WO2023241461A1 - 污水检查井的液位预测方法、装置、电子设备及存储介质 - Google Patents

污水检查井的液位预测方法、装置、电子设备及存储介质 Download PDF

Info

Publication number
WO2023241461A1
WO2023241461A1 PCT/CN2023/099232 CN2023099232W WO2023241461A1 WO 2023241461 A1 WO2023241461 A1 WO 2023241461A1 CN 2023099232 W CN2023099232 W CN 2023099232W WO 2023241461 A1 WO2023241461 A1 WO 2023241461A1
Authority
WO
WIPO (PCT)
Prior art keywords
inspection well
sewage inspection
liquid level
time series
deep learning
Prior art date
Application number
PCT/CN2023/099232
Other languages
English (en)
French (fr)
Inventor
谢娟
Original Assignee
阿里云计算有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 阿里云计算有限公司 filed Critical 阿里云计算有限公司
Publication of WO2023241461A1 publication Critical patent/WO2023241461A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/20—Administration of product repair or maintenance
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A20/00—Water conservation; Efficient water supply; Efficient water use
    • Y02A20/152—Water filtration

Definitions

  • This application relates to the field of cloud computing technology, and in particular to a liquid level prediction method, device, electronic equipment and storage medium for sewage inspection wells.
  • the traditional hydrodynamic model prediction method relies on the basic data of the pipe network, such as the length, material, roughness, specific location, etc. of each pipe in the pipe network. Due to frequent urban renovation activities, the basic data of the pipe network changes frequently, and it is difficult to obtain higher cost. At present, the basic data of the pipeline network is seriously insufficient, causing the prediction accuracy of the hydrodynamic model to fail to meet the requirements of practical applications.
  • Embodiments of the present application provide a liquid level prediction method, device, electronic device and storage medium for sewage inspection wells to improve the liquid level prediction accuracy of sewage inspection wells and reduce prediction costs.
  • embodiments of the present application provide a liquid level prediction method for sewage inspection wells, including:
  • the reference factor values include the actual liquid level time series of the sewage inspection well, the actual rainfall time series in the area where the sewage inspection well is located, and the forecast of the area where the sewage inspection well is located. Rainfall time series;
  • the future liquid level time series of the sewage inspection well is predicted based on the reference factor value.
  • embodiments of the present application provide a liquid level prediction device for sewage inspection wells, including:
  • the acquisition module is used to obtain the reference factor values related to the future liquid level time series of the sewage inspection well.
  • the reference factor values include the actual liquid level time series of the sewage inspection well, the actual rainfall time series of the area where the sewage inspection well is located, and the sewage inspection Forecast rainfall time series for the area where the well is located;
  • the prediction module is used to predict the risk of overflow in sewage inspection wells based on reference factor values. Taking the factor values into account, the future liquid level time series of the sewage inspection well is predicted.
  • embodiments of the present application provide an electronic device, including a memory, a processor, and a computer program stored on the memory.
  • the processor executes the computer program, it implements the method provided by any embodiment of the present application.
  • embodiments of the present application provide a computer-readable storage medium.
  • a computer program is stored in the computer-readable storage medium.
  • the computer program is executed by a processor, the method provided by any embodiment of the present application is implemented.
  • the liquid level prediction method, device, electronic equipment and storage medium of the sewage inspection well obtain reference factor values related to the future liquid level time series of the sewage inspection well.
  • the reference factor values include the actual liquid level of the sewage inspection well. time series, the actual rainfall time series in the area where the sewage inspection well is located, and the forecast rainfall time series in the area where the sewage inspection well is located; when it is determined that there is an overflow risk in the sewage inspection well based on the reference factor value, based on the reference factor value, the prediction Future level time series of wastewater manholes.
  • the embodiment of this application predicts the future liquid level time series of the sewage inspection well based on the actual liquid level time series of the sewage inspection well, the actual rainfall time series of the area where the sewage inspection well is located, and the forecast rainfall time series of the area where the sewage inspection well is located,
  • the prediction results are highly accurate; no basic data of the pipe network is required, and the prediction cost is low.
  • Figure 1 is a schematic diagram of an application scenario of the liquid level prediction method for sewage inspection wells provided by an embodiment of the present application;
  • Figure 2 is a flow chart of a liquid level prediction method for sewage inspection wells provided by an embodiment of the present application
  • Figure 3 is a schematic diagram of a liquid level prediction method for sewage inspection wells provided by an embodiment of the present application
  • Figure 4 is a schematic diagram of the online training process of the deep learning model provided by an embodiment of the present application.
  • Figure 5 is a schematic diagram of a liquid level prediction device for a sewage inspection well provided by an embodiment of the present application
  • FIG. 6 is a block diagram of an electronic device used to implement an embodiment of the present application.
  • FIG. 1 is a schematic diagram of an application scenario of the liquid level prediction method for sewage inspection wells provided by an embodiment of the present application.
  • the liquid level prediction method for sewage inspection wells provided in this embodiment can be deployed in a cloud server or a local server.
  • the reference factor values related to the future liquid level time series of the sewage inspection well are obtained through the Internet of Things devices, where the Internet of Things devices can include liquid level gauges, rain gauges, etc. IoT devices connect to servers through IoT protocols.
  • the reference factor values include the actual liquid level time series of the sewage inspection well, the actual rainfall time series of the area where the sewage inspection well is located, and the forecast rainfall time series of the area where the sewage inspection well is located.
  • the sewage inspection well Based on the reference factor values, it is determined that the sewage inspection well has an overflow risk.
  • the future liquid level time series of the sewage inspection well is predicted through the deep learning model.
  • Embodiments of the present application provide a liquid level prediction method for sewage inspection wells.
  • Figure 2 is a flow chart of a liquid level prediction method for sewage inspection wells according to an embodiment of the present application. This method can be applied to liquid level prediction for sewage inspection wells.
  • Device which can be deployed in a cloud server, local server or other processing equipment.
  • the method can also be implemented by the processor calling computer-readable instructions stored in the memory. As shown in Figure 2, the method includes:
  • Step S201 Obtain reference factor values related to the future liquid level time series of the sewage inspection well.
  • the reference factor values include the actual liquid level time series of the sewage inspection well, the actual rainfall time series of the area where the sewage inspection well is located, and the location of the sewage inspection well. Forecast rainfall time series for the region.
  • the execution subject is the server.
  • the server can obtain the reference factor values related to the future liquid level time series of the sewage inspection well through Internet of Things devices, liquid level gauges, rain gauges, meteorological systems, etc.
  • the actual liquid level time series includes historical and current inspection well liquid level height data. Sampling is performed according to the preset sampling frequency to obtain multiple liquid level heights.
  • the data form a liquid level time series.
  • the sampling frequency can be on the order of minutes.
  • the area where the sewage inspection well is located can be a range with the sewage inspection well as the center and a preset length as the radius. For example, through the rain gauge within 5 kilometers around the sewage inspection well, the historical and current rainfall data of the area where the sewage inspection well is located can be obtained according to Sampling is performed at a preset sampling frequency, and multiple rainfall data at different times form a rainfall time series. For example, the sampling frequency can be sampled once every 5 minutes.
  • perform data cleaning which may include: deduplication of duplicate data, filtering of data exceeding the preset range, and filtering data in the sequence. Missing values are filled in through interpolation, etc.
  • the reference factors related to the future liquid level time series of the sewage inspection well may also include other influencing factors, which is not limited in this embodiment.
  • Step S202 When it is determined that the sewage inspection well has an overflow risk based on the reference factor value, predict the future liquid level time series of the sewage inspection well based on the reference factor value.
  • the liquid level prediction method of the sewage inspection well obtains reference factor values related to the future liquid level time series of the sewage inspection well.
  • the reference factor values include the actual liquid level time series of the sewage inspection well, the location of the sewage inspection well The actual rainfall time series of the area and the forecast rainfall time series of the area where the sewage inspection well is located; when it is determined that the sewage inspection well has an overflow risk based on the reference factor value, the future liquid level time of the sewage inspection well is predicted based on the reference factor value. sequence.
  • the embodiment of this application predicts the future liquid level time series of the sewage inspection well based on the actual liquid level time series of the sewage inspection well, the actual rainfall time series of the area where the sewage inspection well is located, and the forecast rainfall time series of the area where the sewage inspection well is located,
  • the prediction results are highly accurate; no basic data of the pipe network is required, and the prediction cost is low.
  • step S202 predicting the future liquid level time series of the sewage inspection well based on the reference factor value, includes: inputting the reference factor value into a deep learning model based on the attention mechanism, and obtaining the output of the deep learning model.
