iterations is equal to the preset number. If the convergence condition is detected, such as the training error of the third loss function reaches less than the preset error or the error change tends to be stable, it indicates that the training of the optimization function to be trained is completed, and the iterative training can be stopped at this time. If it is detected that the convergence condition is not reached at present, the second sample can be further obtained to train the optimization function to be trained until the training error of the third loss function is within the preset range. When the training error of the third loss function reaches convergence, the optimization function to be trained can be used as the energy consumption optimization function obtained by final training. The target optimization function can be obtained according to the energy consumption optimization function and at least one predetermined constraint condition.

[0054] For example, the energy consumption under different air volume opening groups can be modeled to construct the optimization function to be trained. For example, the optimization function to be trained can be expressed as y=W T X. Among them, the air volume opening group of n blowers is expressed as The weights and biases (i.e. parameters) corresponding to the n blowers w n+1 Integration is defined as y represents the total energy consumption (i.e. energy consumption). The third loss function can be expressed as: Among them, loss linear represents the loss value, represents the prediction result of the i-th second sample (i.e., predicted energy consumption), y i represents the historical real data of the i-th second sample (i.e., the energy consumption label), and N represents the number of second samples in a batch training. The third loss function can be combined to perform gradient descent on the optimization function to be trained, and W in the optimization function to be trained can be obtained. The optimization function to be trained with W fixed is used as the energy consumption optimization function.

[0055] S140, controlling each blower to deliver air to the biochemical pool according to the corresponding target air volume opening.

[0056] Specifically, each blower can be controlled to adjust its own air volume opening according to the target air volume opening assigned to it, so as to deliver the corresponding air volume to the biochemical pool, provide oxygen to the biochemical pool to support the metabolic activities of microorganisms, and realize sewage treatment.

[0057] The technical solution provided in this embodiment obtains the current operating parameters of the biochemical pool and the current opening of the aeration regulating valve; inputs the current operating parameters and the current opening into the pre-trained blower opening and determination model to obtain the air volume opening and to be processed; the blower opening and determination model is trained based on the dissolved oxygen concentration corresponding to each state node in the state transition trajectory of the biochemical pool; the dissolved oxygen concentration corresponding to the state node is pre-set and / or determined based on the dissolved oxygen concentration corresponding to the previous state node and the air volume opening and to be used; the air volume opening and to be processed are processed according to the target optimization function to obtain the target air volume opening of each blower; and each blower is controlled to deliver air volume to the biochemical pool according to the target air volume opening corresponding to it, thereby solving the problem of the prior art that the air volume opening and to be processed are determined based on the target air volume opening corresponding to the blower. Relying on manual control of the aeration flow of the blower according to the actual sewage treatment needs, resulting in poor aeration control accuracy, high energy consumption and poor sewage treatment effect, it is realized that the current operating parameters of the biochemical pool and the current opening of the aeration regulating valve are processed by using the blower opening and determination model to obtain the corresponding openings and of the to-be-treated air volumes of all blowers, and then, the to-be-treated air volume openings and are processed according to the target optimization function, and the energy consumption of the target air volume opening of each blower is controlled to minimize energy consumption. By adjusting the opening of each blower by the target air volume opening, the flow rate of the transported gas can be accurately controlled, which improves the accuracy of the blower aeration control while ensuring that the biochemical pool obtains the required amount of gas, thereby improving the sewage treatment effect.

[0058] Embodiment 2

[0059] Figure 2 This is a flow chart of a blower aeration control method provided according to the second embodiment of the present invention. On the basis of the above embodiment, before the current operating parameters and the current opening are input into the pre-trained blower opening and determination model to obtain the opening and of the air volume to be processed, the blower opening and determination model can also be trained. The specific implementation method can refer to the technical solution of this embodiment. Among them, the technical terms that are the same as or corresponding to the above embodiment are not repeated here.

[0060] like Figure 2 As shown, the method specifically comprises the following steps:

[0061] S210, confirming a state node in the state transfer trajectory; the state node corresponds to the dissolved oxygen concentration to be used in the biochemical pool.

[0062] The state transition trajectory may include one or more state nodes, and the state node is used to characterize the dissolved oxygen concentration to be used in the biochemical pool.

[0063] In this embodiment, the dissolved oxygen concentration to be used corresponding to the first state node in the state transition trajectory may be preset. Each state node after the first state node may be determined based on the dissolved oxygen concentration corresponding to the previous state node and the air volume opening to be used according to the pre-trained dissolved oxygen concentration prediction model.

[0064] S220. Input the dissolved oxygen concentration to be used, the preset operating parameters of the biochemical pool corresponding to the dissolved oxygen concentration to be used, and the preset opening of the aeration regulating valve into the aeration decision model to obtain opening and distribution information; the opening and distribution information includes multiple air volume openings to be used and.

[0065] The preset operating parameters include but are not limited to the influent parameters and sludge concentration of the biochemical pool.

[0066] In this embodiment, the dissolved oxygen concentration to be used corresponding to the state node, the preset operating parameters of the biochemical pool corresponding to the dissolved oxygen concentration to be used, and the preset opening of the aeration regulating valve can be used as inputs of the aeration decision model to obtain the opening and distribution information, which includes multiple air volume openings to be used.

