Waterproofing and seepage early warning system driven by edge computing in complex temperature and humidity environments

Abstract

This invention relates to the field of pipeline safety monitoring and waterproofing / leakage early warning technology, and particularly to a waterproofing / leakage early warning system driven by edge computing in complex temperature and humidity environments. The system includes the following modules: an environmental and water immersion data acquisition module, used to collect and preprocess temperature, relative humidity, water immersion signals, and auxiliary evidence in real time at edge nodes; and a temperature and humidity coupled dew point determination module, used to calculate the dew point based on temperature and humidity data at edge nodes and generate negative condensation evidence. In this invention, by collecting and preprocessing temperature, humidity, and water immersion data at edge nodes, negative condensation evidence is generated based on the temperature and humidity data and fused with positive water immersion evidence to form a Markov chain. Simultaneously, an environmental baseline is established and thresholds are adaptively adjusted, thereby achieving continuous monitoring and closed-loop determination of the seepage status. This improves upon the problem that traditional pipeline seepage monitoring systems mostly rely on remote servers, leading to delayed judgments and false alarms due to complex environments and static thresholds.

Description

Technical Field

[0001] This invention relates to the field of pipeline safety monitoring and waterproofing / leakage early warning technology, and in particular to a waterproofing / leakage early warning system driven by edge computing in complex temperature and humidity environments. Background Technology

[0002] As urban infrastructure and industrial pipeline networks continue to expand, the safety and reliability of pipeline systems have become crucial. Manual inspections can lead to missed or incomplete checks. Failure to detect leaks or seepage not only wastes resources but can also cause equipment damage, environmental pollution, or safety accidents. Current pipeline monitoring methods, relying heavily on manual inspections or remote centralized monitoring, are insufficient to meet the real-time monitoring needs in complex environments.

[0003] Currently, most traditional pipeline seepage monitoring systems rely on remote servers, which lead to problems such as delayed judgment and false alarms due to complex environments and static thresholds. Summary of the Invention

[0004] To overcome the above shortcomings, this invention provides a waterproof seepage early warning system driven by edge computing in complex temperature and humidity environments. It aims to improve the problem that traditional pipeline seepage monitoring systems mostly rely on remote servers, and the judgment is delayed and false alarms are caused by the complexity of the environment and the static threshold.

[0005] In a first aspect, the present invention provides the following technical solution: a waterproof seepage early warning system driven by edge computing in complex temperature and humidity environments, comprising the following modules:

[0006] The environmental and water immersion data acquisition module is used to collect temperature, relative humidity, water immersion signals and auxiliary evidence in real time at the edge nodes and perform preprocessing.

[0007] The temperature and humidity coupled dew point determination module is used to calculate the dew point based on temperature and humidity data at the edge node and generate negative evidence of condensation.

[0008] The water immersion mutual verification algorithm module is used to extract positive evidence from water immersion signals and auxiliary evidence at edge nodes, and fuse it with condensation negative evidence to generate a Markov chain describing the water seepage state transition, which is used to reflect the water seepage state and provide the water seepage judgment module for judgment.

[0009] The edge adaptive threshold and environmental baseline update module is used to establish an environmental baseline at the edge node and dynamically adjust the Markov chain decision threshold.

[0010] The seepage detection and closed-loop control module is used to determine the seepage status at the edge node based on the Markov chain and threshold, and to trigger closed-loop control or early warning operation in the critical state.

[0011] The micro-heating retest enhancement module is used to micro-heat the sensing probe when determining the critical state, update the positive evidence and Markov chain, and then input it again into the seepage determination module for final determination.

[0012] By adopting the above technical solution, temperature, humidity and water immersion data are collected and preprocessed at the edge nodes. Based on the temperature and humidity data, negative evidence of condensation is generated and fused with positive evidence of water immersion to form a Markov chain. At the same time, an environmental baseline is established and the threshold is adaptively adjusted, thereby realizing continuous monitoring and closed-loop judgment of water seepage status. This improves the problem that traditional pipeline water seepage monitoring systems mostly rely on remote servers, and the judgment is delayed and false alarms are caused by complex environment and static thresholds.

[0013] Preferably, the data acquisition includes:

[0014] Temperature and humidity sensors and water immersion sensors are installed on the outer or inner wall of the pipe.

[0015] Real-time acquisition of temperature, relative humidity, water immersion signals, and auxiliary evidence;

[0016] The collected data is filtered and smoothed to form the input, providing basic data for subsequent judgment.

[0017] Preferably, the temperature and humidity coupled dew point determination includes:

[0018] Receive temperature and humidity data output from the environmental and water immersion data acquisition module;

[0019] Dew point is calculated based on a temperature and humidity coupling model, and negative evidence of condensation is generated based on the dew point determination results.

[0020] The condensed negative evidence is output to the water immersion mutual verification algorithm module.

[0021] Preferably, the water immersion mutual verification algorithm and Markov chain generation include:

[0022] Receive water immersion signals and supporting evidence output from the environmental and water immersion data acquisition module;

[0023] Based on the water immersion signal and supporting evidence, positive evidence of water immersion is extracted;

[0024] The negative evidence of condensation output from the temperature and humidity coupling dew point determination module is fused with the positive evidence.

[0025] Generate a Markov chain describing the seepage state transition and output it to the seepage determination module.

