Sewage pipe network early warning method and system based on distributed sensor and AI prediction

Abstract

The invention discloses a sewage pipe network early warning method and system based on a distributed sensor and AI prediction, and the method comprises the steps: 1, setting a plurality of measurement points in a to-be-detected sewage pipeline, and collecting the rainfall and the flow information of the plurality of measurement points; 2, a sewage pipeline end point total flow prediction model is constructed according to the flow information of the multiple measurement points; 3, optimizing the sewage pipeline end point total flow prediction model according to a preset neural network model; and 4, predicting the end point flow of the sewage pipeline according to the optimized end point total flow prediction model of the sewage pipeline to obtain a flow change trend. Flow prediction of the sewage pipeline end point is achieved, and the flow result can be conveniently predicted according to the dynamic rainfall condition.

Description

Technical Field

[0001] This invention relates to the field of smart water technology, and in particular to a method and system for early warning of sewage pipe networks based on distributed sensors and AI prediction. Background Technology

[0002] In urban drainage system management, accurate monitoring and prediction of sewage pipeline flow is crucial for flood control, drainage, sewage treatment plant scheduling, and the safe operation of the pipeline network. Traditional sewage pipeline level monitoring technology mainly relies on single-point sensors to collect data in real time and trigger alarm mechanisms through fixed thresholds. However, this type of technology has the following significant drawbacks in practical applications:

[0003] 1. Traditional liquid level monitoring only displays instantaneous values ​​and lacks trend prediction capabilities.

[0004] Existing monitoring systems typically only provide real-time displays of current liquid levels or flow rates, and cannot effectively predict future flow trends based on historical data. This makes it difficult for managers to anticipate high water level risks in advance, especially under extreme weather conditions such as heavy rain, which can easily lead to urban flooding or sewage overflows due to delayed response.

[0005] 2. The fixed single-point alarm threshold cannot adapt to dynamic rainfall conditions.

[0006] Current alarm mechanisms mostly rely on sensors installed at fixed points for detection and issue warnings based on preset fixed thresholds (such as liquid level). They fail to consider factors such as changes in rainfall, upstream water flow velocity, and the dynamic load capacity of the pipeline network. During periods of continuous heavy rainfall, the fixed thresholds cannot be dynamically adjusted, often resulting in false alarms or missed alarms, thus reducing the reliability of the early warning system.

[0007] 3. The water intake scheduling of the plant is out of sync with the status of the pipeline network, resulting in a delayed response.

[0008] Wastewater treatment plant influent scheduling is typically based on experience or current sewage pipeline flow rates, lacking effective coordination with real-time pipeline network operation. When the pipeline level rises abnormally, the plant struggles to adjust its treatment capacity in a timely manner, causing sewage to stagnate within the network and exacerbating the risk of pipeline overload. Furthermore, current technologies fail to achieve coordinated scheduling between pipeline nodes, meaning that localized blockages or overflows cannot be resolved through global optimization. Summary of the Invention

[0009] To address the shortcomings of existing wastewater pipeline network early warning methods, such as the lack of trend prediction capabilities due to instantaneous level monitoring, fixed single-point alarm thresholds, inability to adapt to dynamic rainfall conditions, and the disconnect between wastewater treatment plant influent scheduling and pipeline status, resulting in delayed response, this invention provides a wastewater pipeline network early warning method and system based on distributed sensors and AI prediction. This method collects flow information at different locations within the wastewater pipeline by setting multiple measurement points. A total flow prediction model for the wastewater pipeline endpoint is then constructed and optimized using a pre-defined neural network model, resulting in an optimized total flow prediction model for the wastewater pipeline endpoint. The flow prediction at the endpoint of the wastewater pipeline is then implemented based on this optimized model. This invention considers multiple measurement points and the flow delay caused by rainfall when predicting the flow at the endpoint of the wastewater pipeline, facilitating flow prediction based on dynamic rainfall conditions. Furthermore, by predicting the wastewater pipeline flow, wastewater treatment plants can obtain real-time information about the pipeline and take proactive measures.

