Comprehensive monitoring method and system for gas, liquid level and well lid movement in underground inspection well

Abstract

The invention relates to the technical field of underground inspection well comprehensive monitoring scheme design, in particular to a comprehensive monitoring method and system for gas, liquid level and well lid movement in an underground inspection well. Inspection well gas concentration, liquid level height and well lid displacement data are collected through a multi-sensor fusion technology, and data precision is improved through self-adaptive filtering and abnormal value elimination; gas component identification is realized based on a CNN-LSTM deep learning model, and liquid level pressure compensation is carried out by adopting a state space model; normal vibration and abnormal movement of the well lid are distinguished through a machine learning classifier, and the false alarm rate is reduced; data are transmitted through LoRaWAN low-power-consumption communication, and when signals are blocked, the signals are cached and forwarded by edge computing nodes; and finally, a multi-source data fusion model is constructed at a platform end, weighted scoring and graded early warning are carried out in combination with gas toxicity, liquid level changes and well lid states, and the safety monitoring capability of the underground inspection well is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of comprehensive monitoring scheme design for underground manholes, and particularly to a comprehensive monitoring method and system for gas, liquid level, and manhole cover movement in underground manholes. Background Art

[0002] Underground manholes are an important part of urban infrastructure and are widely used in systems such as electricity, communication, gas, and drainage. However, the environment inside the manholes is complex and dangerous, with potential safety hazards such as gas leakage, water accumulation, and illegal intrusion. Traditional manual inspections are difficult to meet the real-time monitoring requirements. The existing technologies have the following deficiencies: Low gas monitoring accuracy: A single gas sensor is vulnerable to cross-interference (such as the impact of volatile organic compounds on methane detection), and there is a lack of a dynamic compensation mechanism, resulting in false alarms or missed alarms.

[0003] Large liquid level measurement error: Ultrasonic liquid level gauges are significantly affected by temperature and air pressure changes. Existing systems do not consider environmental factor compensation, and the measurement error is usually large.

[0004] Poor reliability in manhole cover status identification: The vibration signals collected by inertial measurement units (IMUs) are similar to interference signals such as vehicle rolling and personnel movement. Traditional threshold methods are difficult to distinguish normal vibration from abnormal movement, resulting in a high false alarm rate.

[0005] Insufficient communication stability: The underground manhole environment is enclosed, and signal attenuation is severe. Existing wireless communication technologies (such as Wi-Fi, ZigBee) have a short transmission distance and a high data packet loss rate.

[0006] Single risk assessment mechanism: Traditional systems only trigger warnings based on a single parameter (such as gas concentration) and cannot comprehensively consider associated factors such as the liquid level rising rate and the manhole cover displacement amplitude, resulting in untimely warnings or over-alarming.

[0007] Therefore, there is an urgent need for a comprehensive monitoring method that integrates efficient data fusion, intelligent algorithms, and reliable communication to improve the safety management level of manholes.

[0008] Therefore, the existing technologies still need to be further developed. Summary of the Invention

[0009] The purpose of the present invention is to overcome the above technical deficiencies and provide a comprehensive monitoring method and system for gas, liquid level, and manhole cover movement in underground manholes to solve the problems existing in the prior art.

[0010] To achieve the above technical objectives, according to the first aspect of the present invention, the present invention provides a comprehensive monitoring method for gas, liquid level, and manhole cover movement in underground manholes, including: S1: Collect gas concentration, liquid level height, and manhole cover displacement data of the underground manhole through multi-sensor fusion technology; S2: Apply an adaptive filtering algorithm to suppress noise and remove outliers from the sensor data to improve data accuracy; S3: Identify gas components based on a deep learning model, distinguish target gases from interfering gases, and calculate the concentration of each component; S4: Use a pressure compensation algorithm to eliminate the influence of environmental temperature and pressure fluctuations in liquid level measurement and improve liquid level measurement accuracy; S5: Distinguish normal vibration from abnormal movement of the manhole cover through a machine learning classifier to reduce the false alarm rate; S6: Adopt a low-power wide-area communication technology to transmit the monitoring data to a remote monitoring platform, and enable edge computing nodes to cache and forward data when the signal is blocked; S7: Build a multi-source data fusion model at the monitoring platform end, and conduct comprehensive risk assessment and early warning by combining gas concentration, liquid level change, and manhole cover status.

[0011] Specifically, the multi-sensor fusion technology includes a gas sensor array, an ultrasonic liquid level gauge, and an inertial measurement unit, and data synchronization and joint estimation are achieved through Kalman filtering.