[0026] Preferably, the edge adaptive threshold and environmental baseline update include:

[0027] Receive environmental data and historical environmental data output from the environmental and water immersion data acquisition module, and establish an environmental baseline;

[0028] Dynamically adjust the Markov chain determination threshold generated by the water immersion mutual verification algorithm module;

[0029] The updated threshold is output to the seepage detection module.

[0030] Preferably, the seepage detection and closed-loop control include:

[0031] Receive the Markov chain and edge adaptive threshold output by the water immersion mutual verification algorithm module and the threshold output by the environmental baseline update module;

[0032] The current seepage status is determined based on the Markov chain and the threshold.

[0033] When the seepage condition reaches the threshold or is in a critical state, closed-loop control or early warning operation is triggered.

[0034] Preferably, the micro-heating retest enhancement includes:

[0035] When the seepage detection module determines the critical state, the surface of the sensing probe is slightly heated.

[0036] Temperature, humidity, and water immersion data were re-acquired after slight heating;

[0037] Update the positive evidence and Markov chain extracted by the water immersion mutual verification algorithm module;

[0038] The updated Markov chain is then input into the seepage detection module for final determination.

[0039] Preferably, the independent operation feature of the edge node includes: all modules run independently on the edge node, realizing closed-loop data processing, Markov chain update and water seepage detection, without relying on a remote server.

[0040] Preferably, the dynamic update of the Markov chain state includes:

[0041] The state transition probability is calculated based on the negative evidence of condensation output by the temperature and humidity coupling dew point determination module and the positive evidence extracted by the water immersion mutual verification algorithm module.

[0042] Dynamically update the indicators for determining pipeline seepage status by forming a Markov chain for continuous monitoring;

[0043] The judgment indicators are output to the seepage judgment module to realize closed-loop seepage early warning.

[0044] Secondly, this invention provides the following technical solution: a waterproofing and seepage early warning method driven by edge computing in complex temperature and humidity environments, comprising the following methods:

[0045] S1. Real-time acquisition and preprocessing of temperature, relative humidity, water immersion signals and auxiliary evidence at edge nodes;

[0046] S2. Calculate the dew point based on temperature and humidity data at the edge nodes and generate negative evidence of condensation.

[0047] S3. Extract positive evidence from water immersion signals and auxiliary evidence at edge nodes, and fuse it with condensation negative evidence to generate a Markov chain describing the transition of the seepage state, which is used to reflect the seepage state and provide the seepage determination module for determination.

[0048] S4. Establish an environmental baseline at the edge nodes and dynamically adjust the Markov chain decision threshold.

[0049] S5. Determine the seepage status at the edge node based on the Markov chain and threshold, and trigger closed-loop control or early warning operation in the critical state.

[0050] S6. When determining the critical state, the sensor probe is slightly heated to update the positive evidence and Markov chain, and then input into the seepage determination module again for final determination.

[0051] The present invention has the following beneficial effects:

[0052] 1. In this invention, temperature, humidity and water immersion data are collected and preprocessed at edge nodes. Based on the temperature and humidity data, negative evidence of condensation is generated and fused with positive evidence of water immersion to form a Markov chain. At the same time, an environmental baseline is established and the threshold is adaptively adjusted, thereby realizing continuous monitoring and closed-loop judgment of water seepage status. This improves the problem that traditional pipeline water seepage monitoring systems mostly rely on remote servers, and the judgment is delayed and false alarms are caused by complex environment and static threshold.

[0053] 2. In this invention, condensation negative evidence is generated by calculating the dew point based on temperature and humidity data, and then combined with water immersion signals to form a judgment basis. This improves the problem that traditional pipeline seepage monitoring systems mostly use single water immersion measurements, and the seepage judgment is inaccurate due to environmental condensation interference.

[0054] 3. In this invention, a Markov chain is generated by fusing positive evidence of water immersion and negative evidence of condensation at the edge nodes, thereby enabling continuous tracking and prediction of the seepage status. This improves the problem that traditional pipeline seepage monitoring systems mostly use static threshold judgments, which cannot reflect changes in the status and cause seepage to be missed.

[0055] 4. In this invention, by dynamically establishing an environmental baseline at the edge node and adaptively adjusting the Markov chain judgment threshold, and simultaneously initiating micro-heating retesting to update the judgment link in the critical state, a closed-loop seepage judgment is achieved. This improves the problem that traditional pipeline seepage monitoring systems mostly use fixed thresholds and single retests, which lead to unreliable seepage judgments due to environmental changes. Attached Figure Description

[0056] Figure 1This is a system architecture diagram of the edge computing-driven waterproofing and seepage early warning system for complex temperature and humidity environments proposed in this invention.

[0057] Figure 2 This is a flowchart of the data acquisition and preprocessing process of the edge computing-driven waterproofing and seepage early warning system for complex temperature and humidity environments proposed in this invention.

[0058] Figure 3 This is a flowchart of the temperature and humidity coupling determination process of the waterproof seepage early warning system driven by edge computing in complex temperature and humidity environments proposed in this invention.

[0059] Figure 4 The flowchart of evidence fusion and Markov chain generation for the edge computing-driven waterproof seepage early warning system in complex temperature and humidity environments proposed in this invention is shown.

[0060] Figure 5 The flowchart shows the micro-heating retest enhancement mechanism of the edge computing-driven waterproof seepage early warning system for complex temperature and humidity environments proposed in this invention.