[0067] S230. For each air volume opening to be used, the air volume opening to be used and the dissolved oxygen concentration to be used, as well as the preset operating parameters corresponding to the dissolved oxygen concentration to be used in the biochemical pool and the preset opening of the aeration regulating valve are input into the pre-trained dissolved oxygen concentration prediction model to obtain the predicted dissolved oxygen concentration, and the predicted dissolved oxygen concentration is used as the next state node of the state node.

[0068] The dissolved oxygen concentration prediction model is used to determine the next dissolved oxygen concentration of the biochemical pool after the input air volume opening is executed at the current dissolved oxygen concentration according to the current dissolved oxygen concentration, the operating parameters of the biochemical pool and the input air volume opening.

[0069] In this embodiment, the air volume opening to be used and the dissolved oxygen concentration to be used, as well as the preset operating parameters corresponding to the dissolved oxygen concentration to be used in the biochemical pool and the preset opening of the aeration regulating valve, can be input into the pre-trained dissolved oxygen concentration prediction model to obtain the predicted dissolved oxygen concentration. The predicted dissolved oxygen concentration refers to the dissolved oxygen concentration of the biochemical pool after the input air volume opening and are executed under the dissolved oxygen concentration to be used. Further, the predicted dissolved oxygen concentration can be used as the next state node of the state node.

[0070] Exemplarily, the dissolved oxygen concentration a to be used of state node 1 is input into the aeration decision model to obtain a plurality of air volume openings to be used and; then the air volume openings to be used and the dissolved oxygen concentration a to be used, the preset operating parameters corresponding to the biochemical pool when the dissolved oxygen concentration a to be used, and the preset opening of the aeration regulating valve are input into the dissolved oxygen concentration prediction model to obtain the predicted dissolved oxygen concentration b. The predicted dissolved oxygen concentration b is used as the next state node 2 of state node 1.

[0071] It should be noted that each state node in different state transition trajectories may be the same or different, and the air blowing opening and the determination model may be trained according to multiple state transition trajectories respectively. The number of state nodes in each state transition trajectory may be the same or different.

[0072] In this embodiment, after the predicted dissolved oxygen concentration is used as the next state node of the state node, when the number of state nodes in the state transition trajectory does not reach a preset threshold, the predicted dissolved oxygen concentration can be used as the dissolved oxygen concentration to be used, and the steps of inputting the dissolved oxygen concentration to be used into the aeration decision model to obtain the opening and distribution information, obtaining the predicted dissolved oxygen concentration, and determining whether the number of state nodes reaches the preset threshold are re-executed.

[0073] Specifically, after the predicted dissolved oxygen concentration is used as the next state node of the state node and the state node in the state transition trajectory is updated, it is determined whether the number of state nodes in the state transition trajectory reaches a preset threshold. If the preset threshold is not reached, the predicted dissolved oxygen concentration can be used as the dissolved oxygen concentration to be used, and the steps of inputting the dissolved oxygen concentration to be used into the aeration decision model to obtain the opening and distribution information and the predicted dissolved oxygen concentration are re-executed according to the dissolved oxygen concentration to be used, and the predicted dissolved oxygen concentration is used as the next state node of the state node, until the number of state nodes in the state transition trajectory reaches the preset threshold.

[0074] S240. For each state node, the first dissolved oxygen concentration corresponding to the state node is input into the value network model to obtain the expected return attributes corresponding to the aeration opening and the action executed at the first dissolved oxygen concentration, and determine the evaluation results of each air volume opening to be used and the corresponding action executed at the first dissolved oxygen concentration.

[0075] Among them, the first dissolved oxygen concentration is the dissolved oxygen concentration to be used and / or the dissolved oxygen concentration in the predicted dissolved oxygen concentration. That is to say, the dissolved oxygen concentration corresponding to each state node can be called the first dissolved oxygen concentration. The value network model is used to estimate the expected return attribute after taking an action (i.e., aeration opening and action) in a given state (i.e., the first dissolved oxygen concentration). The expected return attribute can be used to characterize the expected benefit of taking an average action in the current state. The action evaluation result can be used to characterize the expected benefit after taking a certain action (i.e., the air volume opening to be used and) in a given state (i.e., the first dissolved oxygen concentration).

[0076] In this embodiment, the first dissolved oxygen concentration corresponding to the state node can be used as the input of the value network model to obtain the expected return attribute corresponding to the aeration opening and action executed at the first dissolved oxygen concentration. At the same time, the evaluation results of each air volume opening to be used and the corresponding action can be determined at the first dissolved oxygen concentration. The model parameters of the aeration decision model can be modified by evaluating the quality of the action based on the expected return attribute and the action evaluation results.

[0077] Optionally, determining each opening of the air volume to be used and the corresponding action evaluation result to be executed respectively at the first dissolved oxygen concentration includes: for the first training, based on the opening and distribution information, determining each opening of the air volume to be used and the corresponding action evaluation result to be executed respectively at the first dissolved oxygen concentration; for the training after the first training, based on the opening and distribution information and the historical opening and distribution information obtained from the previous training, determining each opening of the air volume to be used and the corresponding action evaluation result to be executed respectively at the first dissolved oxygen concentration.

[0078] In this embodiment, when the aeration decision model is trained for the first time, the output probability of each air volume opening and corresponding to the dissolved oxygen concentration to be used output by the aeration decision model can be used as the action evaluation result corresponding to the corresponding air volume opening and at the dissolved oxygen concentration to be used. When the aeration decision model is trained again, the opening and distribution information and the historical opening and distribution information obtained in the previous training can be difference processed to obtain the difference corresponding to each air volume opening and; then, the difference corresponding to the air volume opening and is exponentially calculated, and the exponential result is used as the action evaluation result of the air volume opening and.