  • the deep learning model based on the attention mechanism gives different weights to the peaks of the training samples.
  • the future liquid level time series of sewage inspection wells can be predicted through a deep learning model based on the attention mechanism.
  • the actual liquid level time series of the sewage inspection well, the actual rainfall time series of the area where the sewage inspection well is located, and the forecast rainfall time series of the area where the sewage inspection well is located are numerically input into the deep learning model based on the attention mechanism, and the model outputs sewage Predicted values of future liquid level time series for inspection wells.
  • the data is preprocessed, specifically including: cleaning and normalizing the historical liquid level time series, historical rainfall and liquid level time series, and forecast rainfall time series to form a multi-dimensional input vector.
  • the deep learning model can be any neural network model that can achieve the prediction function.
  • the deep learning model can be a recurrent neural network (RNN).
  • RNN recurrent neural network
  • the attention mechanism refers to adding an attention mechanism layer to the deep learning model.
  • the peak value of the sample is given different weights. For example, compared with a liquid level sample that is not a peak value, the liquid level is Setting a larger weight for the peak value of the sample can make the model pay more attention to the peak value of the liquid level sample than other samples during the training process.
  • the deep learning model based on the attention mechanism is trained based on training samples.
  • the training samples include the sample historical liquid level time series in the sewage inspection well and the sample historical rainfall time series in the area where the sewage inspection well is located. .
  • the historical liquid level time series is historical liquid level height data obtained through actual measurement.
  • the sample historical rainfall time series in the area where the sewage inspection well is located can be the historical rainfall data obtained from actual measurements.
  • the liquid level prediction method for sewage inspection wells also includes:
  • the deep learning model can continue to be adaptively trained and updated online.
  • the future liquid level time series of the sewage inspection well is obtained. Predicted value. After a period of time, the actual value of the future liquid level time series of the sewage inspection well is obtained through actual measurement. The actual value and the predicted value are subtracted. The obtained difference is compared with the preset threshold. If the difference exceeds If the threshold is preset, the training sample is updated. Optionally, the actual value of the future liquid level time series of the sewage inspection well can be added to the training sample to obtain the updated training sample. Use the updated training sample. Train deep learning models. In this embodiment, by updating the training samples and using the updated training samples to train the deep learning model, the prediction effect of the deep learning model can be improved.
  • the liquid level prediction method for sewage inspection wells also includes:
  • the current deep learning model For each prediction of the updated deep learning model, determine the first absolute error between the predicted value output by the deep learning model and the actual value for the current prediction, and the second absolute error between the predicted value output by the deep learning model and the actual value for the previous prediction. Error; when the first absolute error is less than or equal to the second absolute error, the current deep learning model is determined as a model for predicting the future liquid level time series of the sewage inspection well.
  • the updated deep learning model is used for online prediction, and the parameters, weights, preset thresholds and absolute errors between the predicted values and actual values of the deep learning model are saved, and subsequently updated based on adaptive learning.
  • the absolute error can be the absolute value of the difference between the predicted value and the actual value.
  • the liquid level prediction method for sewage inspection wells also includes:
  • For each prediction of the updated deep learning model determine the first absolute error between the predicted value output by the deep learning model and the actual value for the current prediction, and the second absolute error between the predicted value output by the deep learning model and the actual value for the previous prediction. error; when the first absolute error is greater than the second absolute error, update the training samples of the current deep learning model, and use the updated training samples to train the deep learning model.
  • the updated deep learning model is used for online prediction, and the parameters, weights, preset thresholds and absolute errors between the predicted values and actual values of the deep learning model are saved, and subsequently updated based on adaptive learning.
  • For each prediction of the model calculate the first absolute error between the predicted value output by the deep learning model and the actual value for the current prediction, and the second absolute error between the predicted value output by the deep learning model and the actual value for the previous prediction, and divide the first Compare the absolute error with the second absolute error. If the first absolute error is greater than the second absolute error, it means that the accuracy of the prediction results of the deep learning model does not meet the requirements. Then the training samples of the current deep learning model can be updated. If you choose, you can add new training samples, or replace some of the training samples, and use the updated training samples to train the deep learning model.
  • the risk of overflow in a sewage manhole is determined in the following way:
  • the reference factor values are compared with the overflow warning threshold respectively. If at least one of the reference factor values exceeds the overflow warning threshold, it is determined that there is an overflow risk in the sewage inspection well.
  • the reference factor values related to the future liquid level time series of the sewage inspection well are obtained.
  • the reference factor values include the actual liquid level time series of the sewage inspection well, the actual rainfall time series of the area where the sewage inspection well is located, and the sewage inspection Forecast rainfall time series in the area where the well is located, each value in the actual liquid level time series of the sewage inspection well, the actual rainfall time series in the area where the sewage inspection well is located, and the forecast rainfall time series in the area where the sewage inspection well is located are combined with the overflow time series respectively.
  • the overflow warning threshold If at least one of the reference factor values exceeds the overflow warning threshold, it is determined that the sewage inspection well has an overflow risk, and the future liquid level time series of the sewage inspection well is further predicted, and overflow prediction information is generated. Send it to the user terminal of the pipeline network operation and maintenance personnel to notify the pipeline network operation and maintenance personnel in advance.
  • the overflow warning threshold set the percentage of the well depth as the overflow warning threshold. If the liquid level exceeds the overflow warning threshold multiple times in a row, the overflow warning will be triggered.
  • FIG 3 is a schematic diagram of a liquid level prediction method for a sewage inspection well provided by an embodiment of the present application.
  • the server obtains the liquid level time series in the sewage inspection well through the liquid level gauge in the sewage inspection well, including historical liquid level data and current liquid level data; it obtains the rainfall time series through the rain gauge within the scope of the sewage inspection well. This includes historical rainfall data and current rainfall data; rainfall forecast time series are obtained through the meteorological system.
  • the liquid level time series in the sewage inspection well, the rainfall time series obtained from the rain gauge within the inspection well, and the rainfall forecast time series are used for data cleaning, including: deduplication of duplicate data, filtering of data exceeding the preset range, and Missing values in the sequence are filled in through interpolation, etc.
  • Figure 4 is a schematic diagram of the online training process of the deep learning model provided by an embodiment of the present application.
  • the server obtains the actual value of the historical liquid level time series of the sewage inspection well (the original historical liquid level time series as shown in the figure), and the historical rainfall time series of the area where the sewage inspection well is located (the original historical rainfall time as shown in the figure). sequence), and forecast future rainfall time series, preprocessing these data, specifically including: cleaning and normalizing the historical liquid level time series, historical rainfall liquid level time series, and forecast rainfall time series to form a multi-dimensional input vector . Input the multi-dimensional input vector into the deep learning model for prediction, and obtain the prediction results for T hours in the future.
  • the actual value of the liquid level time series for T hours in the future is obtained.
  • the prediction results are compared with the actual results, and the prediction results are calculated and
  • the error of the actual result determines whether the error between the predicted result and the actual result exceeds the preset threshold. If the error does not exceed the preset threshold, continue to use the current deep learning model for prediction; if the error exceeds the preset threshold, it means that the prediction result deviates too much. If it is large, the training sample set is updated, the measured results are returned to the training sample set, and other untrained historical measured data are also added to the training sample set.
  • the embodiment of this application performs early warning prediction based on historical monitoring data and meteorological forecast rainfall data, avoiding the high-cost acquisition of basic data of the drainage pipe network and hydrodynamic modeling work. Based on the prediction error, the training samples and model parameters are updated to perform adaptive learning. , the prediction results are highly accurate and the prediction cost is low.
  • the embodiments of the present application also provide a liquid level prediction device for a sewage inspection well.
  • the liquid level prediction device of the sewage inspection well may include:
  • the acquisition module 501 is used to obtain reference factor values related to the future liquid level time series of the sewage inspection well.
