Flow control system and method for excess pressure power generation of water plant

Abstract

The invention belongs to the technical field of power generation control, and provides a flow control system and method for excess pressure power generation of a water plant. Acquiring a purified water flow data set and a purified water pressure data set of an original purified water discharge pipeline of the sewage treatment plant within a specified duration, and predicting to obtain a purified water discharge stability level; a flow regulation and control model is obtained according to purified water discharge stability grade matching; the flow control model receives the first real-time flow of the main pipeline flowmeter, the real-time water pressure data of the pressure meter and the second real-time flow of the bypass pipeline flowmeter, the flow control model outputs the predicted first opening degree of the main pipeline electric valve and the predicted second opening degree of the bypass pipeline outlet electric valve, and the first opening degree and the second opening degree are executed. The stability of the output electric power of the water purification residual pressure power generation device can be ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of power generation control, and in particular, to a flow control system and method for power generation using the residual pressure of a water treatment plant. Background Art

[0002] In today's society, with the acceleration of the industrialization process and the continuous growth of the population, the demand for water resources and the treatment scale are becoming increasingly large. As a key facility to ensure the safety of urban water supply, the operation process of a water purification plant involves a large amount of water pressure operations. Currently, in the water purification process, after the raw water undergoes a series of complex purification treatments, purified water with a certain pressure is generated. Usually, during the process of sending the purified water to the user side, most of the residual pressure carried by it is directly consumed or wasted and not effectively utilized.

[0003] Currently, technologies for generating electricity using the residual pressure of purified water have emerged. As Figure 1 shown, the system includes two pipelines: the main pipeline (the main pipeline is the original pipeline for discharging purified water from the sewage treatment plant) and the bypass pipeline (the bypass pipeline is the pipeline for installing the residual pressure power generation device of the sewage treatment plant). The electricity generated by the residual pressure power generation device of the purified water needs to meet the grid connection requirements, which requires ensuring the stability of the water head (the water head refers to the vertical height difference (potential energy) when discharging purified water and is the core energy source for driving the water turbine to generate electricity). The stable control of the water head is achieved by regulating the opening degrees of the electric valve of the main pipeline and the electric valve at the outlet of the bypass pipeline. Currently, the existing technology for regulating the opening degrees of the electric valve of the main pipeline and the electric valve at the outlet of the bypass pipeline is mainly rule-based regulation, that is, based on the corresponding relationship between the real-time flow rate detected by the flow meter and the rules to select the corresponding valve regulation parameters. Since the formulated rules often cannot achieve a very high level of refinement, the regulation effect of this regulation method is relatively rough, and the power generation power of the residual pressure power generation device of the purified water is not stable enough.

[0004] Therefore, how to perform more accurate flow regulation on the residual pressure power generation system of purified water to ensure power generation stability is a technical problem that urgently needs to be solved at present. Summary of the Invention

[0005] In response to this, the present invention provides a flow control method, system, electronic device, computer storage medium, and computer program product for power generation using the residual pressure of a water treatment plant to solve the above technical problems.

[0006] The present invention discloses a flow control method for power generation using the residual pressure of a water treatment plant, and the method includes the following steps: Obtain the purified water flow rate data set and the purified water pressure data set of the original pipeline for discharging purified water from the sewage treatment plant within a specified time period, and predict the stable level of purified water discharge based on the purified water flow rate data set and the purified water pressure data set; A flow rate regulation model is obtained by matching according to the stable level of the purified water discharge. Specifically, if the stable level of the purified water discharge is higher than a preset level, the first flow rate regulation model is selected as the flow rate regulation model; otherwise, the second flow rate regulation model is selected as the flow rate regulation model. Among them, the regulation accuracy of the first flow rate regulation model is lower than that of the second flow rate regulation model; Receive the first real-time flow rate of the main pipeline flowmeter and the real-time water pressure data of the pressure gauge, as well as the second real-time flow rate of the bypass pipeline flowmeter. Input the first real-time flow rate, the real-time water pressure data, the second real-time flow rate, and a preset water head into the flow rate regulation model. The flow rate regulation model outputs the first opening degree of the main pipeline electric valve and the second opening degree of the bypass pipeline outlet electric valve, and executes the first opening degree and the second opening degree.