[0079] For example, the dissolved oxygen concentration b corresponding to the state node 2 is input into the aeration decision model to obtain the opening and distribution information P k+1 ;P k+1 The difference between the historical opening and distribution information obtained from the previous training is (P k+1 -P k). The opening and distribution information before and after training corresponding to the same state node are used to evaluate the current opening of each air volume to be used and the corresponding action evaluation results under the first dissolved oxygen concentration, thereby improving the accuracy of the evaluation.

[0080] S250: Based on the first loss function, each action evaluation result and the expected reward attribute, the model parameters of the aeration decision model are modified to obtain a trained aeration decision model.

[0081] In this embodiment, loss processing can be performed on each action evaluation result and expected return attribute according to the first loss function to obtain an error value, and the model parameters in the aeration decision model can be corrected according to the error value. Alternatively, the step size of the correction parameter can be determined according to the loss value, and the model parameters in the aeration decision model can be corrected according to the step size. For example, if the error between each action evaluation result and the expected return attribute is large, the step size of the correction parameter can be increased; if the error between each action evaluation result and the expected return attribute is small, the step size of the correction parameter can be reduced.

[0082] In this embodiment, based on the first loss function, each action evaluation result and the expected return attribute, the model parameters of the aeration decision model are modified, including: determining the dominance based on the expected return attribute and the second dissolved oxygen concentration corresponding to the state node after the state node; based on the first loss function, each action evaluation result and the dominance, the model parameters of the aeration decision model are modified.

[0083] The second dissolved oxygen concentration is the dissolved oxygen concentration to be used and / or the dissolved oxygen concentration in the predicted dissolved oxygen concentration. The advantage is used to evaluate the increment of the expected benefit of taking a specific action (i.e., the sum of the air volume opening to be used) compared with the average action under a specific state (i.e., the first dissolved oxygen concentration).

[0084] In this embodiment, a dominance estimation algorithm can be used to evaluate the dominance according to the expected reward attribute and the second dissolved oxygen concentration corresponding to the state node after the state node. Then, the first loss function is used to modify the model parameters of the aeration decision model in combination with the evaluation result and dominance of each action.

[0085] In this embodiment, the method of determining the dominance can be: determining the cumulative reward based on the second dissolved oxygen concentration corresponding to the state node after the state node; determining the dominance based on the expected return attribute and the cumulative reward. The cumulative reward can be used to represent the sum of all rewards obtained from the state node in the state transition trajectory to a certain terminal state.

[0086] Specifically, the dissolved oxygen concentration corresponding to the first state node after the state node can be determined as the second dissolved oxygen concentration. The intermediate value can be determined based on the second dissolved oxygen concentration. Optionally, if the second dissolved oxygen concentration is within a preset range, the intermediate value can be determined as a preset maximum reward value; if the second dissolved oxygen concentration is less than a preset threshold, the intermediate value can be determined as a preset minimum reward value; if the second dissolved oxygen concentration is not within the preset range and is not less than the preset threshold, the second dissolved oxygen concentration and the preset dissolved oxygen control target value can be differenced to obtain an intermediate value. Furthermore, the cumulative reward can be determined by combining the intermediate value and the third dissolved oxygen concentration corresponding to all state nodes after the first state node after the state node. Furthermore, the difference between the cumulative reward and the expected return attribute can be used as the advantage.

[0087] Exemplarily, determine the state transition trajectory D k The tth state node s in k t The calculation formula for the cumulative reward can be expressed as: Among them, R t represents the cumulative reward, s k t+1 Represents the state transition trajectory D k The t+1th state node in s k t The next state node; λ is the attenuation coefficient; r t Expressed as an intermediate value; the calculation formula for determining the intermediate value can be expressed as:

[0088] [a,b] represents the preset range, A1 represents the preset maximum reward value; target represents the preset dissolved oxygen control target value (such as 2); c represents the preset threshold value; B1 represents the preset minimum reward value. For example, a is 1.5, b is 2.5, and A1 is 100; it can be understood that: for a≤s k t+1 When ≤b, r is given t A maximum reward of 100. C is 0.5, and the preset minimum reward value is -100; it can be understood as: k t+1 <0.5, given r t A very small penalty of -100. The advantage of this setting is that it can improve the prediction accuracy of the model while making the training process more stable.

[0089] The calculation formula for determining the dominance can be expressed as: Indicates that the state node s k tThe expected return attribute obtained by inputting into the value network model. It should be noted that in the technical solution provided in this embodiment, the value network model and the aeration decision model may both include 5 fully connected hidden layers, with 256 nodes in each layer.

[0090] In this embodiment, based on the first loss function, each action evaluation result and the dominance, the model parameters of the aeration decision model are modified, including: for each action evaluation result, based on the preset parameters in the first loss function and the action evaluation result, a preset interval is determined; based on the preset interval, the action evaluation results are aligned to obtain a first evaluation result; based on the dominance and the action evaluation result, a second evaluation result is determined; based on the first evaluation result and the first evaluation result, a first loss value is determined; based on the first loss value, a parameter update step is determined, and based on the parameter update step, the model parameters of the aeration decision model are modified.

[0091] The preset parameter may be 1. The parameter update step size may refer to the size of adjusting the model parameters in each iteration or training, and controls the amplitude of the model parameter update.