[0061] Figure 6 This is a flowchart of the waterproofing and seepage early warning method driven by edge computing in complex temperature and humidity environments proposed in this invention. Detailed Implementation

[0062] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] Example 1:

[0064] In a first embodiment of the present invention, the present invention provides a waterproof seepage early warning system driven by edge computing in complex temperature and humidity environments, such as... Figures 1-5 As shown, it includes the following modules:

[0065] The environmental and water immersion data acquisition module is used to collect temperature, relative humidity, water immersion signals and auxiliary evidence in real time at the edge nodes and perform preprocessing.

[0066] Furthermore, data collection includes:

[0067] Temperature and humidity sensors and water immersion sensors are installed on the outer or inner wall of the pipe.

[0068] Real-time acquisition of temperature, relative humidity, water immersion signals, and auxiliary evidence;

[0069] The collected data is filtered and smoothed to form the input, providing basic data for subsequent judgment.

[0070] Specifically, the environmental and water immersion data acquisition module is used to acquire raw data related to the surrounding environment and water seepage of the pipeline in real time at the edge node, and to preprocess the data to ensure that the subsequent judgment algorithm can receive stable and reliable input.

[0071] Sensors are deployed by fixing temperature sensors, humidity sensors, and water immersion sensors on the outer or inner wall of the pipe to acquire ambient temperature T(t), relative humidity RH(t), and water immersion signal w(t), respectively. In addition, auxiliary environmental sensors can be configured to output supplementary evidence A(t).

[0072] Real-time data acquisition: During operation, edge nodes synchronously acquire the outputs of the aforementioned sensors at a fixed sampling period ∆t, forming a raw data set. ; where t represents the sampling time.

[0073] Data filtering and smoothing: The raw acquired data contains high-frequency noise and instantaneous abnormal fluctuations, therefore preprocessing is required. Filtering: Using the moving average method, with the sampling window length set to n, the filtered data is as follows: Where x(t) represents any type of sensor data, x f (t) represents the filtered result. Smoothing: A first-order exponential smoothing method is used. Let the smoothing coefficient be α∈(0,1). Then the smoothed data is: ;where x s (t) represents the smoothed time series result, which effectively reduces the impact of abrupt changes on subsequent judgments. After filtering and smoothing, the input data is formed: ; where D stable (t) represents the stable data set formed at time t after filtering and smoothing.

[0074] By deploying temperature, humidity, and water immersion sensors on the pipe surface, comprehensive monitoring of the environment and seepage status can be achieved; real-time acquisition ensures data timeliness, enabling the system to promptly detect seepage trends; filtering and smoothing remove instantaneous noise and abnormal fluctuations, ensuring the stability and accuracy of the input data; the processed D... stable (t) provides a reliable basis for subsequent modules, effectively improving the credibility of subsequent condensation determination and Markov chain determination.

[0075] The temperature and humidity coupled dew point determination module is used to calculate the dew point based on temperature and humidity data at the edge node and generate negative evidence of condensation.

[0076] Furthermore, the temperature and humidity coupled dew point determination includes:

[0077] Receive temperature and humidity data output from the environmental and water immersion data acquisition module;

[0078] Dew point is calculated based on a temperature and humidity coupling model, and negative evidence of condensation is generated based on the dew point determination results.

[0079] The condensed negative evidence is output to the water immersion mutual verification algorithm module.

[0080] Specifically, the temperature and humidity coupled dew point determination module is used to calculate the dew point temperature based on temperature and relative humidity data at the edge node and generate negative evidence of condensation. The negative evidence of condensation is used to characterize the possibility of condensation caused by temperature and humidity conditions in the environment, thereby avoiding misjudging condensation as pipe leakage.

[0081] Receive temperature and humidity data, and obtain a stable set of temperature and humidity data from the environmental and water immersion data acquisition module: ; where T s (t) represents the stable temperature data at time t; RH s (t) represents the stable relative humidity data at time t.

[0082] Temperature and humidity coupled dew point calculation, based on empirical formula to calculate dew point temperature T d (t): ;in ;T d (t) represents the dew point temperature at time t; a and b are constant parameters, usually a=17.27, b=237.7; This is a temperature and humidity coupling function, representing the combined effect of temperature and humidity conditions on the dew point.

[0083] Negative evidence for condensation is generated by calculating the dew point temperature T. d (t) and ambient temperature T s (t) Comparison: ; where E cond (t) represents the negative evidence of condensation at time t; E cond (t)=1 indicates the possibility of condensation; E cond (t)=0 indicates that there is no possibility of condensation.

[0084] Data output: The condensed negative evidence is output to the water immersion mutual evidence algorithm module as input conditions for subsequent fusion judgment.

[0085] The dew point temperature T is calculated based on temperature and humidity data. d (t), which realizes the quantitative determination of the effect of temperature and humidity conditions on condensation phenomena; negative evidence of condensation E cond (t) It can effectively distinguish between condensation and seepage in pipeline monitoring, avoiding misjudgment; the output data structure is as follows: As the input to the water immersion mutual verification algorithm module, it forms a mutual verification relationship with the water immersion signal and auxiliary evidence, improving the accuracy of the overall judgment; in terms of process, it realizes a closed loop of receiving temperature and humidity data, calculating dew point, generating negative evidence of condensation, and outputting it for mutual verification.

