Water supply dispatching system and water transfer method

Abstract

The invention relates to a water supply dispatching system and a water transfer method, and belongs to the technical field of urban water supply. The water supply dispatching system comprises a water supply pipe network, a dispatching module and a control module. The water supply network is provided with a plurality of water supply nodes, each water supply node is provided with a pressure sensor, and the pressure sensors are used for detecting the water supply pressure of the corresponding water supply nodes; a plurality of water transfer schemes are stored in the scheduling module; the control module is configured to obtain detection data of the pressure sensor corresponding to the current water supply node so as to obtain water supply pressure data of the current water supply node; constructing a water pressure prediction model according to the water supply pressure data; calculating a pressure prediction value of the current water supply node according to the water pressure prediction model; if the pressure predicted value is smaller than a first threshold value, selecting a corresponding water transfer scheme from a scheduling module according to the pressure predicted value; and calculating the water supply pressure of the current water supply node after water transfer of the selected water transfer scheme, and determining a final water transfer scheme from the selected water transfer schemes according to a calculation result.

Description

Technical Field

[0001] The present invention belongs to the technical field of urban water supply, and in particular relates to a water supply scheduling system and a water regulation method. Background Art

[0002] The urban water supply system is an important part of urban infrastructure, and is responsible for providing clean water for various users such as residents, businesses, and industries. The water supply system uses dispatching instructions to evenly allocate water resources from water extraction and production to the municipal water supply network to continuously meet the water market demand and realize the water supply dispatching process. With the acceleration of urbanization and the continuous growth of population, the water supply dispatching of urban water supply systems faces increasing pressure and challenges. Water supply dispatching needs to find a balance between responding to emergencies and ensuring daily water supply. In a multi-stage pumping station system, water supply dispatching faces more complex challenges. The existence of multi-stage pumping stations increases the complexity of the system, and optimized dispatching is required at multiple levels to ensure stable and efficient water supply. The multi-stage pumping station water supply system needs to achieve efficient system integration and coordinated dispatching to ensure the coordinated operation of each water supply node.

[0003] Traditional water supply scheduling methods mainly rely on manual experience and simple rule systems, which have many shortcomings when facing complex and changeable water supply demands and multiple water source conditions. Traditional methods are difficult to obtain and process a large amount of water supply-related data in real time, and cannot achieve comprehensive monitoring and refined management of the water supply system. In addition, due to the lack of advanced algorithms and uneven technology or waste of resources. The water supply system involves multiple links and departments, and traditional methods are difficult to achieve cross-departmental and cross-system information sharing and collaborative scheduling.

[0004] Therefore, it is of great significance to design a system and method that can effectively solve the problem of urban water supply scheduling with multi-stage pumping stations. Summary of the invention

[0005] In view of the deficiencies existing in the related art, the present invention provides a water supply scheduling system and a water adjustment method to predict whether a water supply node is short of water and adjust water to the water supply node short of water.

[0006] The present invention provides a water supply scheduling system, comprising:

[0007] The water supply network is used to transport water; the water supply network is provided with a number of water supply nodes, each of which is provided with a pressure sensor, and the pressure sensor is used to detect the water supply pressure of the corresponding water supply node;

[0008] A scheduling module, which stores several water diversion schemes;

[0009] The control module is configured to: obtain detection data of a pressure sensor corresponding to a current water supply node to obtain water supply pressure data of the current water supply node; and construct a water pressure prediction model according to the water supply pressure data;

[0010] Calculate the pressure prediction value of the current water supply node according to the water pressure prediction model;

[0011] If the pressure prediction value is less than the first threshold, a corresponding water diversion scheme is selected from the scheduling module according to the pressure prediction value;

[0012] Calculate the water supply pressure of the current water supply node after water is diverted by the selected water diversion scheme, and determine the final water diversion scheme from the selected water diversion schemes based on the calculation results.

[0013] The present technical solution establishes a water pressure prediction model through the historical water supply pressure data of the water supply node, uses the water pressure prediction model to predict the water supply pressure of the water supply node at a certain moment, and compares the predicted value of the water supply pressure with the first threshold value of the water supply node to determine whether the water supply pressure of the water supply node needs to be adjusted. If adjustment is required, a water diversion scheme that meets the requirements is selected from the scheduling module, and the water diversion schemes that meet the requirements are compared, and the optimal water diversion scheme is selected from the water diversion schemes that meet the requirements.

[0014] In some of the embodiments, the water supply node includes several levels of water supply nodes, the current water supply node can supply water to the lower-level water supply node, and the upper-level water supply node can supply water to the current water supply node; when the current water supply node is short of water, the water supply pressure of the upper-level water supply node is adjusted to increase the water supply amount from the upper-level water supply node to the current water supply node.