[0007] In some embodiments, the specified duration is determined by the following method: Obtain the inner wall roughness of the original pipeline, and obtain the layout information of the original pipeline. Evaluate the pipeline layout complexity according to the layout information; Perform a matching operation on the inner wall roughness of the pipeline and the pipeline layout complexity with a matching relationship to obtain the specified duration. Among them, in the matching relationship, both the inner wall roughness of the pipeline and the pipeline layout complexity are positively correlated with the specified duration.

[0008] In some embodiments, predicting the stable level of the purified water discharge based on the purified water flow rate dataset and the purified water pressure dataset includes: Obtain an intercept duration by matching according to the specified duration. Respectively intercept a target purified water flow rate dataset and a target purified water pressure dataset from the purified water flow rate dataset and the purified water pressure dataset according to the intercept duration. Among them, the intercept duration is the duration starting from the end moment of the specified duration and towards the starting moment of the purified water discharge; Use a convolutional network to extract features from the target purified water flow rate dataset and the target purified water pressure dataset, splice the extracted features, obtain a purified water fluctuation feature, and input the purified water fluctuation feature into a prediction model to obtain the stable level of the purified water discharge.

[0009] In some embodiments, the second flow rate regulation model is trained by the following method: Collect and construct a training dataset. Each piece of training data in the training dataset includes: the real-time flow rate and real-time water pressure of the original pipeline, the real-time flow rate of the bypass pipeline, the opening degree of the main pipeline electric valve and the opening degree of the bypass pipeline outlet electric valve, and label data. Among them, the label data is a stability evaluation value of the preset water head; Divide each piece of training data in the training data set into a training set and a test set. Use the training set to train the second flow regulation model. After the training reaches a preset level, use the test set to test the second flow regulation model. If the test is passed, end the training; if the test is not passed, continue to use other training data in the training set for training until the test is passed.

[0010] In some embodiments, the training data used for training the prediction model is obtained based on multiple groups of historical discharge data of the original water purification discharge pipeline of the sewage treatment plant. The historical discharge data includes the water purification flow rate data and the water purification pressure data corresponding to the specified duration and a certain period after the specified duration.

[0011] The present invention also discloses a flow control system for residual pressure power generation in a water plant. The system includes a water purification discharge stability level prediction unit, a flow regulation model screening unit, and a flow regulation unit. The water purification discharge stability level prediction unit obtains the water purification flow rate data set and the water purification pressure data set of the original water purification discharge pipeline of the sewage treatment plant within a specified duration, and predicts the water purification discharge stability level based on the water purification flow rate data set and the water purification pressure data set. The flow regulation model screening unit matches a flow regulation model according to the water purification discharge stability level. Specifically, if the water purification discharge stability level is higher than the preset level, select the first flow regulation model as the flow regulation model; otherwise, select the second flow regulation model as the flow regulation model. Among them, the regulation accuracy of the first flow regulation model is lower than that of the second flow regulation model. The flow regulation unit receives the first real-time flow rate of the main pipeline flowmeter, the real-time water pressure data of the pressure gauge, and the second real-time flow rate of the bypass pipeline flowmeter, and inputs the first real-time flow rate, the real-time water pressure data, the second real-time flow rate, and a preset water head to the flow regulation model. The flow regulation model outputs the first opening degree of the main pipeline electric valve and the second opening degree of the bypass pipeline outlet electric valve, and executes the first opening degree and the second opening degree.

[0012] In some embodiments, the water purification discharge stability level prediction unit is used for: Match an interception duration based on the specified duration, and respectively intercept a target water purification flow rate data set and a target water purification pressure data set from the water purification flow rate data set and the water purification pressure data set according to the interception duration. Among them, the interception duration is the duration starting from the end moment of the specified duration and towards the starting moment of the water purification discharge. Use a convolutional network to extract features from the target clean water flow dataset and the target clean water pressure dataset, splice the extracted features to obtain clean water fluctuation features, and input the clean water fluctuation features into a prediction model to obtain the stable level of clean water discharge.

[0013] The present invention also discloses an electronic device, which includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor. When the computer program is executed, it implements the method described in any one of the preceding items.