[0086] The water immersion mutual verification algorithm module is used to extract positive evidence from water immersion signals and auxiliary evidence at edge nodes, and fuse it with condensation negative evidence to generate a Markov chain describing the water seepage state transition, which is used to reflect the water seepage state and provide the water seepage judgment module for judgment.

[0087] Furthermore, the water immersion mutual verification algorithm and Markov chain generation include:

[0088] Receive water immersion signals and supporting evidence output from the environmental and water immersion data acquisition module;

[0089] Extracting positive evidence of water immersion based on water immersion signals and supporting evidence;

[0090] The negative and positive evidence of condensation output from the temperature and humidity coupling dew point determination module are fused together.

[0091] Generate a Markov chain describing the seepage state transition and output it to the seepage determination module.

[0092] Specifically, the water immersion mutual verification algorithm module is used to receive water immersion signals and auxiliary evidence at the edge node, extract positive water immersion evidence, and fuse it with the negative condensation evidence output by the temperature and humidity coupling dew point determination module to generate a Markov chain model describing the water seepage state transition, which is then output to the water seepage determination module to realize the probabilistic description of the water seepage state.

[0093] Receive input data, including water immersion signals and supporting evidence from the environmental and water immersion data acquisition module: ;W s (t) represents the water immersion signal at time t; A s (t) represents auxiliary evidence, such as optical reflection or pressure information. Simultaneously, it receives negative condensation evidence output from the temperature and humidity coupled dew point determination module. .

[0094] Extracting positive evidence of water immersion, and forming positive evidence based on the strength of the water immersion signal and supporting evidence:

[0095] ; where E water (t) represents positive evidence of water immersion at time t; f(·) is the evidence fusion function, such as a threshold determination or weighting function. When E water When (t)=1, it indicates the presence of signs of flooding; when E water When (t)=0, it indicates that there is no sign of water immersion.

[0096] The fusion of positive and negative evidence involves cross-referencing and integrating positive evidence of water immersion with negative evidence of condensation.

[0097] ; where E fusion (t) represents the fusion evidence at time t; when Econd When (t)=1, it indicates a high probability of condensation, thus suppressing positive evidence of water immersion; when E cond When (t)=0, retain positive evidence of water immersion.

[0098] Markov chain modeling: Establishing a Markov chain describing the seepage state transition based on fused evidence sequences. ;where S t ∈ {0, 1}, representing the seepage state, where 0 indicates no seepage and 1 indicates seepage; P ij This represents the probability of transitioning from state i to state j; the transition probability is obtained from the statistics of the fused evidence sequence. Output the Markov chain state transition matrix: .

[0099] Data output will fuse evidence E fusion The outputs of (t) and the Markov chain transition matrix M are sent to the seepage determination module as input conditions for determining the seepage state.

[0100] E extracted from water immersion signals and supporting evidence water (t) effectively quantifies signs of water immersion; compared with negative evidence of condensation E cond The fusion of (t) avoids misjudgments of water immersion caused by condensation; the Markov chain constructed based on the fused evidence can characterize the transition law of seepage state in the time dimension, forming a probabilistic description of the state; the output data structure is as follows: The output serves as the input to the water seepage determination module, forming a closed-loop process of extracting positive evidence of water immersion, fusing negative evidence, modeling state transitions, and outputting to the determination.

[0101] The edge adaptive threshold and environmental baseline update module is used to establish an environmental baseline at the edge node and dynamically adjust the Markov chain decision threshold.

[0102] Furthermore, edge-adaptive thresholding and environmental baseline updates include:

[0103] Receive environmental data and historical environmental data output from the environmental and water immersion data acquisition module, and establish an environmental baseline;

[0104] Dynamically adjust the Markov chain determination threshold generated by the water immersion mutual verification algorithm module;

[0105] The updated threshold is output to the seepage detection module.

[0106] Specifically, the edge adaptive threshold and environmental baseline update module is used to receive environmental data and historical environmental data output by the environmental and water immersion data acquisition modules at the edge nodes, establish an environmental baseline, and dynamically adjust the Markov chain judgment threshold generated by the water immersion mutual verification algorithm module based on the environmental baseline. The updated threshold is then output to the seepage judgment module to adapt to the judgment requirements under different environmental conditions.

[0107] Receive input data, including real-time environmental data from the environmental and water immersion data acquisition module: Where T(t) represents temperature data; H(t) represents humidity data; and L(t) represents environmental auxiliary information such as light intensity or pressure. Historical environmental data is received. .

[0108] Establish an environmental baseline by calculating the environmental baseline based on historical environmental data: ;

[0109] Among them B env This represents the environmental baseline, reflecting the long-term average environmental level; mean(·) is the statistical average operation. To increase stability, a sliding window update can be used: ; where α∈(0,1) are the update coefficients; this formula achieves adaptive correction of the baseline by the new input data.

[0110] The threshold is dynamically adjusted, and the environmental baseline is used to correct the Markov chain decision threshold generated by the water immersion mutual verification algorithm. Where θ0 represents the initial judgment threshold; , representing the deviation of the current environment from the baseline; β is the threshold correction coefficient, used to control the impact of the deviation on the threshold. The adjustment process ensures that the threshold can adaptively change with environmental fluctuations, thereby improving the reliability of the judgment.

[0111] The updated threshold is output and the updated judgment threshold θ(t) is output to the seepage judgment module as the input condition for the seepage state judgment.

[0112] enter: Output: .