[0015] In some of the embodiments, the water supply node that increases the water supply to the current water supply node is the target node; the target node is the superior of the current water supply node;

[0016] The control module is further configured to: calculate the water supply pressure of all water supply nodes at the same level of the current water supply node after the target node supplies water to the current water supply node, and form the calculation results into pressure regulation data; based on the KNN algorithm, calculate the Manhattan distance between the pressure regulation data and the second threshold in each water diversion scheme, and determine the final water diversion scheme from the selected water diversion schemes according to the Manhattan distance.

[0017] In some of the embodiments, the scheduling module includes a water supply pressure scheduling model, which is set in one-to-one correspondence with the water diversion scheme. The water supply pressure scheduling model is used to calculate the water supply pressure of all water supply nodes at the same level of the current water supply node after water diversion according to the corresponding water diversion scheme.

[0018] Some of the embodiments further include a storage module, which is used to store detection data of the pressure sensor; the control module is further configured to: obtain historical detection data of the pressure sensor from the storage module, and calculate the first threshold value.

[0019] In some embodiments, the first threshold value includes a lower limit value V l and upper limit value V u , lower limit V l and upper limit value V u Together they define the first range, with a lower limit value V l and upper limit value V u Corresponding to the lower limit and upper limit of the first range;

[0020] The control module is further configured to: when the pressure prediction value of the current water supply node is greater than the upper limit value V u When the water volume of the current water supply node is sufficient, the water volume to the next level water supply node can be increased; when the pressure prediction value of the current water supply node is less than the lower limit value V l When the predicted pressure value of the current water supply node is within the first range, it is determined that the water volume of the current water supply node is insufficient and water needs to be transferred from the upper-level water supply node. When the predicted pressure value of the current water supply node is within the first range, it is determined that the water volume of the current water supply node is normal and no water transfer is needed.

[0021] In some embodiments, different water supply nodes correspond to different second thresholds, and the second threshold is the lower limit value V of the corresponding water supply node. l and upper limit value V u The average value of .

[0022] In some of the embodiments, the water pressure prediction model is constructed based on the network structure of TimeGPT, the input of the water pressure prediction model is the historical water supply pressure value of the water supply node, and the output of the water pressure prediction model is the pressure prediction value of the water supply node at a certain moment in the future.

[0023] In addition, the present invention also provides a water regulation method, comprising the above-mentioned water supply scheduling system, and the water regulation method comprises the following steps:

[0024] Deploy pressure sensors at each water supply node of the water supply network, acquire and store detection data of the pressure sensors, and obtain historical detection data of each water supply node;

[0025] According to the historical detection data of each water supply node, a respective water pressure prediction model is constructed;

[0026] Calculate the predicted pressure of the corresponding water supply node according to the water pressure prediction model;

[0027] If the predicted pressure is less than the first threshold, it is determined that the target water supply node has insufficient water, and a water diversion scheme that meets the requirements is selected from the scheduling module;

[0028] Calculate the pressure of all water supply nodes at the same level of the current water supply node after the current water supply node transfers water, and determine the final water transfer plan from multiple water transfer plans based on the calculation results.

[0029] The present technical solution establishes a water pressure prediction model through the historical water supply pressure data of the water supply node, uses the water pressure prediction model to predict the water supply pressure of the water supply node at a certain moment, and compares the predicted value of the water supply pressure with the first threshold value of the water supply node to determine whether the water supply pressure of the water supply node needs to be adjusted. If adjustment is required, a water diversion scheme that meets the requirements is selected from the scheduling module, and the water diversion schemes that meet the requirements are compared, and the optimal water diversion scheme is selected from the water diversion schemes that meet the requirements.

[0030] In some of the embodiments, when determining the final water diversion plan, based on the KNN algorithm, the Manhattan distance between the pressure of all water supply nodes at the same level of the current water supply node after the current water supply node diverts water and the second threshold is calculated, and the final water diversion plan is determined from the selected water diversion plans based on the Manhattan distance.

[0031] Based on the above technical scheme, the water supply scheduling system in the embodiment of the present invention can predict the pressure prediction value of the water supply node, and compare the pressure prediction value with the first threshold value to determine whether the water supply pressure of the water supply node needs to be adjusted. When the water supply node is short of water, a water diversion scheme that meets the requirements is selected, and the water diversion schemes that meet the requirements are compared, and the optimal water diversion scheme is selected from the water diversion schemes that meet the requirements; not only can the water supply pressure regulation of the water supply node be realized to ensure the water diversion effect, but also the water diversion efficiency is high. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0033] Figure 1 It is a structural schematic diagram of water supply nodes at different levels in one embodiment of the water supply scheduling system of the present invention;

[0034] Figure 2 It is a structural schematic diagram of a multi-level water supply node in one embodiment of a water supply scheduling system of the present invention;

[0035] Figure 3 A flow chart of judging whether a water supply node needs water adjustment in one embodiment of a water supply scheduling system of the present invention;

[0036] Figure 4 The following is a flowchart of a control module in one embodiment of a water supply scheduling system of the present invention. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments. 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 making creative work are within the scope of protection of the present invention.