[0014] The present invention also discloses a computer storage medium, which stores a computer program. When the computer program is executed, it implements the method described in any one of the preceding items.

[0015] The present invention also discloses a computer program product, which has computer program code. When the computer program code is executed, it implements the method described in any one of the preceding items.

[0016] The beneficial effects of the present invention are as follows: The flow control method of the present invention can significantly improve the pertinence of flow regulation by accurately evaluating the stability of clean water discharge and flexibly selecting a suitable flow regulation model, can better adapt to the fluctuating changes of clean water discharge, stabilize the power generation power, improve the power quality, facilitate grid connection, and can also reduce the impact loss of equipment and extend the service life of equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 is a schematic diagram of the basic principle of a clean water residual pressure power generation system; Figure 2 is a flowchart of a flow control method for residual pressure power generation in a water plant disclosed in an embodiment of the present invention; Figure 3 is a schematic diagram of the structure of a flow control system for residual pressure power generation in a water plant disclosed in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] The following specific embodiments illustrate the implementation manners of the present application. Those skilled in the art can easily understand other advantages and effects of the present application from the content disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts belong to the scope protected by the present application.

[0020] In addition, the technical features involved in different implementation manners of the present application described below can be combined with each other as long as they do not conflict with each other.

[0021] The control method and control system of the present invention are based on Figure 1 the shown water purification residual pressure power generation system and the principles described in the background art.

[0022] As Figure 2 shown, an embodiment of the present invention discloses a flow control method for residual pressure power generation in a waterworks. The method includes the following steps: S1, obtain a water purification flow rate data set and a water purification pressure data set of the original pipeline for water purification discharge in a sewage treatment plant within a specified time period, and predict the water purification discharge stability level based on the water purification flow rate data set and the water purification pressure data set.

[0023] A pressure gauge and a main pipeline flow meter are arranged in the original pipeline for water purification discharge in the sewage treatment plant. Within the specified time period when starting to discharge water purification, multiple groups of data collected by the pressure gauge and the main pipeline flow meter are received, respectively constituting the water purification flow rate data set and the water purification pressure data set. These time series data reflect the dynamic changes of water purification discharge.

[0024] Then, a prediction model is pre-constructed. The above-mentioned water purification flow rate data set and water purification pressure data set are used as the input data of the prediction model, and the water purification discharge stability level of this water purification discharge operation is predicted by the prediction model. The water purification discharge stability level reflects the fluctuation amplitude, that is, the stability, of the water purification water flow in the original pipeline in terms of unit flow rate and water pressure in the predicted water purification discharge operation (the corresponding water purification discharge stability levels in the next multiple time periods can be obtained through multi-step prediction). The higher the water purification discharge stability level, the smaller the fluctuation amplitude, that is, the higher the stability, of the water purification water flow in terms of unit flow rate and water pressure in the next time period. On the contrary, it indicates that the fluctuation amplitude of the water purification water flow in terms of unit flow rate and water pressure is larger, that is, the stability is lower.

[0025] It should be noted that the prediction model can be constructed based on algorithms such as LSTM and CNN. The present invention does not make specific limitations on this.

[0026] S2. Based on the obtained stable level of purified water discharge, a flow rate regulation model is matched. Specifically, if the stable level of purified water discharge is higher than the preset level, the first flow rate regulation model is selected as the flow rate regulation model; otherwise, the second flow rate regulation model is selected as the flow rate regulation model. Among them, the regulation accuracy of the first flow rate regulation model is lower than that of the second flow rate regulation model.

[0027] After obtaining the stable level of purified water discharge, it is compared with the preset level that is determined in advance. The preset level is a measurement standard determined based on long-term practical experience and expectations for system performance. If the predicted stable level of purified water discharge is higher than the preset level, it indicates that the current purified water discharge state is relatively stable, and the fluctuations of water flow and pressure are within an acceptable small range. In this case, the first flow rate regulation model is selected. The regulation accuracy of this model is relatively low, but its advantage lies in that the calculation process is relatively simple, it can respond quickly, has low requirements for the computing power of the equipment, and can reduce the wear on the equipment caused by frequent fine regulation. It is sufficient to meet the flow rate control requirements under stable operating conditions, thereby reducing the system operation cost.