[0010] To achieve the above objectives, the technical solution of the present invention is as follows:

[0011] The first aspect of this invention proposes a wastewater pipe network early warning method based on distributed sensors and AI prediction, comprising:

[0012] Step 1: Set up multiple measurement points in the sewage pipe to be tested to collect rainfall and flow information at the multiple measurement points; the measurement points include the end point of the sewage pipe and multiple measurement points in the sewage pipe, so as to collect flow information at multiple locations;

[0013] Step 2: Construct a prediction model for the total flow rate at the end of the sewage pipeline based on the flow rate information at multiple measurement points, which facilitates the preliminary prediction of the total flow rate at the end of the sewage pipeline;

[0014] Step 3: Optimize the total flow prediction model at the end of the sewage pipeline based on the preset neural network model to improve the accuracy of the prediction results;

[0015] Step 4: Based on the optimized total flow prediction model at the end of the sewage pipeline, predict the flow rate at the end of the sewage pipeline to obtain the flow rate change trend.

[0016] Furthermore, the flow information includes instantaneous flow rate and cumulative flow rate change.

[0017] Furthermore, the total flow prediction model at the end of the sewage pipeline is expressed by the following formula:

[0018]

[0019] Where s(t) is the predicted total flow rate at the end of the sewage pipeline, and v iLet be the instantaneous flow rate at the i-th measurement point, Δt be the transmission time between adjacent measurement points, α be the rainfall-flow conversion coefficient, and r be the instantaneous flow rate. i Let be the rainfall, i be the i-th measurement point, and u be the... i Let t be the cumulative flow change every 10 minutes during the transmission time from the i-th measurement point to the endpoint, and t be the current time.

[0020] Furthermore, step three includes:

[0021] The historical flow information of all measurement points and their corresponding timestamps are input into a preset neural network model to obtain residual correction terms, which are used to compensate for the nonlinear effects not captured by the total flow prediction model at the end of the sewage pipeline.

[0022] The optimized prediction model for total flow at the end of the sewage pipeline is obtained by combining the residual correction term and the prediction model for total flow at the end of the sewage pipeline, which facilitates the improvement of prediction accuracy.

[0023] The optimized prediction model for the total flow rate at the end of the sewage pipeline is expressed by the following formula:

[0024] s pred (t)=s(t)+Δs(t)

[0025] Among them, s pred Δs(t) represents the optimized predicted total flow rate at the end of the sewage pipeline, and Δs(t) represents the residual correction term.

[0026] Furthermore, step three also includes optimizing the total flow prediction model at the end of the sewage pipeline by adjusting the rainfall-flow conversion coefficient and the transmission time between adjacent measurement points according to a preset neural network model. The above process specifically includes:

[0027] The historical flow information of all measurement points and their corresponding timestamps are input into a preset neural network model to obtain the corrected rainfall-flow conversion coefficient and the transmission time between adjacent measurement points.

[0028] The prediction model for the total flow at the end of the sewage pipeline is updated based on the corrected rainfall-flow conversion coefficient and the transmission time between adjacent measurement points, resulting in an optimized prediction model for the total flow at the end of the sewage pipeline, which facilitates improved prediction accuracy.

[0029] Furthermore, the preset neural network model includes a temporal feature extraction layer, a fully connected layer, and an activation function;

[0030] The temporal feature extraction layer is used to capture temporal dependencies, and the temporal feature extraction layer includes a long short-term memory network or a transformer.

[0031] The fully connected layer is used to map the output of the temporal feature extraction layer to residuals or parameter adjustment amounts;

[0032] The activation function is used to process the output of the fully connected layer to prevent gradient vanishing.

[0033] The second aspect of this invention proposes a wastewater pipe network early warning system based on distributed sensors and AI prediction, comprising:

[0034] A collection module is used to set up multiple measurement points in the sewage pipe to be inspected to collect rainfall and flow information at the multiple measurement points; the measurement points include the end point of the sewage pipe and multiple measurement points in the sewage pipe, which facilitates the collection of flow information at multiple locations;

[0035] The physical model building module is used to build a prediction model of the total flow at the end of the sewage pipeline based on the flow information at multiple measurement points, which facilitates the preliminary prediction of the total flow at the end of the sewage pipeline.