[0038] In the description of the present invention, it should be understood that the terms "center", "lateral", "longitudinal", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0039] The terms "first", "second", and "third" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Thus, a feature defined as "first", "second", and "third" may explicitly or implicitly include one or more of the features.

[0040] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0041] In an illustrative embodiment of the water supply scheduling system of the present invention, the water supply scheduling system is used to optimize the water volume in the water supply network, adjust the water volume distribution in the water supply network, ensure the efficiency of water supply, the water pressure stability at the user end and the safety of the network.

[0042] The above-mentioned water supply scheduling system includes a water supply network, which is used to transport water; the water supply network includes a number of water supply nodes, which are the converging points or diversion points of water flows in the water supply network.

[0043] As attached Figure 1As shown, the water supply node includes several levels of water supply nodes, and the several levels of water supply nodes are arranged in sequence according to the direction and path of the water flow; wherein, the current water supply node can supply water to the lower-level water supply node, and the upper-level water supply node can supply water to the current water supply node; when the current water supply node is short of water, the water supply pressure of the upper-level water supply node is adjusted to increase the water supply amount from the upper-level water supply node to the current water supply node.

[0044] As attached Figure 2 As shown, water supply nodes are usually divided into primary water supply nodes, secondary water supply nodes, ... N-level water supply nodes. The primary water supply node is connected to the secondary water supply node, and the secondary water supply node is connected to the tertiary water supply node. The primary water supply node can supply water to the secondary water supply node, and the secondary water supply node can supply water to the tertiary water supply node, and so on; the water flows from the primary water supply node through the secondary water supply node, the tertiary water supply node...N-level water supply nodes in turn. It can be considered that the input end of the primary water supply node is the source of the water flow, and the output end of the N-level water supply node is the user end.

[0045] In practical applications, water supply nodes include water supply pumping stations and water supply plants, etc. When other conditions remain unchanged, the higher the water supply pressure of the water supply node, the greater the water supply volume of the water supply node. The water supply pressure can be used as an indicator to reflect the water supply volume of the water supply node.

[0046] Water supply pressure reflects the flow dynamics of water and directly affects whether water can be delivered to the user end. Therefore, the water supply pressure of the water supply node affects the efficiency of water supply, the water pressure stability at the user end, and the safety of the pipe network. Among them, the water supply pressure of the water supply node refers to the water pressure measured at a position or measurement point in the water supply pipeline.

[0047] Since there are several water supply nodes in the water supply network, and the water supply pressures of different water supply nodes are different, in the above-mentioned water supply scheduling system, each water supply node is provided with a pressure sensor, and the pressure sensor is used to detect the water pressure of the corresponding water supply node; the pressure sensors are also set to several numbers, and the pressure sensors are set one by one corresponding to the water supply nodes.

[0048] The water supply scheduling system includes a control module, which is connected to a pressure sensor. Figure 3 As shown, the control module is configured to: obtain the detection data of the pressure sensor to obtain the water supply pressure data of the water supply node; construct a water pressure prediction model according to the detection data of the pressure sensor; calculate the pressure prediction value of the water supply node according to the water pressure prediction model; determine whether the water supply node needs to be adjusted according to the pressure prediction value of the water supply node; if the pressure prediction value is less than a first threshold value, it is determined that the water pressure of the water supply node is insufficient and water adjustment is required; if the pressure prediction value is greater than or equal to the first threshold value, it is determined that the water pressure of the water supply node is sufficient and water adjustment is not required.

[0049] It should be noted that, since the water supply pressures of different water supply nodes are different, the water supply pressure data of different water supply nodes are different, and thus the water pressure prediction models corresponding to different water supply nodes are different.

[0050] That is, the control module constructs a corresponding water pressure prediction model according to the water supply pressure data of the water supply node, and uses the corresponding water pressure prediction model to predict the water supply pressure of the water supply node at a certain moment.

[0051] It should be noted that the water supply pressure data of a water supply node is usually the historical water supply pressure. The historical water supply pressure refers to the pressure data collected from the water supply node over a period of time in the past. Through these historical data, the law of pressure change of the water supply node can be obtained, thereby generating a water pressure prediction model.

[0052] The change in water supply pressure of a water supply node is related to time. The water supply pressure of a water supply node is different at different times. Therefore, the input data of the water pressure prediction model includes at least the water pressure value of the water supply node and the time point corresponding to the water pressure value, so that the water pressure prediction model can predict the water supply pressure of the water supply node at a certain moment and realize real-time prediction of the water supply node.