[0028] On the contrary, if the stable level of purified water discharge is lower than the preset level, it means that the discharge state is unstable and the changes in water flow and pressure are relatively complex. At this time, the second flow rate regulation model with higher regulation accuracy is enabled. This model can more sensitively capture the subtle changes in flow rate and pressure and give more accurate regulation strategies to cope with the unstable discharge situation, ensuring that the system can operate stably under complex and changeable conditions and ensuring that the generated electricity can meet the grid connection requirements.

[0029] It should be noted that the first flow rate regulation model can be constructed based on classification control algorithms such as SVM and random forest, while the second flow rate regulation model is based on, for example, a deep Q network.

[0030] S3. Receive the first real-time flow rate of the main pipeline flowmeter, the real-time water pressure data of the pressure gauge, and the second real-time flow rate of the bypass pipeline flowmeter. Input the first real-time flow rate, the real-time water pressure data, the second real-time flow rate, and the preset water head into the flow rate regulation model. The flow rate regulation model outputs the first opening degree of the main pipeline electric valve and the second opening degree of the bypass pipeline outlet electric valve, and executes the first opening degree and the second opening degree.

[0031] During the power generation process of the clean water residual pressure power generation device, the main pipeline flowmeter continuously measures and feeds back the first real-time flow rate, the pressure gauge real-time collects and feeds back the water pressure data, and the bypass pipeline flowmeter provides the second real-time flow rate. At the same time, a water head is specified in advance according to the power generation target, for example, 30 meters. These real-time data and the preset water head are input into the flow rate regulation model selected in the previous step. The flow rate regulation model fully considers the mutual relationship among the flow rate, pressure, and water head, and finally outputs the first opening degree of the main pipeline electric valve and the second opening degree of the bypass pipeline outlet electric valve. These two opening degree values precisely determine the flow rate distribution of the water flow in the two pipelines. Through the precise control of the water flow, the stable control of the water head is achieved, ensuring the stability of the electric power generated by the clean water residual pressure power generation device.

[0032] Taking the second flow rate regulation model, the deep Q network, as an example for illustration: construct a deep neural network as the Q function estimator, with the input being the current state of the system (including the first real-time flow rate, real-time water pressure data, second real-time flow rate, and preset water head), and the output being the Q values of various possible actions (different valve opening degree combinations) in this state. Through continuous iterative training, the network learns the valve opening degree adjustment strategy that can obtain the maximum cumulative reward (such as making the power generation power most stable) in different states.

[0033] The flow rate control method of the present invention can significantly improve the pertinence of flow rate regulation by precisely evaluating the stability of clean water discharge and flexibly selecting a suitable flow rate regulation model. It can better adapt to the fluctuating changes of clean water discharge, stabilize the power generation power, improve the power quality, facilitate grid connection, reduce the impact loss of equipment, and extend the service life of equipment.

[0034] In some embodiments, the specified duration is determined in the following manner: Obtain the inner wall roughness of the original pipeline and the layout information of the original pipeline, and evaluate the pipeline layout complexity according to the layout information; Perform a matching operation on the inner wall roughness of the pipeline and the pipeline layout complexity with the matching relationship to obtain the specified duration; wherein, in the matching relationship, both the inner wall roughness of the pipeline and the pipeline layout complexity are positively correlated with the specified duration.

[0035] In the embodiments of the present invention, the inner wall roughness of the pipeline is one of the key factors affecting the water flow state. The inner wall of the original pipeline is periodically detected in advance by, for example, a roughness meter. The inner wall roughness reflects the unevenness of the inner wall surface of the pipeline, which directly affects the friction between the water flow and the pipe wall. The rougher the inner wall (the more serious the corrosion and scaling), the greater the energy loss of the water flow, the worse the water flow stability, and the longer the time required to reach the stable state.