[0036] The optimization module is used to optimize the prediction model of the total flow at the end of the sewage pipeline based on the preset neural network model, so as to improve the accuracy of the prediction results.

[0037] The prediction module is used to predict the flow rate at the end of the sewage pipeline based on the optimized total flow rate prediction model at the end of the sewage pipeline, and to obtain the flow rate change trend.

[0038] A third aspect of the present invention provides an electronic device including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the wastewater network early warning method based on distributed sensors and AI prediction as described in the first aspect above.

[0039] A fourth aspect of the present invention provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device on which the storage medium is located to perform the wastewater network early warning method based on distributed sensors and AI prediction as described in the first aspect above.

[0040] The beneficial effects of this invention are:

[0041] This invention combines a sewage pipeline end-point total flow prediction model with a neural network model. This not only preserves the domain knowledge of the sewage pipeline end-point total flow prediction model (such as delayed superposition), but also uses a data-driven method to capture complex nonlinear effects. This invention is significantly superior to pure physical models (which lack flexibility) or pure data-driven models (which require massive amounts of data and have poor interpretability), and is particularly suitable for the intelligent transformation of water systems.

[0042] This invention considers multiple measurement points and the flow delay of rainfall to achieve flow prediction at the end of the sewage pipeline, which facilitates the prediction of flow results based on dynamic rainfall conditions. Furthermore, by predicting the flow of the sewage pipeline, sewage treatment plants can obtain sewage pipeline information at all times and take countermeasures in advance. Attached Figure Description

[0043] Figure 1 A flowchart of a sewage pipe network early warning method based on distributed sensors and AI prediction provided in an embodiment of the present invention.

[0044] Figure 2 This is an architecture diagram of a sewage pipe network early warning system based on distributed sensors and AI prediction, provided for an embodiment of the present invention. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the embodiments of this invention will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0046] Example 1

[0047] like Figure 1 As shown, the wastewater pipe network early warning method based on distributed sensors and AI prediction includes:

[0048] S101: Multiple measuring points are set in the sewage pipe to be tested to collect rainfall and flow information at the multiple measuring points; the measuring points include the end point of the sewage pipe and multiple measuring points in the sewage pipe.

[0049] Specifically, this invention sets up 26 measuring points in the sewage pipe to be tested, and installs a flow sensor (such as a vortex flow meter) at each measuring point to collect flow information from the 26 measuring points. The flow information includes instantaneous flow rate and cumulative flow rate change. It also collects rainfall data for the current and next two hours based on weather forecasts.

[0050] S102: Construct a prediction model for the total flow at the end of the sewage pipeline based on flow information from multiple measurement points.

[0051] S103: Optimize the total flow prediction model at the end of the sewage pipeline based on the preset neural network model.

[0052] S104: Based on the optimized prediction model for total flow at the end of the sewage pipeline, the flow at the end of the sewage pipeline is predicted to obtain the flow change trend.

[0053] This invention constructs a total flow prediction model for the end point of a sewage pipeline and optimizes it using a neural network model. By combining the physical model (total flow prediction model for the end point of the sewage pipeline) with the neural network, it retains the linear characteristics of the physical model while capturing complex nonlinear characteristics. The resulting optimized total flow prediction model for the end point of the sewage pipeline can accurately predict the total flow at the end point of the sewage pipeline, which facilitates the dynamic adjustment of the alarm threshold based on the predicted total flow and enables staff to provide early warnings.

[0054] Example 2

[0055] Based on the above embodiments, the present invention proposes a total flow prediction model for the end point of a sewage pipeline, specifically including:

[0056] To determine the flow rate trend at the end point z of the sewage pipeline, it is necessary to comprehensively consider the delayed superposition effect of the instantaneous flow rate v, the cumulative flow rate change u, and the rainfall r at each measurement point. Assuming the transmission time between adjacent measurement points is Δt, the model is constructed as follows:

[0057] The transmission time from each measurement point i to the endpoint z is:

[0058]

[0059] Where, τ i Let Δt be the transmission time from measurement point i to the endpoint z, where i is the i-th measurement point and Δt is the transmission time between adjacent measurement points.