[0053] The water supply scheduling system includes a storage module, which is used to store data detected by the pressure sensor. The storage module is connected to a control module, which is used to obtain historical detection data of the pressure sensor stored in the storage module.

[0054] The detection data of the pressure sensor is raw data. Directly using the raw data for model training can easily lead to large model errors. The raw data needs to be preprocessed. The control module uses the processed data to generate a water pressure prediction model to ensure the prediction accuracy of the water pressure prediction model.

[0055] In data preprocessing, the original data must first be cleaned to remove noise, missing values, and outliers to ensure data quality. Data cleaning includes the use of moving average filtering, Kalman filtering, Fourier transform, and other methods to remove noise. For missing value processing, linear interpolation methods can be used. Outlier processing uses statistical methods (such as the 3σ principle) to identify outliers and processes them through smoothing methods (such as local weighted regression) or median replacement techniques. Through these steps, data quality and reliability can be ensured, thereby ensuring the accuracy of the water pressure prediction model, and then ensuring the prediction accuracy of the water pressure prediction model.

[0056] TimeGPT is a large-scale language model specifically designed for processing time series data. It combines advanced natural language processing (NLP) technology with time series analysis methods to efficiently capture and predict patterns in data changes over time.

[0057] In the above water supply scheduling system, a water pressure prediction model is constructed based on the network structure of TimeGPT. The input of the water pressure prediction model is the historical water supply pressure value of the water supply node, and the output of the water pressure prediction model is the pressure prediction value of the water supply node at a certain moment in the future.

[0058] It should be noted that, since the TimeGPT model requires time series data, the pressure data and time information are collated and constructed to form an input sequence for predicting future pressure value data.

[0059] After the raw data is preprocessed, the processed data is used to train the TimeGPT neural network water pressure prediction model. The TimeGPT model can capture the dynamic laws of pressure changes in the water supply network, thereby making accurate pressure predictions.

[0060] In the process of building the water pressure prediction model, a large amount of historical data is used to optimize the model so that it can learn the complex nonlinear relationships and time dependencies of the water supply network under good operating conditions.

[0061] In some embodiments, the water pressure prediction model can predict the pressure conditions in the next 15 minutes and output the predicted pressure value.

[0062] In this embodiment, the TimeGPT water pressure prediction model outputs the pressure value predicted by the TimeGPT model of the water supply node according to the set time step, that is, the expected pressure value of the pipe network at that moment.

[0063] TimeGPT can effectively process and predict time series data through the transformer model and self-attention mechanism. The TimeGPT network structure can efficiently process and predict the pressure data in the water supply network by introducing the transformer encoder and multi-head self-attention mechanism. This structure not only improves the accuracy of the prediction, but also has good scalability and application prospects. By predicting the real-time pressure data of the water supply node, it can effectively ensure the stable operation of the water supply system, improve the user's water experience, and provide strong support for the optimization management of the water supply system.

[0064] The network structure of TimeGPT mainly includes the following key parts:

[0065] Data preprocessing module: Data standardization, standardize the input time series data to ensure that data with different features are on the same scale. Time window segmentation, segment the time series data into time windows of fixed length to generate training samples.

[0066] Position Encoding Module: Uses sine and cosine functions to generate position encodings to embed the information of time steps into the input data, allowing the model to recognize the order of time steps.

[0067] Transformer Encoder Module: Multi-head self-attention mechanism: Capture long-range dependencies in time series data through parallel calculations of multiple attention heads. Each attention head calculates the correlation (attention weight) of each time step in the input sequence with other time steps, and obtains a new representation through weighted summation. Convolutional Neural Network: Each encoder layer contains a convolutional neural network to extract high-order features.

[0068] Output layer: After being processed by multiple transformer encoder layers, the final output layer predicts the time series data. The output result is the predicted pressure value for the future time step.

[0069] In the model training and optimization part, mean square error is used as the loss function to measure the gap between the predicted value and the true value. At the same time, the Adam optimization algorithm is used to update the model parameters through the gradient descent method to gradually reduce the prediction error. L2 regularization and Dropout technology are introduced to prevent model overfitting and improve the generalization ability of the model.

[0070] The network structure of TimeGPT is common knowledge in the art and will not be described in detail here.

[0071] In the above water supply scheduling system, the first threshold is obtained based on the pressure output by the water supply node at the same time in the past period of time, that is, the first threshold is obtained based on the historical data of the water supply pressure of the water supply node.

[0072] It should be noted that, since the historical data of water supply pressures of different water supply nodes are different, the first thresholds corresponding to different water supply nodes are also different.

[0073] The control module is configured to: obtain historical detection data of the pressure sensor from the storage module, and calculate a first threshold corresponding to the water supply node according to the historical water supply pressure data of the water supply node.

[0074] The pressure of each water supply node in the water supply network has its stable operating range, and the water supply pressure of the water supply node is variable. Therefore, the first threshold of the same water supply node has different values ​​at different times.