[0036] The pipeline layout information covers the pipeline orientation, the presence of components such as elbows and tees, as well as their quantity and distribution. After obtaining this information through on-site surveys, reviewing pipeline design drawings, etc., the complexity of the pipeline layout is evaluated. A complex pipeline layout means that the water flow needs to frequently change direction and velocity during transportation, which easily generates vortex and impact phenomena, and it takes more time for the water flow to stabilize. A set of matching relationships is established in advance. This set of matching relationships clarifies the connection between the inner wall roughness of the pipeline, the complexity of the pipeline layout, and a specified duration. In this set of matching relationships, both the inner wall roughness of the pipeline and the complexity of the pipeline layout are positively correlated with the specified duration. That is to say, the higher the inner wall roughness of the pipeline and the greater the complexity of the pipeline layout, the longer the corresponding specified duration.

[0037] The above positive correlation is established based on a large amount of experimental data, actual engineering experience, and relevant fluid mechanics theories. For example, in the experiments of simulating the water flow in pipelines with different roughnesses and layout complexities, the time required for the water flow to stabilize from the start of discharge is recorded. After multiple experiments and data analysis, the quantitative relationship between the two and the duration is obtained. Substitute the obtained inner wall roughness data of the pipeline and the evaluated pipeline layout complexity data into the established matching relationships for matching operations simultaneously, and finally obtain the specified duration corresponding to the characteristics of this pipeline. The clean water flow rate dataset and clean water pressure dataset obtained based on this specified duration can more accurately reflect the true fluctuation characteristics of the discharged clean water under the conditions of this pipeline after overcoming external interference factors.

[0038] It should be noted that although the clean water flow rate dataset and clean water pressure dataset within the specified duration also contain the previous data affected by the inner wall roughness of the pipeline and the complexity of the pipeline layout, they already contain sufficient subsequent data that can be used to reflect the true fluctuation characteristics of the clean water. Therefore, as a whole, it can meet the subsequent analysis requirements.

[0039] In some embodiments, predicting the stable level of clean water discharge based on the clean water flow rate dataset and the clean water pressure dataset includes: Based on the specified duration, an intercept duration is matched. According to the intercept duration, a target clean water flow rate dataset and a target clean water pressure dataset are respectively intercepted from the clean water flow rate dataset and the clean water pressure dataset; wherein, the intercept duration is the duration starting from the end moment of the specified duration and towards the start moment of clean water discharge; Use a convolutional network to extract features from the target clean water flow rate dataset and the target clean water pressure dataset, splice the extracted features, obtain the clean water fluctuation features, and input the clean water fluctuation features into a prediction model to obtain the stable level of clean water discharge.

[0040] In the embodiment of the present invention, the above embodiment directly inputs the clean water flow data set and the clean water pressure data set into the flow control model. Although it contains sufficient back-end data that can be used to reflect the real fluctuation characteristics of clean water, the front-end data with poor data quality will still have a certain interference effect, which is not conducive to the precise control of the valve. In response to this problem, this embodiment is set to determine a section of interception time from the specified time. The interception time is a period of time that is based on the end of the specified time and traces back to the start time of clean water discharge. With this setting, a section of recent water flow data that basically excludes the influence of the roughness of the inner wall of the pipeline and the complexity of the pipeline layout can be obtained. The fluctuation characteristics of these water flow data are more correlated with the fluctuation of the clean water discharge itself (that is, the amount of clean water discharged and the change in pressure). Among them, the interception time can be a fixed proportion of the specified time, such as 50%, 40%, etc.

[0041] The convolutional network is used to extract features from the target water purification flow data set and the target water purification pressure data set. For the target water purification flow data set, the convolutional network extracts features such as the flow change trend and the frequency of peak occurrence; for the target water purification pressure data set, the convolutional network extracts features such as the pressure fluctuation amplitude and the pressure change cycle. Then, the two sets of extracted features are spliced ​​to form a comprehensive water purification fluctuation feature.

[0042] The concatenated net water fluctuation characteristics are input into the pre-trained prediction model, which is the model built based on the LSTM, CNN and other algorithms mentioned above. The prediction model predicts and outputs the corresponding stability level of net water discharge. For example, the prediction model may divide the stability level into three levels: high, medium and low (or number the levels with numbers, the larger the number, the higher the corresponding stability level), and judge the stability level of the current water flow based on the input net water fluctuation characteristics.