[0113] The process is as follows: real-time environmental data and historical data → environmental baseline establishment → comparison with current data to obtain deviation → dynamic adjustment of threshold → output of updated threshold.

[0114] By introducing environmental baseline B env It can form a stable reference for long-term environmental conditions; dynamically adjust the threshold θ(t) to compensate for the risk of misjudgment caused by environmental fluctuations; the output adaptive threshold can be directly used as the input of the seepage judgment module to ensure the robustness of the judgment logic; a closed-loop process from environmental data input, baseline construction, threshold correction and threshold output has been established.

[0115] The seepage detection and closed-loop control module is used to determine the seepage status at the edge node based on the Markov chain and threshold, and to trigger closed-loop control or early warning operation in the critical state.

[0116] Furthermore, seepage detection and closed-loop control include:

[0117] Receive the Markov chain and edge adaptive threshold output by the water immersion mutual verification algorithm module and the threshold output by the environmental baseline update module;

[0118] The current seepage status is determined based on Markov chains and thresholds;

[0119] When the seepage reaches a threshold or is in a critical state, closed-loop control or early warning operation is triggered.

[0120] Specifically, the seepage detection and closed-loop control module is used to receive the Markov chain output by the water immersion mutual verification algorithm module and the threshold output by the edge adaptive threshold and environmental baseline update module at the edge node. Based on the input, it determines the current seepage status and triggers closed-loop control or early warning operation when the seepage status reaches the set threshold or enters the critical state, thereby realizing real-time prevention and control of seepage risk.

[0121] Receive input data and the Markov chain output by the water immersion mutual verification algorithm module: M = {P} ij}; where P ij This represents the probability of transitioning from state i to state j; the set of states can be represented as S = {S0, S1, S2}, where S0 represents the no-seepage state, S1 represents the potential-seepage state, and S2 represents the confirmed-seepage state. The threshold θ(t) is the output of the edge adaptive threshold and the environmental baseline update module.

[0122] Calculate the probability of the current seepage state. Based on the Markov chain and state transition probabilities, calculate the probability distribution of the system being in a certain state: ;in , represent the probabilities of the system being in state S0, S1, and S2 at time t, respectively; π(t+1) is the probability distribution of the state at the next time step.

[0123] The seepage status is determined by comparing the probability of seepage status with a threshold: S(t) represents the determination of the seepage state at time t; when the probability of a certain state exceeds the determination threshold θ(t), the system is determined to be in that state.

[0124] Trigger closed-loop control or early warning: When the judgment result is potential seepage state S1 or confirmed seepage state S2, trigger closed-loop operation: Where C(t) represents the control command, with a value of 0 indicating no trigger, a value of 1 indicating a warning operation, and a value of 2 indicating the execution of closed-loop control action.

[0125] enter: Processing flow: Markov chain and threshold input → state probability calculation → comparison with threshold → seepage state determination → output control or warning command; Output: .

[0126] By receiving the Markov chain generated by the water immersion mutual verification algorithm, the dynamic evolution of the seepage state can be reflected at the probabilistic level; combined with the edge adaptive threshold, the dynamic adaptation of the seepage judgment threshold under different environments is realized; the formulaic judgment of the seepage state ensures the quantifiability and interpretability of the seepage state; the closed-loop control or early warning is triggered at the critical state, realizing real-time risk prevention and control, and ensuring that the system has adaptive and proactive intervention capabilities; the whole forms a closed-loop process from Markov chain input, threshold correction, state judgment, and closed-loop control output.

[0127] The micro-heating retest enhancement module is used to micro-heat the sensing probe when determining the critical state, update the positive evidence and Markov chain, and then input it again into the seepage determination module for final determination.

[0128] Furthermore, the micro-heating retest enhancement includes:

[0129] When the seepage detection module determines the critical state, the surface of the sensing probe is slightly heated.

[0130] Temperature, humidity, and water immersion data were re-acquired after slight heating;

[0131] Update the positive evidence and Markov chain extracted by the water immersion mutual verification algorithm module;

[0132] The updated Markov chain is then input into the seepage detection module for final determination.

[0133] Specifically, the micro-heating retest enhancement module is activated when the water seepage determination module determines that the state is critical. It performs micro-heating on the surface of the sensing probe, and after micro-heating, it re-collects temperature, humidity and water immersion data. Based on the retest data, it updates the positive evidence and Markov chain, and then inputs the updated results into the water seepage determination module for final determination.

[0134] The main inputs to the module are: Where S(t) represents the current seepage state, π(t) is the probability distribution of each state, θ(t) is the threshold, and M(t) is a Markov chain. When the condition S(t) = S1 is satisfied and When the time is reached, it is considered to be in a critical state. In the formula, S1 represents the potential seepage state, π1(t) represents the probability of this state, and ε is the allowable approach range.

[0135] The module execution process includes: micro-heating, where the heating unit controls the temperature rise of the probe surface.

[0136] ; where T s (t) represents the probe surface temperature before heating, and ΔT is the set temperature increase value. Increasing the temperature can accelerate the evaporation of moisture on the probe surface.

[0137] Re-collection: After the probe has stabilized from heating, environmental parameters are collected again. Where T'(t) is the temperature data, H'(t) is the humidity data, and W'(t) is the water immersion signal.

[0138] Positive evidence update: Calculate new positive evidence using retest data. In the formula, E + (t+1) represents the updated positive evidence, and f(·) is the evidence extraction function.