[0075] The pressure value distribution of the water supply pressure of the water supply node in a day satisfies the normal distribution, and the first threshold value includes the lower limit value V l and upper limit value V u , lower limit V l and upper limit value V u Together they define the first range, with a lower limit value V l and upper limit value V u The corresponding values ​​are the lower limit and the upper limit of the first range. It can be considered that the first threshold value varies within the first range.

[0076] By determining whether the pressure prediction value of the current water supply node is within the first range, it is determined whether the current water supply node needs water transfer. u When the predicted pressure is within the first range, it is determined that the water volume of the current water supply node is sufficient and water transfer is not required, but the water volume of the downstream water supply node cannot be increased; when the predicted pressure is less than the lower limit value V l When the water volume at the detection point is determined to be insufficient, water needs to be transferred from the upper-level water supply node.

[0077] It should be noted that the lower limit value V corresponding to different water supply nodes l and upper limit value V u are different.

[0078] Lower limit V l and upper limit value V u It can be determined based on the sigma principle, 2sigma principle and 3sigma principle in the normal distribution.

[0079] In the above water supply dispatching system, the 3sigma principle is adopted to complete the preliminary setting by combining historical water consumption statistics, dispatching experience and weather temperature correction. The first lower limit V l and upper limit value V u , δ is the standard deviation of historical water volume data, and the weather correction factor is a 气象 , recent total water output correction factor b 水量 For holidays and weekends, the peak time of users' water consumption is relatively delayed compared with that on weekdays, and the peak value of water consumption is relatively low. The correction factor c for holidays and weekends is the general mathematical form of the threshold calculation as shown in formula (1) to formula (2).

[0080] V n =(μ-3δ)×a 气象 ×b 水量 ×c 节假日 (1)

[0081] V u =(μ+3δ)×a 气象 ×b 水量 ×c 节假日 (2)

[0082] Where: V l →Lower limit value;

[0083] V u →Upper limit value;

[0084] μ——mean value of historical water volume data, m 3 ;

[0085] δ——standard deviation of historical water volume data;

[0086] a 气象 ——weather correction factor;

[0087] b 水量 ——Water volume correction factor;

[0088] c 节假日 ——Correction factors for holidays and weekends.

[0089] Meteorological factor a 气象 The calculation of is as shown in formula (3).

[0090]

[0091] Where: T a ——Average temperature of the day, ℃ ° Water demand factor b 水量 The calculation of is as shown in formula (4).

[0092] b 水量 =1 + 1 / 100( 1 + e -0. 15( w-25) ) (4)

[0093] Where: w is the predicted water demand for the day, m 3 / d °

[0094] Holiday factor c 节假日 The calculation of is as shown in formula (5).

[0095]

[0096] Where: t is the day of the holiday; L is the length of the holiday, d °

[0097] When the current water supply node is short of water, the water supply from one or several water supply nodes at the upper level to the current water supply node can be increased to solve the problem of water shortage in the current water supply node.

[0098] Since there are usually multiple water supply nodes at the previous level, there may be multiple different water diversion schemes when the current water supply node is short of water. The water diversion scheme at least includes information such as the water supply node number and the water supply pressure after pressure regulation of the water supply node.

[0099] Although each water diversion scheme can adjust the water volume of the current water supply node so that the water supply pressure of the current water supply node meets the requirements, the water diversion efficiency, water diversion cost, etc. of different water diversion schemes are different.

[0100] In the above water supply scheduling system, each water diversion scheme is calculated and compared to select the best water diversion scheme as the final water diversion scheme to ensure the overall water diversion efficiency and effect.

[0101] Specifically, the control module is configured to: calculate the water supply pressure of all water supply nodes at the same level of the current water supply node after the current water supply node diverts water, form the calculation results into pressure regulation data, calculate the Manhattan distance between the pressure regulation data and the second threshold in each water diversion scheme, and determine the final water diversion scheme from the selected water diversion schemes based on the Manhattan distance.

[0102] It should be noted that the calculation results of the water supply pressures of all water supply nodes at the same level of the current water supply node after the current water supply node adjusts the water to form the pressure regulation data is a common technical means in this field and will not be repeated here.

[0103] K-Nearest Neighbor (also known as KNN model) is a simple and intuitive machine learning algorithm, which is widely used in classification and regression tasks. The core idea is: given a data point, find the K most similar neighbors in the feature space and predict the output of the data point based on the values ​​of these neighbors. In regression problems, the KNN model makes predictions by calculating the average of the K nearest neighbors.

[0104] Manhattan distance is a way to calculate the distance between two points in multidimensional space. It is defined as:

[0105]

[0106] Among them, x i and i are the coordinates of two data points in the i-th dimension, and n is the number of dimensions.