[0092] Specifically, the preset parameters in the first loss function and the action evaluation result can be subjected to difference processing to obtain a left boundary value; the preset parameters in the first loss function and the action evaluation result can be summed to obtain a right boundary value; and a preset interval can be determined based on the left boundary value and the right boundary value. The action evaluation result is aligned (or normalized) within the preset interval to obtain a first evaluation result. The dominance and the action evaluation result can be multiplied to obtain a second evaluation result. Furthermore, the minimum value between the first evaluation result and the first evaluation result can be taken, and the negative of the minimum value can be used as the first loss value. The parameter update step size can be determined based on the size of the first loss value. Furthermore, the model parameters in the aeration decision model can be adjusted based on the parameter update step size.

[0093] Exemplarily, the first loss function can be expressed as: Among them, actor_loss represents the first loss value; min represents the minimum; clip represents alignment; max represents the maximum; Indicates that in state node s t (i.e., the first dissolved oxygen concentration) and execute a t The action evaluation result corresponding to the action (i.e. the sum of the air volume opening to be used); A represents the dominance.

[0094] In this embodiment, the cumulative reward and expected return attributes can also be processed for loss based on the fourth loss function, and the model parameters of the value network model can be corrected. The fourth loss function can be a mean square error function. Based on the fourth loss function, the mean square error of the cumulative reward and expected return attributes can be calculated as the loss value, and the model parameters of the value network model can be corrected according to the loss value, so as to improve the accuracy of the value network model in evaluating the expected return attributes corresponding to the aeration opening and action under the dissolved oxygen concentration.

[0095] S260, determining the trained aeration decision model as the aeration opening and determination model.

[0096] The technical solution of this embodiment trains an aeration decision model in this way, and uses the trained aeration decision model as a blower opening and determination model, thereby improving the accuracy of the air volume opening and determination of all blowers, meeting the sewage treatment needs of the biochemical pool, and thus improving the sewage treatment effect.

[0097] Embodiment 3

[0098] Figure 3 It is a flow chart of a blower aeration control method provided according to the third embodiment of the present invention. On the basis of the above-mentioned embodiment, before inputting the air volume opening to be used and the dissolved oxygen concentration to be used, as well as the preset operating parameters corresponding to the dissolved oxygen concentration to be used in the biochemical pool and the preset opening of the aeration regulating valve into the pre-trained dissolved oxygen concentration prediction model to obtain the predicted dissolved oxygen concentration, the dissolved oxygen concentration prediction model can also be trained. The specific implementation method can refer to the technical solution of this embodiment. Among them, the technical terms that are the same as or corresponding to the above-mentioned embodiments are not repeated here.

[0099] like Figure 3 As shown, the method specifically comprises the following steps:

[0100] S310, obtaining a plurality of first samples; the first samples include data to be used and dissolved oxygen concentration labels.

[0101] The data to be used include historical operating parameters of the biochemical pool, historical valve opening of the aeration regulating valve, and historical air volume opening of multiple blowers; the historical operating parameters at least include the dissolved oxygen concentration to be processed. The dissolved oxygen concentration tag can be used to characterize the actual dissolved oxygen concentration of the biochemical pool under the data to be used.

[0102] In order to make the obtained dissolved oxygen concentration prediction model have a higher accuracy, the first samples can be obtained as much and as rich as possible, so as to obtain the dissolved oxygen concentration prediction model based on the first sample training. It should be noted that the time interval between the first samples can be determined in combination with the time lag problem existing in the biochemical reaction, and the time interval should not be too long or too short. For example, for the blower aeration system, the length of the time lag is mainly affected by the organisms in the biochemical pool, temperature, etc. For example, the time interval is about 8-15 minutes.

[0103] S320: For each first sample, input the data to be used in the first sample into a dissolved oxygen concentration prediction model to be trained to obtain an estimated dissolved oxygen concentration.

[0104] In this embodiment, the data to be used in the first sample can be used as the input of the dissolved oxygen concentration prediction model to be trained to obtain an estimated dissolved oxygen concentration. The estimated dissolved oxygen concentration represents the dissolved oxygen concentration of the biochemical pool after executing the historical air volume openings of multiple blowers under the dissolved oxygen concentration to be processed in the first sample.

[0105] For example, see Figure 4 , Figure 4 It can be represented as a structural diagram of the dissolved oxygen concentration prediction model; the dissolved oxygen concentration prediction model can be a neural network; the entire neural network consists of 5 layers of hidden layers, input layers and output layers, and each layer of the neural network contains 256 nodes. External variables, actions and current states are used as inputs to the neural network, and the state-action pairs (s t+1 |s t ,a t ) in state s t Next, perform action a t The next state to be transferred to is s t+1 To make a prediction, that is, to predict the state of the next time step, that is, to predict the expected dissolved oxygen concentration. The state of the next time step can be used as the current state of the next time step to continue the prediction. Among them, the state is the dissolved oxygen concentration to be treated in the biochemical pool; the external variables include inlet temperature, inlet chemical oxygen demand, inlet ammonia nitrogen, inlet total phosphorus, inlet flow, biochemical pool sludge concentration and the historical valve opening of the aeration regulating valve; the action is the historical air volume opening of the blower.

[0106] S330: Determine a second loss value based on the second loss function, the dissolved oxygen concentration label in the first sample, and the estimated dissolved oxygen concentration, and modify the model parameters of the dissolved oxygen concentration prediction model based on the second loss value.