  • the reference factor values include the actual liquid level time series of the sewage inspection well, the actual rainfall time series of the area where the sewage inspection well is located, and the sewage Forecast rainfall time series for the area where the manhole is located;
  • the prediction module 502 is used to predict the future liquid level time series of the sewage inspection well based on the reference factor value when it is determined that the sewage inspection well has an overflow risk based on the reference factor value.
  • the liquid level prediction device of the sewage inspection well obtains reference factor values related to the future liquid level time series of the sewage inspection well.
  • the reference factor values include the actual liquid level time series of the sewage inspection well, the location of the sewage inspection well The actual rainfall time series of the area and the forecast rainfall time series of the area where the sewage inspection well is located; when it is determined that the sewage inspection well has an overflow risk based on the reference factor value, the future liquid level time of the sewage inspection well is predicted based on the reference factor value. sequence.
  • the embodiment of this application predicts the future liquid level time series of the sewage inspection well based on the actual liquid level time series of the sewage inspection well, the actual rainfall time series of the area where the sewage inspection well is located, and the forecast rainfall time series of the area where the sewage inspection well is located,
  • the prediction results are highly accurate; no basic data of the pipe network is required, and the prediction cost is low.
  • the prediction module 502 is used to:
  • the deep learning model based on the attention mechanism gives different weights to the peak values of the training samples. .
  • the deep learning model based on the attention mechanism is trained based on training samples.
  • the training samples include the sample historical liquid level time series in the sewage inspection well and the sample historical rainfall time series in the area where the sewage inspection well is located. .
  • the liquid level prediction device of the sewage inspection well also includes an update module for:
  • the training samples are updated, and the updated training samples are used to train the deep learning model to obtain the updated deep learning model.
  • the liquid level prediction device of the sewage inspection well also includes a first error comparison module for:
  • For each prediction of the updated deep learning model determine the first absolute error between the predicted value output by the deep learning model and the actual value for the current prediction, and the second absolute error between the predicted value output by the deep learning model and the actual value for the previous prediction. error;
  • the current deep learning model is determined as a model for predicting the future liquid level time series of the sewage inspection well.
  • the liquid level prediction device of the sewage inspection well also includes a second error comparison module for:
  • For each prediction of the updated deep learning model determine the first absolute error between the predicted value output by the deep learning model and the actual value for the current prediction, and the second absolute error between the predicted value output by the deep learning model and the actual value for the previous prediction. error;
  • the training samples of the current deep learning model are updated, and the updated training samples are used to train the deep learning model.
  • the risk of overflow in a sewage manhole is determined in the following way:
  • the reference factor values are compared with the overflow warning threshold respectively. If at least one of the reference factor values exceeds the overflow warning threshold, it is determined that there is an overflow risk in the sewage inspection well.
  • each module in each device of the embodiment of the present application can be found in the corresponding description in the above method, and have corresponding beneficial effects, which will not be described again here.
  • FIG. 6 is a block diagram of an electronic device used to implement an embodiment of the present application.
  • the electronic device includes: a memory 610 and a processor 620 .
  • the memory 610 stores a computer program that can run on the processor 620 .
  • the processor 620 executes the computer program, the method in the above embodiment is implemented.
  • the number of memory 610 and processor 620 may be one or more.
  • the electronic device also includes:
  • the communication interface 630 is used to communicate with external devices and perform data interactive transmission.
  • the bus may be an Industry Standard Architecture (ISA) bus, a peripheral device interconnection (Peripheral Component Interconnect (PCI) bus or Extended Industry Standard Architecture (EISA) bus, etc.
  • ISA Industry Standard Architecture
  • PCI peripheral device interconnection
  • EISA Extended Industry Standard Architecture
  • the bus can be divided into address bus, data bus, control bus, etc. For ease of presentation, only one thick line is used in Figure 6, but it does not mean that there is only one bus or one type of bus.
  • the memory 610, the processor 620 and the communication interface 630 are integrated on one chip, the memory 610, the processor 620 and the communication interface 630 can communicate with each other through the internal interface.
  • Embodiments of the present application provide a computer-readable storage medium, which stores a computer program. When the program is executed by a processor, the method provided in the embodiment of the present application is implemented.
  • An embodiment of the present application also provides a chip, which includes a processor for calling and running instructions stored in the memory, so that the communication device installed with the chip executes the method provided by the embodiment of the present application.
  • Embodiments of the present application also provide a chip, including: an input interface, an output interface, a processor and a memory.
  • the input interface, the output interface, the processor and the memory are connected through an internal connection path, and the processor is used to execute the code in the memory. , when the code is executed, the processor is used to execute the method provided by the application embodiment.
  • processor can be a central processing unit (Central Processing Unit, CPU), or other general-purpose processor, digital signal processor (Digital Signal Processor, DSP), application specific integrated circuit (Application Specific Integrated Circuit, ASIC), Field Programmable Gate Array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
  • CPU Central Processing Unit
  • DSP Digital Signal Processor
  • ASIC Application Specific Integrated Circuit
  • FPGA Field Programmable Gate Array
  • a general-purpose processor can be a microprocessor or any conventional processor, etc. It is worth noting that the processor may be a processor that supports Advanced RISC Machines (ARM) architecture.
  • ARM Advanced RISC Machines
  • the above-mentioned memory may include read-only memory and random access memory, and may also include non-volatile random access memory.
  • the memory may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory.
  • non-volatile memory can include read-only memory (Read-Only Memory, ROM), programmable read-only memory (Programmable ROM, PROM), erasable programmable read-only memory (Erasable PROM, EPROM), electrically removable memory. Erase programmable read-only memory (Electrically EPROM, EEPROM) or flash memory.
  • Volatile memory may include Random Access Memory (RAM), which acts as an external cache.
  • RAM static random access memory
  • DRAM dynamic random access memory
  • DRAM synchronous dynamic random access memory
  • DDR SDRAM double data rate synchronous dynamic random access Memory
  • Enhanced SDRAM, ESDRAM enhanced synchronous dynamic random access memory
  • Synchlink DRAM, SLDRAM synchronous link dynamic random access memory
  • Direct Rambus RAM Direct Rambus RAM
  • a computer program product includes one or more computer instructions.
  • the computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable device.
  • Computer instructions may be stored in or transmitted from one computer-readable storage medium to another computer-readable storage medium.
  • references to the terms “one embodiment,” “some embodiments,” “an example,” “specific examples,” or “some examples” or the like means that specific features are described in connection with the embodiment or example.
  • structures, materials or features are included in at least one embodiment or example of the present application.
  • the specific features, structures, materials or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
  • those skilled in the art may combine and combine different embodiments or examples and features of different embodiments or examples described in this specification unless they are inconsistent with each other.
  • first and second are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity of indicated technical features. Thus, features defined as “first” and “second” may explicitly or implicitly include at least one of these features. In the description of this application, “plurality” means two or more than two, unless otherwise explicitly and specifically limited.
  • logic and/or steps represented in the flowcharts or otherwise described herein, for example, may be considered a sequenced list of executable instructions for implementing the logical functions, and may be embodied in any computer-readable medium, For use by, or in combination with, instruction execution systems, devices or devices (such as computer-based systems, systems including processors or other systems that can fetch instructions from and execute instructions from the instruction execution system, device or device) or equipment.
  • various parts of the present application may be implemented in hardware, software, firmware, or a combination thereof.
  • various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. All or part of the steps of the method in the above embodiment can be completed by instructing relevant hardware through a program.
  • the program can be stored in a computer-readable storage medium. When executed, the program includes one of the steps of the method embodiment or other steps. combination.
  • each functional unit in various embodiments of the present application can be integrated into a processing module, or each unit can exist physically alone, or two or more units can be integrated into one module.
  • the above integrated modules can be implemented in the form of hardware or software function modules. If the above integrated modules are implemented in the form of software function modules and sold or used as independent products, they can also be stored in a computer-readable storage medium.
  • the storage medium can be a read-only memory, a magnetic disk or an optical disk, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • General Physics & Mathematics (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Operations Research (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Sewage (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