[0043] In some embodiments, the second traffic control model is trained in the following manner: Collect and construct a training data set, each piece of training data in the training data set includes: the real-time flow and real-time water pressure of the original pipeline, the real-time flow of the bypass pipeline, the opening of the main pipeline electric valve and the opening of the bypass pipeline outlet electric valve, the preset water head, and label data; wherein the label data is the stability evaluation value of the preset water head; Each training data in the training data set is divided into a training set and a test set, and the training set is used to train the second traffic control model. After the training reaches a preset level, the test set is used to test the second traffic control model. If the test meets the standard, the training is terminated; if the test does not meet the standard, other training data in the training set are continued to be used for training until the test meets the standard.

[0044] In the embodiment of the present invention, the second flow rate regulation model needs to be pre-trained. When the preset water head is in a stable state, the real-time flow rate and real-time water pressure of the original pipeline are collected, the real-time flow rate of the bypass pipeline is also collected, as well as the opening degrees of the electric valve of the main pipeline and the electric valve at the outlet of the bypass pipeline, and the preset water head at this time is also included. The labeled data is the stability evaluation value of the preset water head. For example, it is the fluctuation range of the preset water head. The lower the fluctuation range, the higher the stability evaluation value. This evaluation value can be manually evaluated or obtained by an automatic evaluator based on the fluctuation of the actual water head.

[0045] The training objective of the second flow rate regulation model is to determine the appropriate opening degree of the regulating valve under a specific purified water flow state (i.e., the real-time flow rate and real-time water pressure of the aforementioned original pipeline, and the real-time flow rate of the bypass pipeline) and the preset water head to maintain the stability of the preset water head (i.e., the fluctuation range is less than the preset range).

[0046] To ensure the effectiveness and generalization ability of model training, the constructed training data set needs to be reasonably divided. A part of the data is used as the training set for training the second flow rate regulation model. The proportion of the data volume in the training set is, for example, 70%-80%. Through these data, the model continuously learns the rules and features in the data and adjusts its own parameters to achieve the initial ability of flow rate regulation. Another part of the data is used as the test set for evaluating the performance of the trained model.

[0047] The second flow rate regulation model is trained using the training set. When the model training reaches a preset level, such as reaching a certain number of training rounds or the error is reduced to a certain threshold, the test set is used to test the model. After the data in the test set is input into the model, the model also outputs the regulation strategy and compares it with the stability evaluation value of the corresponding preset water head in the test set. If the test result of the model meets the standard, that is, the regulation strategy output by the model can make the error between the actual water head and the preset water head within an acceptable range, it indicates that the model has good performance and generalization ability, and the training process ends. On the contrary, if the test does not meet the standard, it means that the model still has deficiencies and needs to continue training with the data in the training set that has not participated in training to further optimize the model parameters and improve the model performance until the test meets the standard. Through such a repeated training and testing process, it is ensured that the second flow rate regulation model can accurately and stably achieve precise regulation of the flow rate according to various real-time water flow data to meet the demand of the purified water residual pressure power generation system for a stable water head. In some embodiments, the training data used for training the prediction model is obtained based on multiple sets of historical discharge data of the original purified water discharge pipeline of the sewage treatment plant, and the historical discharge data includes purified water flow rate data and purified water pressure data corresponding to the specified duration and a certain period after the specified duration.

[0048] In the embodiments of the present invention, the prediction model is based on the historical data of the purified water discharge of the sewage treatment plant for prediction, that is, a large number of sets of historical discharge data of the original purified water discharge pipeline of the sewage treatment plant are collected. The historical discharge data includes water flow data from the start of purified water discharge to near the end of discharge, that is, multiple sets of purified water flow rate data and purified water pressure data sampled according to a preset duration.

[0049] Based on these purified water flow rate data and purified water pressure data, the training data required for training the prediction model is constructed. The trained prediction model can predict the stable level of purified water discharge in the next period (such as 3 minutes, 5 minutes) based on the purified water fluctuation characteristics in the early stage of purified water discharge, so as to select an appropriate first flow rate regulation model or second flow rate regulation model. The prediction of the stable level of purified water discharge by the prediction model can be gradually started, so that the flow rate regulation model to be selected in the next period can be gradually determined.