[0107] Based on the KNN algorithm, the Manhattan distance between the pressure regulation data and the second threshold is calculated, and the final water diversion scheme is determined from the selected water diversion schemes according to the Manhattan distance.

[0108] It should be noted that different water supply nodes correspond to different second thresholds, and the second threshold is the lower limit value V of the corresponding water supply node. l and upper limit value V u The average value of .

[0109] In practical applications, water diversion plans are usually prepared in advance based on historical data or historical experience. When a water supply node is short of water, one or more water diversion plans that meet the requirements are selected from a known water diversion plan database.

[0110] The water supply scheduling system includes a scheduling module, which stores several water diversion schemes. When the control module determines that the water pressure at the detection point is insufficient, the control module is configured to select a water diversion scheme from the scheduling module.

[0111] The scheduling module includes a water supply pressure scheduling model, which is set in one-to-one correspondence with the water diversion scheme. The water supply pressure scheduling model is used to calculate the water supply pressure of all water supply nodes at the same level of the current water supply node after the water is diverted according to the corresponding water diversion scheme. The water supply pressure scheduling model can be used to calculate the water supply pressure of all water supply nodes at the same level of the current water supply node after the current water supply node diverts water.

[0112] In this embodiment, a water supply pressure scheduling model is obtained by training a neural network model. In the neural network model, the neural network model does not have a specific mathematical relationship expression, but obtains the mapping relationship between input and output through learning and training of historical data.

[0113] Specifically, the historical water supply pressures of all water supply nodes at the same level as the target node are used as input, and the historical water supply pressures of all water supply nodes at the same level as the current water supply node at the corresponding moment are used as output. Through continuous training and adjustment, a water supply pressure scheduling model is obtained; through the water supply pressure scheduling model, the water supply pressures of all water supply nodes at the same level as the current water supply node can be calculated after the current water supply node adjusts the water.

[0114] In this embodiment, for the convenience of description, the current water supply node is called a detection point, and the water supply node that supplies water to the current water supply node is called a target node, wherein the target node is located at the upper level of the detection point. By adjusting the water supply pressure of the target node, the amount of water provided by the target node to the detection point is increased.

[0115] In model training, the historical pressure value of the target node is first used as the input of the BP pressure fitting model, and the pressure data of the detection point and its water supply nodes at the same level at the corresponding time is used as the model output. Through continuous adjustment and training, a fitting model that can accurately reflect the pressure changes of the actual water supply network is formed, which can predict the pressure changes of the detection point under these pressure regulation conditions; then based on the water supply pressure scheduling model fitted by the BP (back propagation) neural network, the above collected data is used to train the water supply pressure scheduling model. Historical data can provide basic information for subsequent model training and prediction. The collected data cannot be used directly, and the quality and consistency of the data are ensured through preprocessing. During the preprocessing process, the pressure data of adjacent nodes at the same time point in the past are sorted. The time series and correlation of the data are ensured, providing more accurate input for the model. The historical pressure data of adjacent nodes are input into the scheduling model of the BP neural network, and the BP neural network is used to fit the dynamic data relationship between the pressure values ​​of each adjacent node.

[0116] That is, the BP neural network is used to fit the pressure effect produced by the pressurization scheme. The content of the BP neural network is as follows:

[0117] The general activation function is logsig, and its expression is:

[0118]

[0119] So in the hidden layer, the data from the previous layer is received and processed as follows

[0120] logsig(X i W ij +X i+1 W ij+1 +β i )

[0121] The expression of y in the table view above is:

[0122]

[0123] The pressure prediction value of the detection point is input into the water supply pressure scheduling model, and the water supply pressure scheduling model will output the water supply pressure of all water supply nodes at the same level as the detection point.

[0124] like Figure 4 As shown, the control principle of the above-mentioned water supply scheduling system is: the control module obtains the detection data of the pressure sensor of each water supply node, and stores the data in the storage module to obtain the water supply pressure history data of each water supply node; the water pressure prediction model of each water supply node is constructed according to the water supply pressure history data of each water supply node; the pressure prediction value of the current water supply node at a certain moment in the future is calculated according to the water pressure prediction model of the current water supply node; if the pressure prediction value is within the first range, the current water supply node does not need to transfer water; if the pressure prediction value is greater than the upper limit value of the first range, the current water supply node does not need to transfer water, and the water supply to the lower-level water supply node can be increased; if the pressure prediction value is less than the lower limit value of the first range, the current water supply node needs to transfer water, and a water transfer scheme that meets the requirements is selected from the scheduling module according to the pressure prediction value, and the pressure regulation data in each water transfer scheme is calculated through the water supply pressure scheduling model, and the Manhattan distance between the pressure regulation data and the second threshold is calculated, so as to select the final water transfer scheme from the water transfer scheme that meets the requirements.