本申请提供了一种污水检查井的液位预测方法、装置、电子设备及存储介质，涉及云计算技术领域。方法包括：获取与污水检查井的未来液位时间序列相关的参考因素数值，参考因素数值包括污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列；在基于参考因素数值确定污水检查井存在溢流风险的情况下，基于参考因素数值，预测污水检查井的未来液位时间序列。本申请实施例，根据污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列，预测污水检查井的未来液位时间序列，预测结果准确度高；不需要管网基础数据，预测成本低。

Description

污水检查井的液位预测方法、装置、电子设备及存储介质
本申请要求于2022年06月14日提交中国专利局、申请号为202210674810.X、申请名称为“污水检查井的液位预测方法、装置、电子设备及存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及云计算技术领域，尤其涉及一种污水检查井的液位预测方法、装置、电子设备及存储介质。
背景技术
随着城市化进程的加速，城市改造活动增加。城市排污管网容易出现老化、雨污混接、堵塞等问题，经常造成城市污水管网系统超负荷运行、或被异常堵塞，导致发生检查井溢流，造成不良的社会和环境影响。提前对污水检查井进行预警预测，辅助管网运维人员提早对污水管网进行运维和清梳，对提升市民安全以及减少环境污染有着重要意义。
传统的水动力模型预测方法，依赖于管网基础数据，例如，管网中的各管道的长度、材质、粗糙程度、具体位置等，由于城市改造活动频繁，导致管网基础数据经常变化，获取成本较高。目前，管网基础数据严重不足，导致水动力模型预测精度不能满足实际应用的要求。
发明内容
本申请实施例提供一种污水检查井的液位预测方法、装置、电子设备及存储介质，以提高污水检查井的液位预测精度，降低预测成本。
第一方面，本申请实施例提供了一种污水检查井的液位预测方法，包括：
获取与污水检查井的未来液位时间序列相关的参考因素数值，参考因素数值包括污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列；
在基于参考因素数值确定污水检查井存在溢流风险的情况下，基于参考因素数值，预测污水检查井的未来液位时间序列。
第二方面，本申请实施例提供了一种污水检查井的液位预测装置，包括：
获取模块，用于获取与污水检查井的未来液位时间序列相关的参考因素数值，参考因素数值包括污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列；
预测模块，用于在基于参考因素数值确定污水检查井存在溢流风险的情况下，基于参 考因素数值，预测污水检查井的未来液位时间序列。
第三方面，本申请实施例提供一种电子设备，包括存储器、处理器及存储在存储器上的计算机程序，处理器在执行计算机程序时实现本申请任一实施例提供的方法。
第四方面，本申请实施例提供一种计算机可读存储介质，计算机可读存储介质内存储有计算机程序，计算机程序被处理器执行时实现本申请任一实施例提供的方法。
与现有技术相比，本申请具有如下优点：
本申请实施例提供的污水检查井的液位预测方法、装置、电子设备及存储介质，获取与污水检查井的未来液位时间序列相关的参考因素数值，参考因素数值包括污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列；在基于参考因素数值确定污水检查井存在溢流风险的情况下，基于参考因素数值，预测污水检查井的未来液位时间序列。本申请实施例，根据污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列，预测污水检查井的未来液位时间序列，预测结果准确度高；不需要管网基础数据，预测成本低。
上述概述仅仅是为了说明书的目的，并不意图以任何方式进行限制。除上述描述的示意性的方面、实施方式和特征之外，通过参考附图和以下的详细描述，本申请进一步的方面、实施方式和特征将会是容易明白的。
附图说明
在附图中，除非另外规定，否则贯穿多个附图相同的附图标记表示相同或相似的部件或元素。这些附图不一定是按照比例绘制的。应该理解，这些附图仅描绘了根据本申请公开的一些实施方式，而不应将其视为是对本申请范围的限制。
图1为本申请一实施例提供的污水检查井的液位预测方法应用场景的示意图；
图2为本申请一实施例提供的污水检查井的液位预测方法的流程图；
图3为本申请一实施例提供的污水检查井的液位预测方法的示意图；
图4为本申请一实施例提供的深度学习模型的线上训练过程的示意图；
图5为本申请一实施例提供的污水检查井的液位预测装置的示意图；
图6为用来实现本申请实施例的电子设备的框图。
具体实施方式
在下文中，仅简单地描述了某些示例性实施例。正如本领域技术人员可认识到的那样，在不脱离本申请的精神或范围的情况下，可通过各种不同方式修改所描述的实施例。因此，附图和描述被认为本质上是示例性的而非限制性的。
为便于理解本申请实施例的技术方案，以下对本申请实施例的相关技术进行说明，以下相关技术作为可选方案与本申请实施例的技术方案可以进行任意结合，其均属于本申请 实施例的保护范围。
为了更清楚地展示本申请实施例中提供的污水检查井的液位预测方法，首先介绍可用于实现该方法的应用场景。
图1为本申请一实施例提供的污水检查井的液位预测方法应用场景的示意图。本实施例提供的污水检查井的液位预测方法，可以部署于云服务器或本地服务器中。通过物联网设备获取与污水检查井的未来液位时间序列相关的参考因素数值，其中，物联网设备可以包括液位计、雨量计等。物联网设备通过物联网协议连接到服务器。参考因素数值包括污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列，在基于参考因素数值确定污水检查井存在溢流风险的情况下，进行溢流预警，基于参考因素数值，通过深度学习模型预测污水检查井的未来液位时间序列。发送溢流预测信息到管网运维人员的用户终端进行消息通知，提前通知管网运维人员进行辅助决策和运维工作。
本申请实施例提供了一种污水检查井的液位预测方法，图2是本申请一实施例的污水检查井的液位预测方法的流程图，该方法可以应用于污水检查井的液位预测装置，该装置可以部署于云服务器、本地服务器或其它处理设备中。在一些可能的实现方式中，该方法还可以通过处理器调用存储器中存储的计算机可读指令的方式来实现。如图2所示，该方法包括：
步骤S201，获取与污水检查井的未来液位时间序列相关的参考因素数值，参考因素数值包括污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列。
本申请实施例中，执行主体为服务器。服务器可以通过物联网设备，液位计、雨量计、气象系统等，获取与污水检查井的未来液位时间序列相关的参考因素数值。
在污水检查井内安装液位计，获取污水检查井的实际液位时间序列，实际液位时间序列包括历史和当前检查井液面高度数据，按照预设采样频率进行采样，得到多个液面高度数据组成液位时间序列。例如，采样频率可以是分钟级的。