[0060] The instantaneous flow rate at measurement point i reaches the end point z of the sewage pipe after a delay, and its contribution is: v i (t-τ i ), where v i Let t be the instantaneous flow rate at the i-th measurement point, and t be the current time.

[0061] Transmission time τ i The cumulative change is:

[0062]

[0063] Among them, u i This represents the cumulative flow change every 10 minutes during the transmission time from the i-th measurement point to the endpoint.

[0064] Regarding the contribution of rainfall, the rainfall is converted into a flow increment using a rainfall-flow conversion coefficient, specifically expressed as: α·r i (t-τ i Where α is the rainfall-to-flow conversion coefficient, and r i This refers to rainfall.

[0065] The total flow rate at the end of the sewage pipeline is the sum of the contributions from all measurement points, specifically expressed by the following formula:

[0066]

[0067] In summary, the prediction model for the total flow rate at the end of the sewage pipeline is expressed by the following formula:

[0068]

[0069] According to the formula, the instantaneous flow rate at each measurement point is affected by directly adding the current value of each measurement point to the endpoint z after a delay. i Positive / negative values ​​and transmission time τ i Together, they determine the acceleration / deceleration trend of flow changes. The effect of rainfall on point z is delayed by τ. i It becomes apparent over time and is directly proportional to the intensity of rainfall.

[0070] Preferably, when calculating the total flow rate at the end of the sewage pipeline based on the prediction model, data alignment is required. Specifically, for each time point t, the delay time of each measurement point i is calculated, and v within the corresponding time window is extracted. i (t-τ i ), u i (t-τ i ) and r i (t-τ i And if τ i For non-integer multiples of the time step, linear interpolation is used to obtain the feature values.

[0071] Preferably, the input features also need to be constructed in advance, with the input feature tensor dimension being [Batch, Time, Measurement Point, Features (Instantaneous Flow Rate, Cumulative Flow Rate Change, Rainfall)]. Furthermore, if the transmission time between adjacent measurement points Δt = 2, and the transmission from measurement point a to the sewage pipe endpoint z takes 26 × 2 = 52 minutes, then the input time window must cover at least 52 minutes of historical data.

[0072] Preferably, when training the total flow prediction model at the end of the sewage pipeline, the actual value of the flow meter set at the end of the sewage pipeline z is used as the supervision label.

[0073] Example 3

[0074] Based on the above embodiments, the present invention proposes a process for optimizing the total flow prediction model at the end of a sewage pipeline according to a preset neural network model, specifically including:

[0075] The neural network model includes a temporal feature extraction layer, a fully connected layer, and an activation function.

[0076] Temporal feature extraction layers are used to capture temporal dependencies. These layers include Long Short-Term Memory (LSTM) networks or Transformers.

[0077] Fully connected layers are used to map the output of the temporal feature extraction layer to residuals or parameter adjustments.

[0078] Activation functions are used to process the output of fully connected layers to prevent gradient vanishing. Activation functions include ReLU or Swish.

[0079] As one implementation method, the residual correction term of the total flow prediction model at the end of the sewage pipeline can be obtained through a neural network model, specifically including:

[0080] The historical flow information of all measurement points and their corresponding timestamps are input into a preset neural network model to obtain residual correction terms;

[0081] The optimized total flow prediction model for the sewage pipeline endpoint is obtained by combining the residual correction term and the total flow prediction model at the end of the sewage pipeline.

[0082] The optimized prediction model for the total flow rate at the end of the sewage pipeline is expressed by the following formula:

[0083] s pred (t)=s(t)+Δs(t)

[0084] Among them, s pred Δs(t) represents the optimized predicted total flow rate at the end of the sewage pipeline, and Δs(t) represents the residual correction term.

[0085] As another implementation method, an optimized prediction model for the total flow rate at the end of a sewage pipeline can be obtained by optimizing two parameters—the rainfall-flow conversion coefficient and the transmission time between adjacent measurement points—using a neural network model. Specifically, this includes:

[0086] The historical flow information of all measurement points and their corresponding timestamps are input into a preset neural network model to obtain the corrected rainfall-flow conversion coefficient and the transmission time between adjacent measurement points.