[0125] The above-mentioned water supply scheduling system establishes a water pressure prediction model through the historical water supply pressure data of the water supply node, uses the water pressure prediction model to predict the water supply pressure of the water supply node at a certain moment, and compares the predicted value of the water supply pressure with the first threshold of the water supply node to determine whether the water supply pressure of the water supply node needs to be adjusted. If adjustment is required, a water diversion scheme that meets the requirements is selected from the scheduling module, and the water diversion schemes that meet the requirements are compared, and the optimal water diversion scheme is selected from the water diversion schemes that meet the requirements.

[0126] Based on the above water supply scheduling system, the present invention also provides a water adjustment method, which includes the above water supply scheduling system. The water adjustment method includes the following steps:

[0127] Deploy pressure sensors at each water supply node of the water supply network, acquire and store detection data of the pressure sensors, and obtain historical detection data of each water supply node;

[0128] According to the historical detection data of each water supply node, a respective water pressure prediction model is constructed;

[0129] Calculate the predicted pressure of the corresponding water supply node according to the water pressure prediction model;

[0130] If the predicted pressure is less than the first threshold, it is determined that the target water supply node has insufficient water, and multiple water transfer schemes are selected from the scheduling module;

[0131] The water supply pressure of all water supply nodes at the same level of the current water supply node after the current water supply node transfers water is calculated, and a final water transfer scheme is determined from multiple water transfer schemes based on the calculation results.

[0132] When determining the final water diversion plan, the calculation results of the water supply pressures of all water supply nodes at the same level of the current water supply node after the current water supply node diverts water are used to form pressure regulation data; based on the KNN algorithm, the Manhattan distance between the pressure regulation data and the second threshold is calculated, and the final water diversion plan is determined from the selected water diversion plans according to the Manhattan distance.

[0133] It should be noted that in actual applications, the control module does not only determine whether a certain water supply node needs water diversion, but makes a judgment on all water supply nodes to achieve the overall reliable operation of the water supply network.

[0134] The following takes the water supply node A22 as an example of the detection point to introduce the water regulation method of the above water supply scheduling system in detail.

[0135] The pressure sensor that detects the pressure value of the water supply node A22 stores its detection data in the storage module;

[0136] The control module generates a water pressure prediction model based on the historical detection data stored in the storage module, and uses the water pressure prediction model to predict the water supply pressure of the water supply node A22 at a certain moment to obtain the pressure prediction value of the water supply node A22 at a certain moment.

[0137] When the pressure prediction value is within the first range, the water volume of the water supply node A22 meets the requirements of the water supply network and no water adjustment is required.

[0138] When the predicted pressure value is greater than the upper limit of the first range, the water volume of the water supply node A22 is sufficient, which not only meets the requirements of the water supply network, but also can supply water to the lower-level water supply nodes when the lower-level water supply nodes are short of water.

[0139] When the pressure prediction value is less than the lower limit of the first range, the water volume of the water supply node A22 is insufficient to meet the requirements of the water supply network, and water needs to be transferred from the upper-level water supply node.

[0140] Among them, water can be directly transferred from water supply node A2, water can also be transferred from water supply node A1, water can also be transferred from water supply node A3, water can also be transferred from water supply node A2 and water supply node A1, etc. There can be many water transfer schemes that meet the requirements. These water transfer schemes are pre-stored in the scheduling module, and the control module selects a water transfer scheme that meets the requirements from the scheduling module according to the pressure prediction value.

[0141] Assume that there are three solutions that meet the requirements. The first solution is to transfer water from water supply node A2, the second solution is to transfer water from water supply node A2 and water supply node A1, and the third solution is to transfer water from water supply node A3.

[0142] Calculate the pressure values ​​of water supply nodes A11, A12, A13, A21, A22, A23, A31, A32 and A33 after water is transferred by water supply node A22 in the first solution, form pressure adjustment data with the calculation results, and calculate the Manhattan distance between the pressure adjustment data and the second threshold value;

[0143] Calculate the pressure values ​​of water supply nodes A11, A12, A13, A21, A22, A23, A31, A32 and A33 after water is transferred by water supply node A22 in the second scheme, form pressure adjustment data with the calculation results, and calculate the Manhattan distance between the pressure adjustment data and the second threshold value;

[0144] Calculate the pressure values ​​of water supply nodes A11, A12, A13, A21, A22, A23, A31, A32 and A33 after water is transferred by water supply node A22 in the third solution, form pressure adjustment data with the calculation results, and calculate the Manhattan distance between the pressure adjustment data and the second threshold value;

[0145] According to the Manhattan distance calculated in each water diversion scheme, the optimal water diversion scheme is selected from the three water diversion schemes.

[0146] Finally, it should be noted that: the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0147] The above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or some technical features can be replaced by equivalents without departing from the spirit of the technical solution of the present invention, which should be included in the scope of the technical solution for protection of the present invention.