通过污水检查井所在区域的雨量计，获取污水检查井所在区域的实际降雨时间序列。污水检查井所在区域可以是以污水检查井为中心，预设长度为半径的范围，例如，通过污水检查井周围5公里范围的雨量计，获取历史和当前污水检查井所在区域的雨量数据，按照预设采样频率进行采样，不同时间的多个雨量数据组成降雨时间序列，例如，采样频率可以是5分钟进行一次采样。
通过气象系统获得污水检查井所在区域的预报降雨时间序列，例如，获取污水检查井周围5公里范围的未来1小时(10分钟间隔)预报降雨数据。
可选的，获取到与污水检查井的未来液位时间序列相关的参考因素数值之后，进行数据清洗，具体可以包括：重复数据去重、对超过预设范围的数据进行过滤、对于序列中的缺值通过插值进行补全等。
需要说明的是，与污水检查井未来的液位时间序列相关的参考因素还可以包括其他影响因素，本实施例对此不作限定。
步骤S202，在基于参考因素数值确定污水检查井存在溢流风险的情况下，基于参考因素数值，预测污水检查井的未来液位时间序列。
基于参考因素数值中的各项确定污水检查井是否存在溢流风险，如果存在溢流风险，则预测未来液位时间序列，例如，未来1小时内，10分钟间隔一次的液面高度数据，并生成溢流预测信息发送到用户终端，提前通知管网运维人员。
本申请实施例提供的污水检查井的液位预测方法，获取与污水检查井的未来液位时间序列相关的参考因素数值，参考因素数值包括污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列；在基于参考因素数值确定污水检查井存在溢流风险的情况下，基于参考因素数值，预测污水检查井的未来液位时间序列。本申请实施例，根据污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列，预测污水检查井的未来液位时间序列，预测结果准确度高；不需要管网基础数据，预测成本低。
其中，污水检查井的未来液位时间序列的具体预测方式见如下实施例：
在一种可能的实现方式中，步骤S202，基于参考因素数值，预测污水检查井的未来液位时间序列，包括：将参考因素数值输入基于注意力机制的深度学习模型，得到深度学习模型输出的污水检查井的未来液位时间序列的预测值，基于注意力机制的深度学习模型对训练样本的峰值赋予不同的权重。
在实际应用中，可以通过基于注意力机制的深度学习模型预测污水检查井的未来液位时间序列。将污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列等参考因素数值输入基于注意力机制的深度学习模型，模型输出污水检查井的未来液位时间序列的预测值。
可选的，数据输入模型之前，对数据进行预处理，具体包括：对历史液位时间序列、历史雨量液位时间序列、预报雨量时间序列进行清洗和归一化处理，形成多维输入向量。
其中，深度学习模型可以是任意的可以实现预测功能的神经网络模型，可选的，深度学习模型可以是循环神经网络(Recurrent Neural Network,RNN)。注意力机制是指在深度学习模型中增加注意力机制层，通过设置函数关注液位样本的峰值，对样本的峰值赋予不同的权重，例如，与不是峰值的液位样本相比，为液位样本的峰值设置较大的权重，可以使模型在训练过程中对液位样本的峰值的关注度高于其他样本。
在一种可能的实现方式中，基于注意力机制的深度学习模型是基于训练样本训练得到的，训练样本包括污水检查井内的样本历史液位时间序列、污水检查井所在区域的样本历史降雨时间序列。
在实际应用中，获取多个历史液位时间序列，历史液位时间序列为经过实际测量得到的历史液位高度数据。将多个历史液位时间序列分为两个部分：第一液位时间序列和第二 液位时间序列，第二液位时间序列中的时间晚于第一液位时间序列，可以理解为，第二液位时间序列可以作为第一液位时间序列的未来预测值对应的真实值。污水检查井所在区域的样本历史降雨时间序列可以为实际测量得到的历史降雨数据。
在一种可能的实现方式中，污水检查井的液位预测方法还包括：
获取污水检查井的未来液位时间序列的实际值；在未来液位时间序列的预测值与实际值的差值超过预设阈值的情况下，对训练样本进行更新，并利用更新后的训练样本训练深度学习模型，得到更新后的深度学习模型。
在实际应用中，利用污水检查井内的样本历史液位时间序列、污水检查井所在区域的样本历史降雨时间序列训练得到深度学习模型之后，可以将深度学习模型在线上继续进行自适应训练和更新。
将污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列输入深度学习模型之后，得到污水检查井的未来液位时间序列的预测值，经过一段时间之后，通过实际测量得到污水检查井的未来液位时间序列的实际值，将实际值和预测值做减法运算，得到的差值与预设阈值进行比较，如果差值超过预设阈值，则对训练样本进行更新，可选的，可以将实际测量得到污水检查井的未来液位时间序列的实际值添加到训练样本，得到更新之后的训练样本，利用更新后的训练样本训练深度学习模型。本实施例中，通过对训练样本进行更新，利用更新后的训练样本训练深度学习模型，可以提高深度学习模型的预测效果。
在一种可能的实现方式中，污水检查井的液位预测方法还包括：
对于更新后的深度学习模型的每一次预测，确定当前次预测深度学习模型输出的预测值与实际值的第一绝对误差，以及前一次预测深度学习模型输出的预测值与实际值的第二绝对误差；在第一绝对误差小于或等于第二绝对误差的情况下，将当前的深度学习模型确定为预测污水检查井的未来液位时间序列的模型。
在实际应用中，利用更新后的深度学习模型进行线上预测，保存深度学习模型的参数、权重、预设阈值以及预测值和实际值之间的绝对误差，后续根据自适应学习进行更新。
对于模型的每一次预测，计算当前次预测深度学习模型输出的预测值与实际值的第一绝对误差，以及前一次预测深度学习模型输出的预测值与实际值的第二绝对误差，将第一绝对误差和第二绝对误差进行比较，如果第一绝对误差小于或等于第二绝对误差，则说明深度学习模型的预测结果的准确度符合要求，则将当前的深度学习模型作为预测污水检查井的未来液位时间序列的模型。其中，绝对误差可以是预测值与实际值的差值的绝对值。
在一种可能的实现方式中，污水检查井的液位预测方法还包括：
对于更新后的深度学习模型的每一次预测，确定当前次预测深度学习模型输出的预测值与实际值的第一绝对误差，以及前一次预测深度学习模型输出的预测值与实际值的第二绝对误差；在第一绝对误差大于第二绝对误差的情况下，对当前的深度学习模型的训练样本进行更新，并利用更新后的训练样本训练深度学习模型。
在实际应用中，利用更新后的深度学习模型进行线上预测，保存深度学习模型的参数、权重、预设阈值以及预测值和实际值之间的绝对误差，后续根据自适应学习进行更新。
对于模型的每一次预测，计算当前次预测深度学习模型输出的预测值与实际值的第一绝对误差，以及前一次预测深度学习模型输出的预测值与实际值的第二绝对误差，将第一绝对误差和第二绝对误差进行比较，如果第一绝对误差大于第二绝对误差，则说明深度学习模型的预测结果的准确度不符合要求，则对当前的深度学习模型的训练样本进行更新，可选的，可以增加新的训练样本，或者替换掉一部分训练样本，并利用更新后的训练样本训练深度学习模型。