[0050] As Figure 3 shown, the embodiments of the present invention also disclose a flow control system for residual pressure power generation in a water plant. The system includes a purified water discharge stable level prediction unit, a flow rate regulation model screening unit, and a flow rate regulation unit; The purified water discharge stable level prediction unit obtains a purified water flow rate data set and a purified water pressure data set of the original purified water discharge pipeline of the sewage treatment plant within a specified duration, and predicts the purified water discharge stable level based on the purified water flow rate data set and the purified water pressure data set; The flow rate regulation model screening unit matches a flow rate regulation model according to the purified water discharge stable level. Specifically, if the purified water discharge stable level is higher than a preset level, the first flow rate regulation model is selected as the flow rate regulation model, otherwise the second flow rate regulation model is selected as the flow rate regulation model; wherein, the regulation accuracy of the first flow rate regulation model is lower than that of the second flow rate regulation model; Receives the first real-time flow rate of the main pipeline flow meter and the real-time water pressure data of the pressure gauge, as well as the second real-time flow rate of the bypass pipeline flow meter, inputs the first real-time flow rate, the real-time water pressure data, the second real-time flow rate and a preset water head to the flow rate regulation model, and the flow rate regulation model outputs the first opening degree of the main pipeline electric valve and the second opening degree of the bypass pipeline outlet electric valve, and executes the first opening degree and the second opening degree.

[0051] In some embodiments, the stable level prediction unit for purified water discharge is configured to: Obtain an intercepted duration based on the specified duration match, and respectively intercept a target purified water flow data set and a target purified water pressure data set from the purified water flow data set and the purified water pressure data set according to the intercepted duration; wherein, the intercepted duration is a duration starting from the end moment of the specified duration and towards the starting moment of the purified water discharge. Use a convolutional network to extract features from the target purified water flow data set and the target purified water pressure data set, splice the extracted features, obtain a purified water fluctuation feature, and input the purified water fluctuation feature into a prediction model to obtain the stable level of the purified water discharge.

[0052] An embodiment of the present invention also discloses an electronic device, which includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, and when the computer program is executed, it implements the method described in any of the previous items.

[0053] An embodiment of the present invention also discloses a computer storage medium, which stores a computer program, and when the computer program is executed, it implements the method described in any of the previous items.

[0054] An embodiment of the present invention also discloses a computer program product, which has computer program code, and when the computer program code is executed, it implements the method described in any of the previous items.

[0055] It should be noted that in the description of the present invention, terms such as "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of the present invention, unless otherwise specified, the meaning of "a plurality" is two or more. Any process or method description shown in the flowchart or described in other ways herein can be understood as representing a module, segment, or part of executable instructions including one or more steps for implementing a specific logical function or process, and the scope of the preferred embodiments of the present invention includes additional implementations, where the functions can be executed in a substantially simultaneous manner or in the reverse order according to the involved functions, rather than in the order shown or discussed, which should be understood by those skilled in the technical field to which the embodiments of the present invention belong.

[0056] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), and the like.

[0057] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A flow control method for water plant residual pressure power generation, characterized in that: The method comprises the following steps: Obtaining a clean water flow data set and a clean water pressure data set of an original clean water discharge pipeline of a sewage treatment plant within a specified time period, and predicting a clean water discharge stability level based on the clean water flow data set and the clean water pressure data set; A flow control model is obtained according to the matching of the clean water discharge stability level. Specifically, if the clean water discharge stability level is higher than a preset level, a first flow control model is selected as the flow control model, otherwise a second flow control model is selected as the flow control model; wherein the control accuracy of the first flow control model is lower than that of the second flow control model; Receive the first real-time flow rate of the main pipeline flowmeter and the real-time water pressure data of the pressure gauge, as well as the second real-time flow rate of the bypass pipeline flowmeter, and input the first real-time flow rate, the real-time water pressure data, the second real-time flow rate and the preset water head into the flow control model. The flow control model outputs the first opening of the main pipeline electric valve and the second opening of the bypass pipeline outlet electric valve, and executes the first opening and the second opening.