[0107] In this embodiment, the dissolved oxygen concentration label and the estimated dissolved oxygen concentration in the first sample can be subjected to loss processing according to the second loss function, and the obtained loss value can be used as the second loss value. Furthermore, the model parameters of the dissolved oxygen concentration prediction model can be adjusted according to the second loss value.

[0108] Exemplarily, the second loss function can be expressed as: Where N represents the number of first samples in a batch training. represents the prediction result of the i-th first sample (i.e. the expected dissolved oxygen concentration), Represents the historical real data of the i-th first sample (ie, the dissolved oxygen concentration label).

[0109] S340, taking the convergence of the second loss function as a training goal, and obtaining a trained dissolved oxygen concentration prediction model.

[0110] In this embodiment, the training error of the second loss function, that is, the loss parameter, can be used as a condition for detecting whether the second loss function has reached convergence at present, such as whether the training error is less than the preset error or whether the error change trend tends to be stable, or whether the current number of iterations is equal to the preset number. If the detection reaches the convergence condition, such as the training error of the second loss function reaches less than the preset error or the error change tends to be stable, it indicates that the training of the dissolved oxygen concentration prediction model to be trained is completed, and the iterative training can be stopped at this time. If it is detected that the convergence condition is not reached at present, the first sample can be further obtained to train the dissolved oxygen concentration prediction model until the training error of the second loss function is within the preset range. When the training error of the second loss function reaches convergence, the dissolved oxygen concentration prediction model at this time can be used as the dissolved oxygen concentration prediction model obtained by the final training. Based on the dissolved oxygen concentration prediction model, the predicted dissolved oxygen concentration is determined according to the air volume opening to be used and the dissolved oxygen concentration to be used, as well as the preset operating parameters corresponding to the dissolved oxygen concentration to be used in the biochemical pool and the preset opening of the aeration regulating valve.

[0111] The technical solution of this embodiment is to obtain multiple first samples; input the data to be used in the first samples into the dissolved oxygen concentration prediction model to be trained to obtain the expected dissolved oxygen concentration, and then determine the second loss value based on the second loss function, the dissolved oxygen concentration label in the first sample and the expected dissolved oxygen concentration, and correct the model parameters of the dissolved oxygen concentration prediction model based on the second loss value; take the convergence of the second loss function as the training goal to obtain a trained dissolved oxygen concentration prediction model, and determine the dissolved oxygen concentration corresponding to each state node in the state transition trajectory based on the dissolved oxygen concentration prediction model, so as to improve the reliability and accuracy of the dissolved oxygen concentration determination.

[0112] Embodiment 4

[0113] Figure 51 is a schematic diagram of the structure of a blower aeration control device according to Embodiment 4 of the present invention. The device is configured in a sewage treatment system, wherein the sewage treatment system includes a biochemical pool and a plurality of blowers connected by an aeration pipe; an aeration regulating valve is provided on the aeration pipe; Figure 5 As shown, the device includes: a data acquisition module 410, an air volume opening and determination module 420, a target air volume opening determination module 430 and a control module 440.

[0114] Among them, the data acquisition module 410 is used to obtain the current operating parameters of the biochemical pool and the current opening of the aeration regulating valve; wherein the current operating parameters include at least dissolved oxygen concentration, water inlet parameters and sludge concentration; the air volume opening and determination module 420 is used to input the current operating parameters and the current opening into the pre-trained blower opening and determination model to obtain the air volume opening and to be processed; wherein the blower opening and determination model is trained based on the dissolved oxygen concentration corresponding to each state node in the state transition trajectory of the biochemical pool; the dissolved oxygen concentration corresponding to the state node is pre-set and / or determined based on the dissolved oxygen concentration corresponding to the previous state node and the air volume opening and to be used; the target air volume opening determination module 430 is used to process the air volume opening and to be processed according to the target optimization function to obtain the target air volume opening of each blower; the control module 440 is used to control each blower to deliver air volume to the biochemical pool according to the target air volume opening corresponding to it.

[0115] The technical solution of this embodiment obtains the current operating parameters of the biochemical pool and the current opening of the aeration regulating valve; inputs the current operating parameters and the current opening into the pre-trained blower opening and determination model to obtain the to-be-processed air volume opening and; the blower opening and determination model is obtained by training based on the dissolved oxygen concentration corresponding to each state node in the state transition trajectory of the biochemical pool; the dissolved oxygen concentration corresponding to the state node is pre-set and / or determined based on the dissolved oxygen concentration corresponding to the previous state node and the to-be-used air volume opening and; processes the to-be-processed air volume opening and and according to the target optimization function to obtain the target air volume opening of each blower; controls each blower to deliver air volume to the biochemical pool according to the corresponding target air volume opening, thereby solving the problem of the prior art relying on Artificially controlling the aeration flow of the blower according to the actual sewage treatment needs leads to problems such as poor aeration control accuracy, high energy consumption and poor sewage treatment effect. By using the blower opening and determination model to process the current operating parameters of the biochemical pool and the current opening of the aeration regulating valve, the sum of the to-be-treated air volume openings corresponding to all blowers is obtained. Then, the sum of the to-be-treated air volume openings is processed according to the target optimization function, and the energy consumption of the target air volume opening of each blower is controlled to minimize energy consumption. By adjusting the opening of each blower according to the target air volume opening, the flow rate of the transported gas can be accurately controlled, which improves the accuracy of the blower aeration control while ensuring that the biochemical pool obtains the required amount of gas, thereby improving the sewage treatment effect.