在一种可能的实现方式中，污水检查井存在溢流风险是通过以下方式确定的：
将参考因素数值分别与溢流预警阈值进行比较，若参考因素数值中的至少一项超过溢流预警阈值，则确定污水检查井存在溢流风险。
在实际应用中，获取与污水检查井的未来液位时间序列相关的参考因素数值，参考因素数值包括污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列，将污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列中的各个数值，分别与溢流预警阈值进行比较，如果参考因素数值中的至少一项超过溢流预警阈值，则确定污水检查井存在溢流风险，则进一步预测污水检查井的未来液位时间序列，并生成溢流预测信息发送到管网运维人员的用户终端，提前通知管网运维人员。
可选的，设置井深的百分比作为溢流预警阈值，液位连续多次超过溢流预警阈值，触发溢流预警。
为了使本申请技术方案的过程更加易于理解，下面通过几个具体的实施例对本申请技术方案的实现过程进行详细介绍。
图3为本申请一实施例提供的污水检查井的液位预测方法的示意图。如图3所示，服务器通过污水检查井内液位计获取污水检查井内的液位时间序列，其中包括历史液位数据和当前液位数据；通过污水检查井所在范围的雨量计获取降雨时间序列，其中包括历史降雨数据和当前降雨数据；通过气象系统获取降雨预报时间序列。将污水检查井内的液位时间序列，检查井所在范围的雨量计获取降雨时间序列，以及降雨预报时间序列进行数据清洗，具体包括：重复数据去重、对超过预设范围的数据进行过滤、对于序列中的缺值通过插值进行补全等。将清洗之后的数据中的各个数值分别与溢流预警阈值进行比较，设置井深的百分比作为溢流预警阈值，如果液位连续多次超过溢流预警阈值，则确定污水检查井存在溢流风险，则进行溢流预警，启动液位预测，将清洗之后的数据构造成多维输入向量，输入基于注意力机制的深度学习模型，进行溢流预测，得到污水检查井的未来液位时间序列的预测值，生成溢流事件预测诊断信息，发送到管网运维人员的用户终端，进行移动端消息通知，以便于管网运维人员进行辅助决策和运维工作。
图4为本申请一实施例提供的深度学习模型的线上训练过程的示意图。如图4所示， 服务器获取污水检查井的历史液位时间序列的实际值(如图中所示的原始历史液位时间序列)、污水检查井所在区域的历史雨量时间序列(如图中所示的原始历史雨量时间序列)、以及预报未来雨量时间序列，对这些数据进行预处理，具体包括：对历史液位时间序列、历史雨量液位时间序列、预报雨量时间序列进行清洗和归一化处理，形成多维输入向量。将多维输入向量输入深度学习模型进行预测，得到未来T小时的预测结果，在经过一段时间之后，得到未来T小时液位时间序列的实际值，将预测结果与实际结果进行比较，计算预测结果与实际结果的误差，判断预测结果与实际结果的误差是否超过预设阈值，如果误差不超过预设阈值，则继续采用当前的深度学习模型进行预测；如果误差超过预设阈值，说明预测结果偏差过大，则更新训练样本集，将实测结果回归到训练样本集中，将其它未训练的历史实测数据也加入训练样本集。进行自适应训练，将原来的深度神经网络模型的超参数、权重、预测值和实际值的绝对误差，更新为新的训练的初值，进行自适应学习训练，直到当前次训练的预测绝对误差小于等于上一次训练的预测绝对误差，则训练结束，更新深度学习模型；如果当前次训练的预测绝对误差大于上一次训练的预测绝对误差，则等待新的监测数据，更新训练样本，重新进行自适应训练。
本申请实施例，基于历史监测数据和气象预报降雨数据进行预警预测，避免高成本的获取排水管网基础数据以及水动力建模工作，根据预测误差，更新训练样本和模型参数，进行自适应学习，预测结果准确度高，预测成本低。
与本申请实施例提供的方法的应用场景以及方法相对应地，本申请实施例还提供一种污水检查井的液位预测装置。如图5所示，该污水检查井的液位预测装置可以包括：
获取模块501，用于获取与污水检查井的未来液位时间序列相关的参考因素数值，参考因素数值包括污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列；
预测模块502，用于在基于参考因素数值确定污水检查井存在溢流风险的情况下，基于参考因素数值，预测污水检查井的未来液位时间序列。
本申请实施例提供的污水检查井的液位预测装置，获取与污水检查井的未来液位时间序列相关的参考因素数值，参考因素数值包括污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列；在基于参考因素数值确定污水检查井存在溢流风险的情况下，基于参考因素数值，预测污水检查井的未来液位时间序列。本申请实施例，根据污水检查井的实际液位时间序列、污水检查井所在区域的实际降雨时间序列、以及污水检查井所在区域的预报降雨时间序列，预测污水检查井的未来液位时间序列，预测结果准确度高；不需要管网基础数据，预测成本低。
在一种可能的实现方式中，预测模块502，用于：
将参考因素数值输入基于注意力机制的深度学习模型，得到深度学习模型输出的污水检查井的未来液位时间序列的预测值，基于注意力机制的深度学习模型对训练样本的峰值赋予不同的权重。
在一种可能的实现方式中，基于注意力机制的深度学习模型是基于训练样本训练得到的，训练样本包括污水检查井内的样本历史液位时间序列、污水检查井所在区域的样本历史降雨时间序列。
在一种可能的实现方式中，污水检查井的液位预测装置还包括更新模块，用于：
获取污水检查井的未来液位时间序列的实际值；
在未来液位时间序列的预测值与实际值的差值超过预设阈值的情况下，对训练样本进行更新，并利用更新后的训练样本训练深度学习模型，得到更新后的深度学习模型。
在一种可能的实现方式中，污水检查井的液位预测装置还包括第一误差比较模块，用于：
对于更新后的深度学习模型的每一次预测，确定当前次预测深度学习模型输出的预测值与实际值的第一绝对误差，以及前一次预测深度学习模型输出的预测值与实际值的第二绝对误差；
在第一绝对误差小于或等于第二绝对误差的情况下，将当前的深度学习模型确定为预测污水检查井的未来液位时间序列的模型。
在一种可能的实现方式中，污水检查井的液位预测装置还包括第二误差比较模块，用于：
对于更新后的深度学习模型的每一次预测，确定当前次预测深度学习模型输出的预测值与实际值的第一绝对误差，以及前一次预测深度学习模型输出的预测值与实际值的第二绝对误差；
在第一绝对误差大于第二绝对误差的情况下，对当前的深度学习模型的训练样本进行更新，并利用更新后的训练样本训练深度学习模型。
在一种可能的实现方式中，污水检查井存在溢流风险是通过以下方式确定的：
将参考因素数值分别与溢流预警阈值进行比较，若参考因素数值中的至少一项超过溢流预警阈值，则确定污水检查井存在溢流风险。
本申请实施例各装置中的各模块的功能可以参见上述方法中的对应描述，并具备相应的有益效果，在此不再赘述。
图6为用来实现本申请实施例的电子设备的框图。如图6所示，该电子设备包括：存储器610和处理器620，存储器610内存储有可在处理器620上运行的计算机程序。处理器620执行该计算机程序时实现上述实施例中的方法。存储器610和处理器620的数量可以为一个或多个。