[0139] Markov chain correction: Correcting Markov chains based on new evidence. In the formula, M'(t+1) is the updated transition matrix, and g(·) is the correction rule function.

[0140] Finally, the corrected Markov chain is re-input into the seepage detection module: The final judgment state is as follows: ; where h(·) represents the decision function, which is used to compare the probability distribution with the threshold and output the conclusion of whether there is seepage or no seepage.

[0141] The data flow process is as follows: the input is the water seepage judgment result, probability distribution, threshold, and initial Markov chain; after micro-heating, the temperature, humidity, and water immersion data are retested and new data are obtained; positive evidence is extracted and the Markov chain is updated; the updated Markov chain is input into the judgment module again to obtain the final judgment result.

[0142] When there is uncertainty in the initial judgment, the interference caused by factors such as condensation is eliminated by retesting the data, so that the final judgment result is closer to the actual seepage situation and the probability of misjudgment is reduced.

[0143] The independent operation of edge nodes includes the following features: all modules run independently on the local edge node, realizing closed-loop data processing, Markov chain updates and water leakage detection, without relying on remote servers.

[0144] Specifically, the independent operation of edge nodes means that all modules in the system run locally on the edge nodes, completing data acquisition, temperature and humidity coupling dew point determination, water immersion mutual verification algorithm processing, Markov chain update, threshold adaptive adjustment, and seepage determination closed-loop operation without relying on remote servers or external computing resources.

[0145] Data acquisition and preprocessing: edge nodes locally acquire temperature and humidity data and water immersion signals. Where T(t) is temperature, H(t) is humidity, W(t) is water immersion signal, and A(t) is auxiliary evidence. The collected data is filtered and smoothed to form the input: In the formula, filter(·) represents the smoothing and denoising processing function.

[0146] Temperature and humidity coupled dew point determination: calculate the dew point locally and generate negative evidence of condensation. ; where E cond (t) represents the negative evidence of condensation, and f(·) represents the dew point determination function based on the temperature and humidity coupling model.

[0147] Water immersion mutual verification and Markov chain update, receiving positive evidence E water (t) and negative evidence of condensation E cond (t) Fusion to generate Markov chains: ; where M(t) is the matrix describing the probability of seepage state transition, and g(·) is the fusion function.

[0148] Threshold adaptive adjustment, establishing environmental baseline B based on historical environmental data env Calculate the threshold: Where θ0 is the initial threshold and β is the adjustment coefficient.

[0149] The seepage detection loop uses a local Markov chain and a threshold to determine the seepage status. ; where π(t+1) is the state probability distribution, S(t+1) is the seepage determination result, and h(·) is the determination function.

[0150] Micro-heating retest enhancement: Under critical conditions, micro-heating of the probe is performed locally to collect retest data and update positive evidence and Markov chains. The Markov chain, after retesting, is input into the local seepage detection module again.

[0151] enter: Output: The process is as follows: Data collection → Local preprocessing → Local calculation of dew point and condensation negative evidence → Fusion of positive and negative evidence → Markov chain update → Threshold adaptive adjustment → Local seepage judgment → Critical state triggering retest → Output judgment result and update chain.

[0152] Running independently on edge nodes ensures that the system can complete closed-loop determination without relying on remote servers. All data is processed locally; the calculation of seepage probability, threshold adjustment, and final determination form a closed loop, reducing the impact of network latency and communication failures on determination accuracy. Local Markov chain updates and retesting enhance the timeliness and reliability of the determination results.

[0153] Markov chain state dynamic updates include:

[0154] The state transition probability is calculated based on the negative evidence of condensation output by the temperature and humidity coupling dew point determination module and the positive evidence extracted by the water immersion mutual verification algorithm module.

[0155] Dynamically update the indicators for determining pipeline seepage status by forming a Markov chain for continuous monitoring;

[0156] The judgment indicators are output to the seepage judgment module to realize closed-loop seepage early warning.

[0157] Specifically, the Markov chain state dynamic update is used to calculate the state transition probability based on the negative evidence of condensation output by the temperature and humidity coupling dew point determination module and the positive evidence extracted by the water immersion mutual verification algorithm module. It dynamically generates judgment indicators for continuously monitoring pipeline seepage status and outputs the judgment indicators to the seepage determination module to realize closed-loop seepage early warning.

[0158] Receive input data; module input includes ; where E cond (t) represents the negative evidence of condensation generated by the temperature and humidity coupled dew point determination; E water M(t) represents the positive evidence extracted by the water immersion mutual verification algorithm; M(t) represents the Markov chain state transition matrix of the previous time step.

[0159] State transition probability calculation, updating the state transition probability based on input evidence: ;where P ij (t+1) represents the probability of transitioning from state i to state j; f ij (·) represents the effect function of fusing positive and negative evidence on state transition; the denominator is the normalization coefficient, which ensures that the sum of all state transition probabilities is 1.

[0160] Markov chain update: Update the Markov chain with the calculated state transition probabilities. Generate continuously monitored seepage assessment indicators to describe the probability distribution of pipeline seepage status over time: ; where π(t+1) is the updated state probability distribution vector.

[0161] Output the decision criteria, and output the updated Markov chain and state probability vector to the seepage decision module: The seepage detection module determines the threshold based on this indicator and outputs the seepage status or triggers a closed-loop operation.