[0116] Based on the above device, optionally, the target optimization function includes an energy consumption optimization function and at least one constraint condition; the target air volume opening determination module 430 includes:

[0117] An air volume opening set determination unit, configured to process the to-be-processed air volume opening sum according to at least one constraint condition to obtain an air volume opening set; wherein the constraint condition at least includes that the sum of the air volume openings of each of the blowers in the air volume opening set is consistent with the air volume opening sum;

[0118] The energy consumption determination unit is used to determine the energy consumption corresponding to the air volume opening set according to the energy consumption optimization function, and to determine the target air volume opening of each blower based on the air volume opening set when the target is reached, with the goal of minimizing the energy consumption.

[0119] On the basis of the above device, optionally, the device further comprises: a blast opening and determination model determination module, which is used to train and obtain the blast opening and determination model; wherein the blast opening and determination model determination module comprises:

[0120] A state node determination unit, used to determine the state node in the state transfer trajectory; wherein the state node corresponds to the dissolved oxygen concentration to be used in the biochemical pool;

[0121] The opening and distribution information determination unit is used to input the dissolved oxygen concentration to be used, the preset operating parameters of the biochemical pool corresponding to the dissolved oxygen concentration to be used, and the preset opening of the aeration regulating valve into the aeration decision model to obtain the opening and distribution information; wherein the opening and distribution information includes a plurality of air volume openings to be used and;

[0122] A predicted dissolved oxygen concentration determination unit is used to input the sum of the air volume opening to be used, the dissolved oxygen concentration to be used, the preset operating parameters of the biochemical pool corresponding to the dissolved oxygen concentration to be used, and the preset opening of the aeration regulating valve into a pre-trained dissolved oxygen concentration prediction model for each of the air volume opening to be used, to obtain a predicted dissolved oxygen concentration, and use the predicted dissolved oxygen concentration as the next state node of the state node;

[0123] An evaluation unit, for inputting the first dissolved oxygen concentration corresponding to each state node into the value network model, obtaining the expected return attribute corresponding to the aeration opening and the action executed at the first dissolved oxygen concentration, and determining the evaluation result of executing each of the air volume openings to be used and the corresponding action respectively at the first dissolved oxygen concentration;

[0124] an aeration decision model determination unit, configured to modify the model parameters of the aeration decision model based on the first loss function, each of the action evaluation results and the expected reward attribute, so as to obtain a trained aeration decision model;

[0125] The air blowing opening and determination model determination unit is used to determine the trained aeration decision model as the air blowing opening and determination model; wherein the first dissolved oxygen concentration is the dissolved oxygen concentration to be used and / or the dissolved oxygen concentration in the predicted dissolved oxygen concentration.

[0126] On the basis of the above device, optionally, the device further includes:

[0127] A re-execution unit is used to, when the number of state nodes in the state transfer trajectory does not reach a preset threshold, use the predicted dissolved oxygen concentration as the dissolved oxygen concentration to be used, and re-execute the steps of inputting the dissolved oxygen concentration to be used into the aeration decision model, obtaining the opening and distribution information, obtaining the predicted dissolved oxygen concentration, and determining whether the number of state nodes reaches the preset threshold.

[0128] On the basis of the above device, optionally, the evaluation unit includes:

[0129] An initial training unit, for initial training, for determining, based on the opening and distribution information, to respectively execute each of the air volume openings to be used and the corresponding action evaluation results at the first dissolved oxygen concentration;

[0130] A non-first training unit is used to determine, for training after the first training, based on the opening and distribution information and the historical opening and distribution information obtained from the previous training, to execute each of the air volume openings to be used and the corresponding action evaluation results under the first dissolved oxygen concentration.

[0131] Based on the above device, optionally, the aeration decision model determination unit includes:

[0132] A dominance determination unit, configured to determine a dominance based on the expected reward attribute and a second dissolved oxygen concentration corresponding to a state node subsequent to the state node; wherein the second dissolved oxygen concentration is a dissolved oxygen concentration in the to-be-used dissolved oxygen concentration and / or the predicted dissolved oxygen concentration;

[0133] A correction unit is used to correct the model parameters of the aeration decision model based on the first loss function, each action evaluation result and the dominance.

[0134] Based on the above device, optionally, the advantage determination unit includes:

[0135] a cumulative reward determining unit, configured to determine a cumulative reward based on a second dissolved oxygen concentration corresponding to a state node subsequent to the state node;

[0136] The advantage determination subunit is used to determine the advantage based on the expected return attribute and the accumulated reward.

[0137] On the basis of the above device, optionally, the correction unit includes:

[0138] a preset interval determining unit, configured to determine a preset interval for each of the action evaluation results based on a preset parameter in the first loss function and the action evaluation result;

[0139] A first evaluation result determining unit, configured to align the action evaluation results based on the preset interval to obtain a first evaluation result;

[0140] A second evaluation result determining unit, configured to determine a second evaluation result based on the dominance and the action evaluation result;

[0141] a first loss value determining unit, configured to determine a first loss value based on the first evaluation result and the second evaluation result;

[0142] A parameter updating step length determining unit is used to determine a parameter updating step length based on the first loss value, and to modify a model parameter of the aeration decision model based on the parameter updating step length.