该电子设备还包括：
通信接口630，用于与外界设备进行通信，进行数据交互传输。
如果存储器610、处理器620和通信接口630独立实现，则存储器610、处理器620和通信接口630可以通过总线相互连接并完成相互间的通信。该总线可以是工业标准体系结构(Industry Standard Architecture，ISA)总线、外部设备互连(Peripheral  Component Interconnect，PCI)总线或扩展工业标准体系结构(Extended Industry Standard Architecture，EISA)总线等。该总线可以分为地址总线、数据总线、控制总线等。为便于表示，图6中仅用一条粗线表示，但并不表示仅有一根总线或一种类型的总线。
可选的，在具体实现上，如果存储器610、处理器620及通信接口630集成在一块芯片上，则存储器610、处理器620及通信接口630可以通过内部接口完成相互间的通信。
本申请实施例提供了一种计算机可读存储介质，其存储有计算机程序，该程序被处理器执行时实现本申请实施例中提供的方法。
本申请实施例还提供了一种芯片，该芯片包括，包括处理器，用于从存储器中调用并运行存储器中存储的指令，使得安装有芯片的通信设备执行本申请实施例提供的方法。
本申请实施例还提供了一种芯片，包括：输入接口、输出接口、处理器和存储器，输入接口、输出接口、处理器以及存储器之间通过内部连接通路相连，处理器用于执行存储器中的代码，当代码被执行时，处理器用于执行申请实施例提供的方法。
应理解的是，上述处理器可以是中央处理器(Central Processing Unit，CPU)，还可以是其他通用处理器、数字信号处理器(Digital Signal Processor，DSP)、专用集成电路(Application Specific Integrated Circuit，ASIC)、现场可编程门阵列(Field Programmable Gate Array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。通用处理器可以是微处理器或者是任何常规的处理器等。值得说明的是，处理器可以是支持进阶精简指令集机器(Advanced RISC Machines，ARM)架构的处理器。
进一步地，可选的，上述存储器可以包括只读存储器和随机存取存储器，还可以包括非易失性随机存取存储器。该存储器可以是易失性存储器或非易失性存储器，或可包括易失性和非易失性存储器两者。其中，非易失性存储器可以包括只读存储器(Read-Only Memory，ROM)、可编程只读存储器(Programmable ROM，PROM)、可擦除可编程只读存储器(Erasable PROM，EPROM)、电可擦除可编程只读存储器(Electrically EPROM，EEPROM)或闪存。易失性存储器可以包括随机存取存储器(Random Access Memory，RAM)，其用作外部高速缓存。通过示例性但不是限制性说明，许多形式的RAM可用。例如，静态随机存取存储器(Static RAM，SRAM)、动态随机存取存储器(Dynamic Random Access Memory，DRAM)、同步动态随机存取存储器(Synchronous DRAM，SDRAM)、双倍数据速率同步动态随机存取存储器(Double Data Rate SDRAM，DDR SDRAM)、增强型同步动态随机存取存储器(Enhanced SDRAM，ESDRAM)、同步连接动态随机存取存储器(Synchlink DRAM，SLDRAM)和直接内存总线随机存取存储器(Direct Rambus RAM，DR RAM)。
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实 现。当使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行计算机程序指令时，全部或部分地产生按照本申请的流程或功能。计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输。
在本说明书的描述中，参考术语“一个实施例”、“一些实施例”、“示例”、“具体示例”、或“一些示例”等的描述意指结合该实施例或示例描述的具体特征、结构、材料或者特点包括于本申请的至少一个实施例或示例中。而且，描述的具体特征、结构、材料或者特点可以在任一个或多个实施例或示例中以合适的方式结合。此外，在不相互矛盾的情况下，本领域的技术人员可以将本说明书中描述的不同实施例或示例以及不同实施例或示例的特征进行结合和组合。
此外，术语“第一”、“第二”仅用于描述目的，而不能理解为指示或暗示相对重要性或者隐含指明所指示的技术特征的数量。由此，限定有“第一”、“第二”的特征可以明示或隐含地包括至少一个该特征。在本申请的描述中，“多个”的含义是两个或两个以上，除非另有明确具体的限定。
流程图中或在此以其他方式描述的任何过程或方法描述可以被理解为，表示包括一个或更多个用于实现特定逻辑功能或过程的步骤的可执行指令的代码的模块、片段或部分。并且本申请的优选实施方式的范围包括另外的实现，其中可以不按所示出或讨论的顺序，包括根据所涉及的功能按基本同时的方式或按相反的顺序，来执行功能。
在流程图中表示或在此以其他方式描述的逻辑和/或步骤，例如，可以被认为是用于实现逻辑功能的可执行指令的定序列表，可以具体实现在任何计算机可读介质中，以供指令执行系统、装置或设备(如基于计算机的系统、包括处理器的系统或其他可以从指令执行系统、装置或设备取指令并执行指令的系统)使用，或结合这些指令执行系统、装置或设备而使用。
应理解的是，本申请的各部分可以用硬件、软件、固件或它们的组合来实现。在上述实施方式中，多个步骤或方法可以用存储在存储器中且由合适的指令执行系统执行的软件或固件来实现。上述实施例方法的全部或部分步骤是可以通过程序来指令相关的硬件完成，该程序可以存储于一种计算机可读存储介质中，该程序在执行时，包括方法实施例的步骤之一或其组合。
此外，在本申请各个实施例中的各功能单元可以集成在一个处理模块中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个模块中。上述集成的模块既可以采用硬件的形式实现，也可以采用软件功能模块的形式实现。上述集成的模块如果以软件功能模块的形式实现并作为独立的产品销售或使用时，也可以存储在一个计算机可读存储介质中。该存储介质可以是只读存储器，磁盘或光盘等。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟 悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到其各种变化或替换，这些都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以权利要求的保护范围为准。