[0087] The prediction model for the total flow at the end of the sewage pipeline is updated based on the corrected rainfall-flow conversion coefficient and the transmission time between adjacent measurement points, resulting in an optimized prediction model for the total flow at the end of the sewage pipeline.

[0088] Preferably, the present invention also provides a process for predicting the total flow rate at the end of a sewage pipeline and training a neural network, specifically including:

[0089] The optimized prediction model for the total flow rate at the end of the sewage pipeline is transformed into a differentiable computational graph (such as a custom layer in PyTorch / TensorFlow). The delay calculation τ... i Interpolation operations must support automatic differentiation.

[0090] The neural network model and the optimized sewage pipeline terminal total flow prediction model are then jointly trained, specifically including the following steps:

[0091] Physical parameter initialization:

[0092] Δt is initialized to an empirical value (such as the theoretical time step calculated based on pipe length and flow velocity). α is initialized to an empirical value in the range of 0.1 to 1.0.

[0093] Neural network parameter initialization:

[0094] The Xavier initialization method is used to avoid gradient explosion or vanishing during the early stages of training.

[0095] First, freeze the neural network model, using only the physical parameters (Δt and α) as learnable parameters. Then, train the physical parameters using the Adam optimizer with mean squared error (MSE) as the loss function.

[0096] Then, the weights of the neural network model are unfrozen, and the physical parameters and neural network model weights are optimized simultaneously. Specifically, a gradual decay strategy of the learning rate is adopted, and the neural network model and the optimized sewage pipeline terminal total flow prediction model are trained with MSE + physical parameter regularization loss (such as constraining Δt>0).

[0097] Finally, to ensure the accuracy of the neural network model and the optimized sewage pipeline terminal total flow prediction model, time-series cross-validation was performed using a test set. The training and test sets were divided chronologically to avoid future information leakage. MSE, MAE, and peak prediction error (such as peak time deviation) were used as validation metrics.

[0098] Example 4

[0099] Based on the above embodiments, such as Figure 2 As shown, this invention proposes a wastewater pipe network early warning system based on distributed sensors and AI prediction, comprising:

[0100] A collection module is used to set up multiple measurement points in the sewage pipe to be inspected to collect rainfall and flow information at the multiple measurement points; the measurement points include the end point of the sewage pipe and multiple measurement points in the sewage pipe.

[0101] The physical model building module is used to construct a prediction model of the total flow at the end of the sewage pipeline based on flow information from multiple measurement points.

[0102] The optimization module is used to optimize the prediction model of total flow at the end of the sewage pipeline based on a preset neural network model.

[0103] The prediction module is used to predict the flow rate at the end of the sewage pipeline based on the optimized total flow rate prediction model at the end of the sewage pipeline, and to obtain the flow rate change trend.

[0104] It should be noted that the sewage pipe network early warning system based on distributed sensors and AI prediction provided in this embodiment of the invention is for implementing the above-mentioned sewage pipe network early warning method based on distributed sensors and AI prediction. Its specific functions can be referred to in the above-mentioned method embodiments, and will not be repeated here.

[0105] Example 5

[0106] Based on the above embodiments, the present invention proposes an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the sewage pipe network early warning method based on distributed sensors and AI prediction as described in Embodiment 1 above.

[0107] The present invention proposes a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device where the storage medium is located to perform the wastewater pipe network early warning method based on distributed sensors and AI prediction as described in Embodiment 1 above.

[0108] In summary, this invention combines a sewage pipeline end-point total flow prediction model with a neural network model. This retains the domain knowledge (such as delay superposition) of the original sewage pipeline end-point total flow prediction model while utilizing a data-driven approach to capture complex nonlinear effects. This invention is significantly superior to purely physical models (lacking flexibility) or purely data-driven models (requiring massive amounts of data and poor interpretability), and is particularly suitable for the intelligent transformation of water systems. This invention considers multiple measurement points and the flow delay caused by rainfall to achieve sewage pipeline end-point flow prediction, facilitating flow prediction based on dynamic rainfall conditions. Furthermore, by predicting the sewage pipeline flow, sewage treatment plants can obtain real-time sewage pipeline information and take proactive measures.