[0012] Specifically, the inertial measurement unit is fixed to the edge of the manhole cover by pasting or magnetic attraction, and is used to monitor the three-axis acceleration data and angular velocity data of the manhole cover.

[0013] Specifically, the deep learning model adopts a structure combining a convolutional neural network and a long short-term memory network. The input layer receives the original sensor signal, and the output layer is the gas component classification result and concentration value.

[0014] Specifically, the pressure compensation algorithm is based on the real-time data of the temperature and air pressure sensors in the manhole, and dynamically corrects the liquid level measurement value through a state space model.

[0015] Specifically, the machine learning classifier adopts a random forest algorithm, and the training data includes a normal vibration signal feature library and an abnormal movement event sample library.

[0016] Specifically, the edge computing node deploys a lightweight data processing algorithm, temporarily stores the monitoring data during communication interruption, and resumes transmission through a breakpoint resumption mechanism.

[0017] Specifically, the comprehensive risk assessment adopts a weighted scoring mechanism, generates a risk level by combining gas toxicity, liquid level rise rate, and manhole cover displacement amplitude, and triggers a hierarchical early warning.

[0018] Specifically, the weighted score is calculated using the following formula: Where, is the weighted score, is the concentration of toxic gas, is the liquid level rising rate, is the impact energy of the manhole cover, t is the event duration, and the denominator of the formula is the time decay factor, which is used to suppress short-term false alarms and trigger hierarchical early warnings according to the weighted score and the preset hierarchical early warning threshold list.

[0019] According to the second aspect of the present invention, there is provided a comprehensive monitoring system for gas, liquid level, and manhole cover movement in an underground manhole, including: An acquisition module for collecting gas concentration, liquid level height, and manhole cover displacement data in the underground manhole through multi-sensor fusion technology; A control module for suppressing noise and removing outliers from sensor data by using an adaptive filtering algorithm to improve data accuracy; for identifying gas components based on a deep learning model, distinguishing target gases from interfering gases, and calculating the concentration of each component; for eliminating the influence of environmental temperature and pressure fluctuations in liquid level measurement by using a pressure compensation algorithm to improve liquid level measurement accuracy; for distinguishing normal vibration of the manhole cover from abnormal movement by using a machine learning classifier to reduce the false alarm rate; for transmitting monitoring data to a remote monitoring platform by using a low-power wide-area communication technology and enabling an edge computing node to perform data caching and forwarding when the signal is blocked; for constructing a multi-source data fusion model at the monitoring platform end and conducting comprehensive risk assessment and early warning by combining gas concentration, liquid level change, and manhole cover status.

[0020] Beneficial effects: 1. Precise fusion of multi-source data: Multi-sensor fusion eliminates time deviation through Kalman filtering to achieve data synchronization, and the CNN-LSTM model effectively identifies gas components.

[0021] 2. Reduction of liquid level measurement error: The pressure compensation algorithm dynamically corrects the sound speed error and improves the stability of liquid level measurement.

[0022] 3. Decrease in manhole cover false alarm rate: The random forest classifier extracts time-frequency features, reduces the false alarm rate of manhole cover status monitoring, and supports micro-impact detection.

[0023] 4. Improvement of communication reliability: LoRaWAN communication ensures data transmission, and the edge computing node cache ensures that data is not lost and the transmission is restored when the communication is interrupted.

[0024] 5. Enhancement of dynamic risk assessment ability: The weighted scoring model introduces a time decay factor, improves the accuracy of hierarchical early warning, and reduces the missed alarm rate. Description of the drawings

[0025] Figure 1 is a schematic flow chart of the comprehensive monitoring method for gas, liquid level, and manhole cover movement in an underground manhole provided in a specific embodiment of the present invention; Figure 2 It is a schematic diagram of the system composition of the comprehensive monitoring system for gas, liquid level, and manhole cover movement in an underground manhole provided in a specific embodiment of the present invention. Specific embodiments

[0026] In order to enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Based on the embodiments in this application, other similar embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of this application. In addition, the directional terms mentioned in the following embodiments, such as "up", "down", "left", "right", etc. preferred by the present invention are only with reference to the directions of the drawings. Therefore, the directional terms used are for illustration rather than limiting the present invention.

[0027] The present invention will be further described below in conjunction with the drawings and preferred embodiments.

[0028] Please refer to Figure 1 , the present invention provides a comprehensive monitoring method for gas, liquid level, and manhole cover movement in an underground manhole, including: S1: Collect gas concentration, liquid level height, and manhole cover displacement data of the underground manhole through multi-sensor fusion technology.