[0162] Input: Negative evidence of condensation E cond (t), positive evidence of water immersion E water(t), Markov chain M(t) at the previous time step; Processing: Calculate state transition probability → Update Markov chain → Calculate state probability distribution → Output judgment index; Output: Updated Markov chain M(t+1) and state probability distribution π(t+1).

[0163] The dynamically updated Markov chain can continuously reflect the pipeline seepage status locally, and the state probability is corrected in real time by combining negative evidence of temperature and humidity and positive evidence of water immersion. The output judgment index is provided to the seepage judgment module to realize closed-loop judgment and early warning functions. This method avoids the randomness of single judgments, makes the seepage monitoring process continuous and controllable, and improves the reliability of judgment and the ability to adapt to complex temperature and humidity environments.

[0164] Example 2:

[0165] In urban underground water supply network monitoring scenarios, pipelines are located in areas with high humidity, seasonal water level fluctuations, and rainwater infiltration, resulting in complex and fluctuating temperature and humidity environments. The system deploys temperature and humidity sensors, water immersion sensors, and auxiliary environmental sensors on the outer and part of the inner walls of the pipeline. Each edge node operates independently, enabling local data acquisition, temperature and humidity coupled dew point determination, water immersion mutual verification algorithms, Markov chain dynamic updates, threshold adaptive adjustment, and seepage closed-loop determination. The high humidity environment around the pipeline easily generates condensation interference, causing traditional water immersion sensors to misjudge or miss seepage. Pipeline seepage may be slow or localized, making continuous and reliable determination difficult with single measurements. Static thresholds are difficult to adapt to environmental changes, leading to inaccurate warnings. Data processing relies on remote servers, resulting in latency and reliability issues. The lack of retesting methods under critical conditions makes it impossible to verify the reliability of the judgment results. To solve these problems, this invention provides a waterproof seepage early warning method driven by edge computing in complex temperature and humidity environments, the structure of which is as follows... Figure 6 As shown. The specific implementation process of this method is as follows:

[0166] S1. Real-time acquisition and preprocessing of temperature, relative humidity, water immersion signals and auxiliary evidence at edge nodes;

[0167] S2. Calculate the dew point based on temperature and humidity data at the edge nodes and generate negative evidence of condensation.

[0168] S3. Extract positive evidence from water immersion signals and auxiliary evidence at edge nodes, and fuse it with condensation negative evidence to generate a Markov chain describing the transition of the seepage state, which is used to reflect the seepage state and provide the seepage determination module for determination.

[0169] S4. Establish an environmental baseline at the edge nodes and dynamically adjust the Markov chain decision threshold.

[0170] S5. Determine the seepage status at the edge node based on the Markov chain and threshold, and trigger closed-loop control or early warning operation in the critical state.

[0171] S6. When determining the critical state, the sensor probe is slightly heated to update the positive evidence and Markov chain, and then input into the seepage determination module again for final determination.

[0172] Specifically, temperature, relative humidity, water immersion signals, and supporting evidence are collected in real time at edge nodes to form raw environmental data. The collected data is filtered and smoothed to form input, providing the foundational data for subsequent temperature and humidity coupled dew point determination and water immersion mutual verification algorithms. The output data is used in all subsequent modules to ensure the accuracy and reliability of the basic information for water seepage determination.

[0173] The system receives temperature and humidity data output from S1, calculates the dew point locally at the edge node, and generates negative condensation evidence. This negative condensation evidence describes the likelihood of condensation in the environment and provides negative evidence input to the water immersion mutual verification algorithm. The output negative evidence is directly passed to the water immersion mutual verification algorithm module, achieving data closure.

[0174] The system receives the water immersion signal and supporting evidence output from S1 and extracts positive evidence. This positive evidence is then fused with the negative condensation evidence from S2 to generate a Markov chain describing the transition of the seepage state, reflecting the changing patterns of pipeline seepage. The generated Markov chain serves as input to the seepage determination module, enabling continuous monitoring and assessment.

[0175] The system receives environmental and historical data output from S1 and establishes an environmental baseline at the edge nodes. Based on the environmental baseline, it dynamically adjusts the Markov chain decision threshold and outputs the updated threshold to the seepage decision module, providing an adaptive standard for decision-making under different environmental conditions.

[0176] The system receives the Markov chain output from S3 and the threshold output from S4 to determine the water seepage status at the edge node. When a critical state is reached, closed-loop control or an early warning operation is triggered, and the water seepage status information is output to achieve local closed-loop management at the edge node.

[0177] When S5 determines the critical state, the surface of the sensor probe is slightly heated. After the slight heating, temperature, humidity, and water immersion data are collected again. The positive evidence of water immersion and the Markov chain are updated. The updated Markov chain is then input into the water seepage determination module to achieve the final closed-loop determination and ensure the reliability of the determination result.