[0143] On the basis of the above device, optionally, the device further comprises: a dissolved oxygen concentration prediction model determination module, which is used to train and obtain a dissolved oxygen concentration prediction model; wherein the dissolved oxygen concentration prediction model determination module comprises:

[0144] A first sample determination unit is used to obtain a plurality of first samples; wherein the first samples include data to be used and dissolved oxygen concentration labels; the data to be used include historical operating parameters of the biochemical pool, historical valve openings of the aeration regulating valve, and historical air volume openings of the plurality of blowers; the historical operating parameters include at least the dissolved oxygen concentration to be processed;

[0145] an estimated dissolved oxygen concentration determination unit, configured to input the to-be-used data in the first sample into a dissolved oxygen concentration prediction model to be trained for each of the first samples, so as to obtain an estimated dissolved oxygen concentration;

[0146] a second loss value determining unit, configured to determine a second loss value based on a second loss function, a dissolved oxygen concentration label in the first sample, and the estimated dissolved oxygen concentration, and to modify a model parameter of the dissolved oxygen concentration prediction model based on the second loss value;

[0147] The dissolved oxygen concentration prediction model determination unit is used to obtain a trained dissolved oxygen concentration prediction model by taking the convergence of the second loss function as a training target.

[0148] On the basis of the above device, optionally, the device further includes:

[0149] A second sample determination unit, used to obtain a plurality of second samples, wherein the second samples include an air volume opening group and an energy consumption label; the air volume opening group includes a plurality of air volume openings of the blower;

[0150] A predicted energy consumption determination unit, for each of the second samples, inputting the air volume opening group in the second sample into the optimization function to be trained to obtain the predicted energy consumption;

[0151] a third loss value determining unit, configured to determine a third loss value based on a third loss function, the energy consumption label in the second sample and the predicted energy consumption, and to modify parameters of the optimization function to be trained based on the third loss value;

[0152] The target optimization function determination unit is used to obtain the target optimization function with the convergence of the third loss function as the goal.

[0153] The blower aeration control device provided in the embodiment of the present invention can execute the blower aeration control method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0154] Embodiment 5

[0155] Figure 6 Schematic diagram of the structure of an electronic device for implementing the blower aeration control method of an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0156] like Figure 6 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0157] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0158] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the blower aeration control method.

[0159] In some embodiments, the blower aeration control method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the blower aeration control method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the blower aeration control method in any other appropriate manner (e.g., by means of firmware).

[0160] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0161] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0162] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. 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.

[0163] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0164] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by 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), a blockchain network, and the Internet.

[0165] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0166] An embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements the blower aeration control method provided by any embodiment of the present invention.

[0167] In the process of implementation, the computer program product can be written in one or more programming languages ​​or a combination thereof to perform the computer program code of the present invention, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can 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 can 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 can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0168] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0169] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A blower aeration control method, characterized in that: Applied to a sewage treatment system, the sewage treatment system includes a biochemical pool and a plurality of blowers connected by an aeration pipeline; an aeration regulating valve is provided on the aeration pipeline; the method includes: Obtaining the current operating parameters of the biochemical pool and the current opening of the aeration regulating valve; wherein the current operating parameters at least include dissolved oxygen concentration, influent parameters and sludge concentration; The current operating parameters and the current opening are input into a pre-trained blast opening and determination model to obtain the opening and of the air volume to be processed; wherein the blast opening and determination model is trained based on the dissolved oxygen concentration corresponding to each state node in the state transition trajectory of the biochemical pool; the dissolved oxygen concentration corresponding to the state node is preset and / or determined based on the dissolved oxygen concentration corresponding to the previous state node and the opening and of the air volume to be used; Processing the air volume opening to be processed and the air volume opening to be processed according to the target optimization function to obtain the target air volume opening of each of the blowers; Each of the blowers is controlled to deliver air to the biochemical pool according to the corresponding target air volume opening.

2. The method according to claim 1, characterized in that The target optimization function includes an energy consumption optimization function and at least one constraint condition; The air volume opening and the target air volume opening of each blower are processed according to the target optimization function, and the target air volume opening of each blower is obtained, including: Processing the to-be-processed air volume opening sum according to at least one constraint condition to obtain an air volume opening set; wherein the constraint condition at least includes that the sum of the air volume openings of each of the blowers in the air volume opening set is consistent with the air volume opening sum; According to the energy consumption optimization function, the energy consumption corresponding to the air volume opening set is determined, and with the goal of minimizing the energy consumption, the target air volume opening of each blower is determined based on the air volume opening set when the target is reached.