[0109] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A wastewater pipe network early warning method based on distributed sensors and AI prediction, characterized in that, include: Step 1: Set up multiple measuring points in the sewage pipe to be tested, and collect rainfall and flow information at the multiple measuring points; Step 2: Construct a prediction model for the total flow rate at the end of the sewage pipeline based on flow rate information from multiple measurement points; Step 3: Optimize the total flow prediction model at the end of the sewage pipeline based on the preset neural network model; Step 4: Based on the optimized total flow prediction model at the end of the sewage pipeline, predict the flow rate at the end of the sewage pipeline to obtain the flow rate change trend.

2. The sewage pipe network early warning method based on distributed sensors and AI prediction according to claim 1, characterized in that, The traffic information includes instantaneous traffic volume and cumulative traffic volume change.

3. The sewage pipe network early warning method based on distributed sensors and AI prediction according to claim 2, characterized in that, The prediction model for the total flow rate at the end of the sewage pipeline is expressed by the following formula: Where s(t) is the predicted total flow rate at the end of the sewage pipeline, and v i Let be the instantaneous flow rate at the i-th measurement point, Δt be the transmission time between adjacent measurement points, α be the rainfall-flow conversion coefficient, and r be the instantaneous flow rate. i Let be the rainfall, i be the i-th measurement point, and u be the... i Let t be the cumulative flow change every 10 minutes during the transmission time from the i-th measurement point to the endpoint, and t be the current time.

4. The sewage pipe network early warning method based on distributed sensors and AI prediction according to claim 3, characterized in that, Step three includes: The historical flow information of all measurement points and their corresponding timestamps are input into a preset neural network model to obtain residual correction terms; The optimized total flow prediction model for the sewage pipeline endpoint is obtained by combining the residual correction term and the total flow prediction model at the end of the sewage pipeline. The optimized prediction model for the total flow rate at the end of the sewage pipeline is expressed by the following formula: s pred (t)=s(t)+Δs(t) Among them, s pred Δs(t) represents the optimized predicted total flow rate at the end of the sewage pipeline, and Δs(t) represents the residual correction term.

5. The sewage pipe network early warning method based on distributed sensors and AI prediction according to claim 3, characterized in that, Step three also includes optimizing the total flow prediction model at the end of the sewage pipeline by adjusting the rainfall-flow conversion coefficient and the transmission time between adjacent measurement points according to a preset neural network model. The above process specifically includes: The historical flow information of all measurement points and their corresponding timestamps are input into a preset neural network model to obtain the corrected rainfall-flow conversion coefficient and the transmission time between adjacent measurement points. The prediction model for the total flow at the end of the sewage pipeline is updated based on the corrected rainfall-flow conversion coefficient and the transmission time between adjacent measurement points, resulting in an optimized prediction model for the total flow at the end of the sewage pipeline.

6. The sewage pipe network early warning method based on distributed sensors and AI prediction according to claim 4 or 5, characterized in that, The preset neural network model includes a temporal feature extraction layer, a fully connected layer, and an activation function; The temporal feature extraction layer is used to capture temporal dependencies, and the temporal feature extraction layer includes a long short-term memory network or a transformer. The fully connected layer is used to map the output of the temporal feature extraction layer to residuals or parameter adjustment amounts; The activation function is used to process the output of the fully connected layer to prevent gradient vanishing.

7. A sewage pipe network early warning system based on distributed sensors and AI prediction, characterized in that, include: The collection module is used to set up multiple measuring points in the sewage pipe to be inspected and collect rainfall and flow information at the multiple measuring points; The physical model building module is used to construct a prediction model of the total flow at the end of the sewage pipeline based on flow information from multiple measurement points. The optimization module is used to optimize the prediction model of total flow at the end of the sewage pipeline based on a preset neural network model. The prediction module is used to predict the flow rate at the end of the sewage pipeline based on the optimized total flow rate prediction model at the end of the sewage pipeline, and to obtain the flow rate change trend.

8. An electronic device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the wastewater network early warning method based on distributed sensors and AI prediction as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device where the storage medium is located to perform the wastewater pipe network early warning method based on distributed sensors and AI prediction as described in any one of claims 1 to 6.

Images