Claims (10)

  1. 一种污水检查井的液位预测方法，其特征在于，所述方法包括：
    获取与污水检查井的未来液位时间序列相关的参考因素数值，所述参考因素数值包括所述污水检查井的实际液位时间序列、所述污水检查井所在区域的实际降雨时间序列、以及所述污水检查井所在区域的预报降雨时间序列；
    在基于所述参考因素数值确定所述污水检查井存在溢流风险的情况下，基于所述参考因素数值，预测所述污水检查井的未来液位时间序列。
  2. 根据权利要求1所述的方法，其特征在于，所述基于所述参考因素数值，预测所述污水检查井的未来液位时间序列，包括：
    将所述参考因素数值输入基于注意力机制的深度学习模型，得到所述深度学习模型输出的所述污水检查井的未来液位时间序列的预测值，所述基于注意力机制的深度学习模型对训练样本的峰值赋予不同的权重。
  3. 根据权利要求2所述的方法，其特征在于，所述基于注意力机制的深度学习模型是基于所述训练样本训练得到的，所述训练样本包括所述污水检查井内的样本历史液位时间序列、所述污水检查井所在区域的样本历史降雨时间序列。
  4. 根据权利要求2或3所述的方法，其特征在于，所述方法还包括：
    获取所述污水检查井的未来液位时间序列的实际值；
    在所述未来液位时间序列的预测值与所述实际值的差值超过预设阈值的情况下，对所述训练样本进行更新，并利用更新后的训练样本训练所述深度学习模型，得到更新后的深度学习模型。
  5. 根据权利要求4所述的方法，其特征在于，所述方法还包括：
    对于所述更新后的深度学习模型的每一次预测，确定当前次预测深度学习模型输出的预测值与实际值的第一绝对误差，以及前一次预测深度学习模型输出的预测值与实际值的第二绝对误差；
    在所述第一绝对误差小于或等于所述第二绝对误差的情况下，将当前的深度学习模型确定为预测所述污水检查井的未来液位时间序列的模型。
  6. 根据权利要求4所述的方法，其特征在于，所述方法还包括：
    对于所述更新后的深度学习模型的每一次预测，确定当前次预测深度学习模型输出的预测值与实际值的第一绝对误差，以及前一次预测深度学习模型输出的预测值与实际值的第二绝对误差；
    在所述第一绝对误差大于所述第二绝对误差的情况下，对当前的深度学习模型的训练样本进行更新，并利用更新后的训练样本训练深度学习模型。
  7. 根据权利要求1-3任一项所述的方法，其特征在于，所述污水检查井存在溢流风险是通过以下方式确定的：
    将所述参考因素数值分别与溢流预警阈值进行比较，若所述参考因素数值中的至少一项超过所述溢流预警阈值，则确定所述污水检查井存在溢流风险。
  8. 一种污水检查井的液位预测装置，其特征在于，所述装置包括：
    获取模块，用于获取与污水检查井的未来液位时间序列相关的参考因素数值，所述参考因素数值包括所述污水检查井的实际液位时间序列、所述污水检查井所在区域的实际降雨时间序列、以及所述污水检查井所在区域的预报降雨时间序列；
    预测模块，用于在基于所述参考因素数值确定所述污水检查井存在溢流风险的情况下，基于所述参考因素数值，预测所述污水检查井的未来液位时间序列。
  9. 一种电子设备，包括存储器、处理器及存储在存储器上的计算机程序，所述处理器在执行所述计算机程序时实现权利要求1-7中任一项所述的方法。
  10. 一种计算机可读存储介质，所述计算机可读存储介质内存储有计算机程序，所述计算机程序被处理器执行时实现权利要求1-7中任一项所述的方法。
PCT/CN2023/099232 2022-06-14 2023-06-08 污水检查井的液位预测方法、装置、电子设备及存储介质 WO2023241461A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210674810.X 2022-06-14
CN202210674810.XA CN114897272A (zh) 2022-06-14 2022-06-14 污水检查井的液位预测方法、装置、电子设备及存储介质

Publications (1)

Publication Number Publication Date
WO2023241461A1 true WO2023241461A1 (zh) 2023-12-21

Family

ID=82727437

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/099232 WO2023241461A1 (zh) 2022-06-14 2023-06-08 污水检查井的液位预测方法、装置、电子设备及存储介质

Country Status (2)

Country Link
CN (1) CN114897272A (zh)
WO (1) WO2023241461A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114897272A (zh) * 2022-06-14 2022-08-12 阿里云计算有限公司 污水检查井的液位预测方法、装置、电子设备及存储介质
CN116504394B (zh) * 2023-06-21 2024-01-30 天津医科大学朱宪彝纪念医院(天津医科大学代谢病医院、天津代谢病防治中心) 基于多特征融合的辅助医疗方法、装置及计算机存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109978235A (zh) * 2019-03-04 2019-07-05 宁波市气象服务中心 一种基于样本学习的积涝水位预测方法
CN113221439A (zh) * 2021-04-07 2021-08-06 浙江大学 一种基于bp神经网络的排水系统实时率定与动态预测方法
KR20220057740A (ko) * 2020-10-30 2022-05-09 홍익대학교 산학협력단 순환신경망을 이용한 하천 지역 침수 예측 방법
CN114897272A (zh) * 2022-06-14 2022-08-12 阿里云计算有限公司 污水检查井的液位预测方法、装置、电子设备及存储介质

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109978235A (zh) * 2019-03-04 2019-07-05 宁波市气象服务中心 一种基于样本学习的积涝水位预测方法
KR20220057740A (ko) * 2020-10-30 2022-05-09 홍익대학교 산학협력단 순환신경망을 이용한 하천 지역 침수 예측 방법
CN113221439A (zh) * 2021-04-07 2021-08-06 浙江大学 一种基于bp神经网络的排水系统实时率定与动态预测方法
CN114897272A (zh) * 2022-06-14 2022-08-12 阿里云计算有限公司 污水检查井的液位预测方法、装置、电子设备及存储介质

Also Published As

Publication number Publication date
CN114897272A (zh) 2022-08-12

Similar Documents

Publication Publication Date Title
WO2023241461A1 (zh) 污水检查井的液位预测方法、装置、电子设备及存储介质
JP4807565B2 (ja) 流量予測装置
US10402738B2 (en) Bayesian nonparametric method for infrastructure failure prediction
Ruslan et al. Flood water level modeling and prediction using NARX neural network: Case study at Kelang river
JP6845126B2 (ja) 故障確率算出装置、故障確率算出方法及びプログラム
CN109886477B (zh) 水污染的预测方法、装置及电子设备
Lin et al. Integrative modeling of performance deterioration and maintenance effectiveness for infrastructure assets with missing condition data
CN104636874A (zh) 检测业务异常的方法及设备
CN113221439B (zh) 一种基于bp神经网络的排水系统实时率定与动态预测方法
JP2006092058A (ja) 流量予測装置
JP2022510667A (ja) エネルギー消費量を予測するための方法および装置
CN115643193A (zh) 一种网络流量异常检测方法、装置、设备及介质
JP6038011B2 (ja) ダム下流河川流量予測システム
Puig et al. Limnimeter and rain gauge FDI in sewer networks using an interval parity equations based detection approach and an enhanced isolation scheme
JP2001167078A (ja) ダムまたは河川における流量予測方法
Vezzaro et al. Evaluating the performance of a simple phenomenological model for online forecasting of ammonium concentrations at WWTP inlets
WO2024067824A1 (zh) 气缸劣化趋势确定方法及装置、电子设备、存储介质
JP6555391B2 (ja) 予測装置、予測システム、予測方法及びプログラム
CN114157486B (zh) 通信流量数据异常检测方法、装置、电子设备及存储介质
JPH11194803A (ja) ダム出水量予測モデル構築方法及びダム出水量予測方法
CN115267950A (zh) 一种初期雨水量化方法、装置、存储介质及电子设备
JP4836813B2 (ja) 雨水流入量予測装置及び雨水流入量予測方法
Qu et al. Performance prediction for steel bridges using SHM data and bayesian dynamic regression linear model: a novel approach
CN111310960A (zh) 一种能源消耗预测方法及装置
Zhou et al. Model Verification and Validation of a Cable-Stayed Bridge: Interval-Based Uncertainty Quantification of the Model Parameters

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23823029

Country of ref document: EP

Kind code of ref document: A1