3. The method according to claim 1, characterized in that Before inputting the current operating parameters and the current opening into the pre-trained air blast opening and determination model to obtain the air volume opening and to be processed, the method further includes: The training obtains the air blast opening and determines the model; among them, The training to obtain the air blast opening and determine the model includes: Determine a state node in the state transfer trajectory; wherein the state node corresponds to the dissolved oxygen concentration to be used in the biochemical pool; Input the dissolved oxygen concentration to be used, the preset operating parameters of the biochemical pool corresponding to the dissolved oxygen concentration to be used, and the preset opening of the aeration regulating valve into the aeration decision model to obtain opening and distribution information; wherein the opening and distribution information includes a plurality of air volume openings to be used and; For each of the air volume openings to be used, the air volume openings to be used, the dissolved oxygen concentration to be used, the preset operating parameters of the biochemical pool corresponding to the dissolved oxygen concentration to be used, and the preset opening of the aeration regulating valve are input into a pre-trained dissolved oxygen concentration prediction model to obtain a predicted dissolved oxygen concentration, and the predicted dissolved oxygen concentration is used as the next state node of the state node; For each of the state nodes, the first dissolved oxygen concentration corresponding to the state node is input into the value network model to obtain the expected return attribute corresponding to the aeration opening and the action executed at the first dissolved oxygen concentration, and determine the evaluation results of each of the air volume openings to be used and the corresponding actions executed at the first dissolved oxygen concentration; Based on the first loss function, each of the action evaluation results and the expected reward attribute, the model parameters of the aeration decision model are modified to obtain a trained aeration decision model; The trained aeration decision model is determined as the air blowing opening and determination model; The first dissolved oxygen concentration is the dissolved oxygen concentration to be used and / or the predicted dissolved oxygen concentration.

4. The method according to claim 3, characterized in that After taking the predicted dissolved oxygen concentration as the next state node of the state node, the method further includes: When the number of state nodes in the state transition trajectory does not reach a preset threshold, the predicted dissolved oxygen concentration is used as the dissolved oxygen concentration to be used, and the steps of inputting the dissolved oxygen concentration to be used into the aeration decision model to obtain the opening and distribution information, obtaining the predicted dissolved oxygen concentration, and determining whether the number of state nodes reaches the preset threshold are re-executed.

5. The method according to claim 3, characterized in that: The determining to respectively execute each of the air volume openings to be used and the corresponding action evaluation results under the first dissolved oxygen concentration includes: For the first training, based on the opening and distribution information, determining to respectively execute each of the air volume openings to be used and the corresponding action evaluation results under the first dissolved oxygen concentration; For the training after the first training, based on the opening and distribution information and the historical opening and distribution information obtained from the previous training, it is determined to execute each of the air volume openings to be used and the corresponding action evaluation results respectively under the first dissolved oxygen concentration.

6. The method according to claim 3, characterized in that The modifying of the model parameters of the aeration decision model based on the first loss function, each of the action evaluation results and the expected reward attribute comprises: Determine the dominance based on the expected reward attribute and a second dissolved oxygen concentration corresponding to a state node after the state node; wherein the second dissolved oxygen concentration is a dissolved oxygen concentration in the to-be-used dissolved oxygen concentration and / or the predicted dissolved oxygen concentration; Based on the first loss function, each of the action evaluation results and the dominance, the model parameters of the aeration decision model are modified.

7. The method according to claim 6, characterized in that The determining of the dominance based on the expected reward attribute and a second dissolved oxygen concentration corresponding to a state node after the state node includes: determining a cumulative reward based on a second dissolved oxygen concentration corresponding to a state node subsequent to the state node; Based on the expected return attribute and the accumulated reward, a degree of dominance is determined.

8. The method according to claim 6, characterized in that The modifying of the model parameters of the aeration decision model based on the first loss function, each of the action evaluation results and the dominance degree includes: For each of the action evaluation results, determining a preset interval based on a preset parameter in the first loss function and the action evaluation result; Aligning the action evaluation results based on the preset interval to obtain a first evaluation result; determining a second evaluation result based on the dominance and the action evaluation result; determining a first loss value based on the first evaluation result and the first evaluation result; A parameter update step length is determined based on the first loss value, and model parameters of the aeration decision model are modified based on the parameter update step length.

9. The method according to claim 3, characterized in that: Before inputting the air volume opening to be used and the dissolved oxygen concentration to be used, the preset operating parameters of the biochemical pool corresponding to the dissolved oxygen concentration to be used, and the preset opening of the aeration regulating valve into the pre-trained dissolved oxygen concentration prediction model to obtain the predicted dissolved oxygen concentration, the method further includes: The dissolved oxygen concentration prediction model is obtained by training; The training obtains a dissolved oxygen concentration prediction model, including: Acquire multiple first samples; wherein the first samples include data to be used and dissolved oxygen concentration labels; the data to be used include historical operating parameters of the biochemical pool, historical valve openings of the aeration regulating valve, and historical air volume openings of the multiple blowers; the historical operating parameters include at least the dissolved oxygen concentration to be processed; For each of the first samples, inputting the to-be-used data in the first sample into a dissolved oxygen concentration prediction model to be trained to obtain an estimated dissolved oxygen concentration; Determine a second loss value based on a second loss function, a dissolved oxygen concentration label in the first sample, and the estimated dissolved oxygen concentration, and modify a model parameter of the dissolved oxygen concentration prediction model based on the second loss value; The convergence of the second loss function is taken as a training goal to obtain a trained dissolved oxygen concentration prediction model.

10. The method according to claim 1, characterized in that Before processing the air volume opening to be processed and the target air volume opening of each blower according to the target optimization function to obtain the target air volume opening of each blower, the method further includes: Acquire a plurality of second samples, wherein the second samples include an air volume opening group and an energy consumption label; the air volume opening group includes a plurality of air volume openings of the blowers; For each of the second samples, inputting the air volume opening group in the second sample into the optimization function to be trained to obtain predicted energy consumption; Determine a third loss value based on a third loss function, the energy consumption label in the second sample and the predicted energy consumption, and modify the parameters of the optimization function to be trained based on the third loss value; Taking the convergence of the third loss function as a goal, a target optimization function is obtained.

Images