2. A flow control method for water plant residual pressure power generation according to claim 1, characterized in that: The specified duration is determined in the following manner: Obtaining the roughness of the inner wall of the original pipeline and the layout information of the original pipeline, and evaluating the complexity of the pipeline layout according to the layout information; The roughness of the inner wall of the pipeline and the complexity of the pipeline layout are matched with the matching relationship at the same time to obtain the specified time length; wherein, in the matching relationship, the roughness of the inner wall of the pipeline and the complexity of the pipeline layout are both positively correlated with the specified time length.

3. A flow control method for water plant residual pressure power generation according to claim 2, characterized in that: The predicting of the clean water discharge stability level based on the clean water flow data set and the clean water pressure data set includes: Based on the specified time length matching, an interception time length is obtained, and a target clean water flow data set and a target clean water pressure data set are respectively intercepted from the clean water flow data set and the clean water pressure data set according to the interception time length; wherein the interception time length is the time length starting from the end moment of the specified time length and heading toward the start moment of clean water discharge; A convolutional network is used to extract features from the target clean water flow data set and the target clean water pressure data set, and the extracted features are spliced ​​to obtain clean water fluctuation features. The clean water fluctuation features are input into a prediction model to obtain the clean water discharge stability level.

4. A flow control method for water plant residual pressure power generation according to claim 1, characterized in that: The second traffic control model is trained in the following manner: Collect and construct a training data set, each piece of training data in the training data set includes: the real-time flow and real-time water pressure of the original pipeline, the real-time flow of the bypass pipeline, the opening of the main pipeline electric valve and the opening of the bypass pipeline outlet electric valve, and label data; wherein the label data is a stability evaluation value of a preset water head; Each training data in the training data set is divided into a training set and a test set, and the training set is used to train the second traffic control model. After the training reaches a preset level, the test set is used to test the second traffic control model. If the test meets the standard, the training is terminated; if the test does not meet the standard, other training data in the training set are continued to be used for training until the test meets the standard.

5. The flow control method for water plant residual pressure power generation according to claim 1, characterized in that: The training data used for the prediction model training is obtained based on multiple groups of historical discharge data of the original clean water discharge pipeline of the sewage treatment plant, and the historical discharge data includes clean water flow data and clean water pressure data corresponding to the specified time period and a certain period after the specified time period.

6. A flow control system for water plant residual pressure power generation, characterized in that: The system includes a clean water discharge stability level prediction unit, a flow control model screening unit, and a flow control unit; The clean water discharge stability level prediction unit obtains a clean water flow data set and a clean water pressure data set of the original clean water discharge pipeline of the sewage treatment plant within a specified time period, and predicts the clean water discharge stability level based on the clean water flow data set and the clean water pressure data set; The flow control model screening unit obtains a flow control model according to the clean water discharge stability level matching. Specifically, if the clean water discharge stability level is higher than a preset level, the first flow control model is selected as the flow control model, otherwise the second flow control model is selected as the flow control model; wherein the control accuracy of the first flow control model is lower than that of the second flow control model; The flow control unit receives the first real-time flow of the main pipeline flow meter and the real-time water pressure data of the pressure gauge, as well as the second real-time flow of the bypass pipeline flow meter, and inputs the first real-time flow, the real-time water pressure data, the second real-time flow and the preset water head into the flow control model. The flow control model outputs the first opening of the main pipeline electric valve and the second opening of the bypass pipeline outlet electric valve, and executes the first opening and the second opening.

7. A flow control system for water plant residual pressure power generation according to claim 6, characterized in that: The clean water discharge stability level prediction unit is used to: Based on the specified time length matching, an interception time length is obtained, and a target clean water flow data set and a target clean water pressure data set are respectively intercepted from the clean water flow data set and the clean water pressure data set according to the interception time length; wherein the interception time length is the time length starting from the end moment of the specified time length and heading toward the start moment of clean water discharge; A convolutional network is used to extract features from the target clean water flow data set and the target clean water pressure data set, and the extracted features are spliced ​​to obtain clean water fluctuation features. The clean water fluctuation features are input into a prediction model to obtain the clean water discharge stability level.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the computer program implements the method according to any one of claims 1 to 5 when executed.

9. A computer storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed, the method according to any one of claims 1 to 5 is implemented.

10. A computer program product, characterized in that: The computer program product contains computer program codes, which implement the method according to any one of claims 1 to 5 when executed.

Images