[0178] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A complex temperature and humidity environment edge computing driven waterproof seepage early warning system, characterized in that, The system comprises the following modules: An environment and water immersion data acquisition module for collecting temperature, relative humidity, water immersion signals and auxiliary evidence in real time at the edge node and performing preprocessing; A temperature and humidity coupling dew point determination module for calculating dew point based on temperature and humidity data at the edge node and generating condensation negative evidence; A water immersion mutual evidence algorithm module for extracting positive evidence from water immersion signals and auxiliary evidence at the edge node, and fusing the positive evidence with the condensation negative evidence to generate a Markov chain describing the transition of water seepage state, which is used to reflect the water seepage state and output to a water seepage determination and closed-loop control module for determination; An edge adaptive threshold and environment baseline update module for establishing an environment baseline at the edge node and dynamically adjusting the Markov chain determination threshold; A water seepage determination and closed-loop control module for determining the water seepage state at the edge node based on the Markov chain and the threshold, and triggering closed-loop control or warning operation when the water seepage state reaches the threshold or is in a critical state; A micro-heating re-measurement enhancement module for micro-heating the sensing probe when the critical state is determined, updating the positive evidence and the Markov chain, and inputting the updated Markov chain to the water seepage determination and closed-loop control module for final determination.

2. The complex temperature and humidity environment edge computing driven waterproof and water seepage early warning system according to claim 1, characterized in that, Data acquisition includes: Arranging temperature and humidity sensors, water immersion sensors and auxiliary environment sensors on the outer wall or inner wall of the pipeline; Collecting temperature, relative humidity, water immersion signals and auxiliary evidence in real time; Filtering and smoothing the collected data to form input, providing basic data for subsequent determination.

3. The complex temperature and humidity environment edge computing driven water seepage early warning system according to claim 1, characterized in that, Temperature and humidity coupling dew point determination includes: Receiving temperature and humidity data output by the environment and water immersion data acquisition module; Calculating dew point based on the temperature and humidity coupling model, and generating condensation negative evidence according to the dew point determination result; Outputting the condensation negative evidence to the water immersion mutual evidence algorithm module.

4. The complex temperature and humidity environment edge computing driven waterproof and water seepage early warning system according to claim 1, characterized in that, Water immersion mutual evidence algorithm and Markov chain generation includes: Receiving water immersion signals and auxiliary evidence output by the environment and water immersion data acquisition module; Extracting water immersion positive evidence based on the water immersion signals and auxiliary evidence; Fusing the condensation negative evidence output by the temperature and humidity coupling dew point determination module with the positive evidence; Generating a Markov chain describing the transition of water seepage state, and outputting to the water seepage determination and closed-loop control module for determination.

5. The complex temperature and humidity environment edge computing driven waterproof and water seepage early warning system according to claim 1, characterized in that, Edge adaptive threshold and environment baseline update includes: Receiving environment data output by the environment and water immersion data acquisition module and historical environment data, establishing an environment baseline; Dynamically adjusting the Markov chain determination threshold generated by the water immersion mutual evidence algorithm module; Outputting the updated threshold to the water seepage determination and closed-loop control module.

6. The complex temperature and humidity environment edge computing driven waterproof and water seepage early warning system according to claim 1, characterized in that, Water seepage determination and closed-loop control includes: Receiving the Markov chain output by the water immersion mutual evidence algorithm module and the threshold output by the edge adaptive threshold and environment baseline update module; Determining the current water seepage state based on the Markov chain and the threshold; Triggering closed-loop control or warning operation when the water seepage state reaches the threshold or is in a critical state.

7. The complex temperature and humidity environment edge computing driven waterproof and water seepage early warning system according to claim 1, characterized in that, Micro-heating re-measurement enhancement includes: Micro-heating the surface of the sensing probe when the critical state is determined by the water seepage determination and closed-loop control module; Recollecting temperature and humidity and water immersion data after micro-heating; Updating the positive evidence extracted by the water immersion mutual evidence algorithm module and the Markov chain; Inputting the updated Markov chain to the water seepage determination and closed-loop control module for final determination.

8. The complex temperature and humidity environment edge computing driven waterproof and water seepage early warning system according to claim 1, characterized in that, The edge node independent running characteristics include that all modules are independently run locally on the edge node, closed-loop data processing, Markov chain updating and water seepage determination are achieved, and a remote server is not needed to be relied on.

9. The complex temperature and humidity environment edge computing driven waterproof and water seepage early warning system according to claim 1, characterized in that, The Markov chain state dynamic updating includes: calculating state transition probability based on condensation negative evidence output by the temperature and humidity coupling dew point determination module and positive evidence extracted by the water immersion mutual evidence algorithm module; dynamically updating the Markov chain to form a determination index of the water seepage state of the continuous monitoring pipeline; outputting the determination index to the water seepage determination and closed-loop control module to achieve closed-loop water seepage early warning.

10. A waterproof seepage early warning method driven by complex temperature and humidity environment edge computing, characterized in that, The waterproof and water seepage early warning system driven by edge computing in a complex temperature and humidity environment according to any one of claims 1-9 comprises the following method: S1, collecting temperature, relative humidity, water immersion signal and auxiliary evidence in real time on the edge node and preprocessing; S2, calculating dew point based on temperature and humidity data on the edge node and generating condensation negative evidence; S3, extracting positive evidence from the water immersion signal and auxiliary evidence on the edge node, and fusing the positive evidence with the condensation negative evidence to generate a Markov chain describing water seepage state transition, which is used to reflect the water seepage state and output to the water seepage determination and closed-loop control module for determination; S4, establishing an environmental baseline on the edge node and dynamically adjusting the Markov chain determination threshold; S5, determining the water seepage state on the edge node according to the Markov chain and the threshold, and triggering closed-loop control or early warning operation in the critical state; S6, when the critical state is determined, slightly heating the sensing probe, updating the positive evidence and the Markov chain, and inputting them again to the water seepage determination and closed-loop control module for final determination.

Images