Claims

1. A water supply scheduling system, characterized in that: include: A water supply network for conveying water; the water supply network is provided with a plurality of water supply nodes, each of which is provided with a pressure sensor, and the pressure sensor is used to detect the water supply pressure corresponding to the water supply node; A scheduling module, which stores several water diversion schemes; A control module, which is configured to: obtain detection data of the pressure sensor corresponding to the current water supply node to obtain water supply pressure data of the current water supply node; Construct a water pressure prediction model based on water supply pressure data; Calculate the predicted pressure value of the current water supply node according to the water pressure prediction model; If the pressure prediction value is less than a first threshold, selecting a corresponding water diversion scheme from the scheduling module according to the pressure prediction value; The water supply pressure of the current water supply node after water diversion according to the selected water diversion scheme is calculated, and a final water diversion scheme is determined from the selected water diversion schemes according to the calculation results.

2. The water supply scheduling system according to claim 1, characterized in that: The water supply node includes several levels of water supply nodes. The current water supply node can supply water to the lower-level water supply node, and the upper-level water supply node can supply water to the current water supply node. When the current water supply node is short of water, the water supply pressure of the upper-level water supply node is adjusted to increase the water supply volume from the upper-level water supply node to the current water supply node.

3. The water supply scheduling system according to claim 2, characterized in that: The water supply node that increases the water supply to the current water supply node is the target node; the target node is the superior of the current water supply node; The control module is further configured to: calculate the water supply pressure of all the water supply nodes at the same level of the current water supply node after the target node supplies water to the current water supply node, and form the calculation results into pressure regulation data; based on the KNN algorithm, calculate the Manhattan distance between the pressure regulation data and the second threshold in each water diversion scheme, and determine the final water diversion scheme from the selected water diversion schemes according to the Manhattan distance.

4. The water supply scheduling system according to claim 3, characterized in that: The scheduling module includes a water supply pressure scheduling model, which is set in one-to-one correspondence with the water diversion scheme. The water supply pressure scheduling model is used to calculate the water supply pressure of all water supply nodes at the same level of the current water supply node after water diversion corresponding to the water diversion scheme.

5. The water supply scheduling system according to claim 1, characterized in that: It also includes a storage module, which is used to store the detection data of the pressure sensor; the control module is further configured to: obtain the historical detection data of the pressure sensor from the storage module, and calculate the first threshold.

6. The water supply scheduling system according to claim 1 or 3, characterized in that: The first threshold value includes a lower limit value V l and upper limit value V u , the lower limit value V l and the upper limit value V u Together they define a first range, the lower limit value V l and the upper limit value V u corresponding to the lower limit and upper limit of the first range; The control module is further configured to: when the current predicted pressure value of the water supply node is greater than the upper limit value V u When the water volume of the current water supply node is sufficient, the water volume to the water supply node at the next level can be increased; when the pressure prediction value of the current water supply node is less than the lower limit value V l When the predicted pressure value of the current water supply node is within the first range, it is determined that the water volume of the current water supply node is insufficient and water needs to be transferred from the upper water supply node; when the predicted pressure value of the current water supply node is within the first range, it is determined that the water volume of the current water supply node is normal and water transfer is not needed.

7. The water supply scheduling system according to claim 6, characterized in that: The second threshold value corresponding to different water supply nodes is different, and the second threshold value is the lower limit value V corresponding to the water supply node. l and the upper limit value V u The average value of .

8. The water supply scheduling system according to claim 1, characterized in that: The water pressure prediction model is constructed based on the network structure of TimeGPT. The input of the water pressure prediction model is the historical water supply pressure value of the water supply node, and the output of the water pressure prediction model is the pressure prediction value of the water supply node at a certain moment in the future.

9. A water regulation method for a water supply scheduling system according to claim 1, characterized in that: The water transfer method comprises the following steps: Deploy pressure sensors at each water supply node of the water supply network, acquire and store detection data of the pressure sensors, and obtain historical detection data of each water supply node; Constructing respective water pressure prediction models according to the historical detection data of each water supply node; Calculating the predicted pressure corresponding to the water supply node according to the water pressure prediction model; If the predicted pressure is less than the first threshold, it is determined that the water volume of the target water supply node is insufficient, and a water diversion scheme that meets the requirements is selected from the scheduling module; The pressure of all water supply nodes at the same level of the current water supply node after the current water supply node transfers water is calculated, and a final water transfer scheme is determined from multiple water transfer schemes based on the calculation results.

10. The water supply scheduling method according to claim 9, characterized in that: When determining the final water diversion plan, based on the KNN algorithm, the Manhattan distance between the pressure of all water supply nodes at the same level of the current water supply node after the current water supply node diverts water and the second threshold is calculated, and the final water diversion plan is determined from the selected water diversion plans according to the Manhattan distance.

Images