[0029] Specifically, the multi-sensor fusion technology includes a gas sensor array, an ultrasonic liquid level gauge, and an inertial measurement unit, and data synchronization and joint estimation are realized through Kalman filtering.

[0030] Specifically, the inertial measurement unit is fixed on the edge of the manhole cover by pasting or magnetic attraction, and is used to monitor the three-axis acceleration data and angular velocity data of the manhole cover.

[0031] It should be further noted that the sensor deployment scheme designed by the present invention preferably includes the following: Gas sensor: An array composed of MQ-135 (VOC), Figaro TGS2611 ( , MiCS-2614 ( S) is installed 0.5 m above the ground at the top of the manhole, and data is collected every 10 minutes, and the temperature and humidity are synchronously recorded as auxiliary inputs. Data preprocessing: Calculate the mean and standard deviation through a sliding window (window size 15 min, step size 5 min) to eliminate high-frequency noise.

[0032] Liquid level sensor: The ultrasonic liquid level gauge (SI-7021) is installed at the center of the top of the manhole, with a measurement range of 0-3 m and an accuracy of . Compensation mechanism: According to the temperature sensor (DS18B20, accuracy .

[0033] Among them, m / s (speed of sound at 0°C), is the corrected speed of sound, and T is the Celsius temperature in the manhole.

[0034] Manhole cover displacement sensor: The MEMS inertial measurement unit (MPU-9250) is fixed on the edge of the manhole cover by pasting or magnetic attraction to monitor the three-axis acceleration and angular velocity , with a sampling rate of 100 Hz.

[0035] S2: An adaptive filtering algorithm is used to suppress noise and remove outliers from the sensor data to improve data accuracy.

[0036] It should be further noted that the S2 includes: Kalman filter parameter optimization: For the slow-varying characteristics of the gas concentration, the process noise covariance matrix Q is set as a diagonal matrix: It can be understood that The setting of corresponds to the tolerance of minute fluctuations of the three gases, adapting to the slow change characteristics of the gas concentration.

[0037] Observation noise covariance is estimated online adaptively: Among them is the original sensor reading, is the Kalman filter prediction value, and N is the last 10 sampling points.

[0038] Outlier rejection rule: If the data at three consecutive time points deviate from the moving average by more than ( is the real-time standard deviation), it is marked as an outlier and replaced with the predicted value.

[0039] In the preferred embodiment of the present invention, the measurement noise covariance R = 0.5. Setting the R to 0.5 is optimized based on 50 groups of on-site measured data to reduce noise interference.

[0040] S3: Based on the deep learning model, identify the gas components, distinguish the target gas from the interfering gas, and calculate the concentration of each component.

[0041] Specifically, the deep learning model adopts a structure combining a convolutional neural network and a long short-term memory network. The input layer receives the original sensor signal, and the output layer is the gas component classification result and concentration value.

[0042] It should be further noted that regarding the gas component identification, the present invention designs a CNN-LSTM model structure and a training scheme. The specific scheme includes: Input layer: The original gas signals at 128 time points (sampling rate 1 Hz, window length 128 s), normalized to [0, 1].

[0043] CNN part: Convolution layer 1: 32 channels, kernel size 3×3, ReLU activation, output size 126×32; Pooling layer: Max pooling 2×2, output size 63×32; Convolution layer 2: 64 channels, kernel size 3×3, ReLU activation, output size 61×64. It can be understood that the training channels are set to 32 to adapt to the computing power of the embedded GPU (preferably including NVIDIA Jetson Nano in the present invention).

[0044] LSTM part: The number of hidden layer units is 128, and the Dropout rate is set to 0.2 to prevent overfitting; Output layer: Softmax classification, and the target gas categories are CH 4 、H 2 S, CO, and normal air.

[0045] Training configuration: Loss function: Cross-Entropy Loss; Optimizer: Adam (learning rate 0.001, β 1 = 0.9, β 2 = 0.999); Dataset: UCI Gas Sensor Dataset, expanded to 10,000 samples by simulating real environment disturbances such as rotation and adding Gaussian noise.

[0046] Specifically, the distribution of the training data is shown in Table 1: Table 1 Training data distribution Number of training rounds: Early stopping strategy (terminate if the validation set loss does not decrease for 3 consecutive rounds, with an average training time of 5 hours).

[0047] S4: Use the pressure compensation algorithm to eliminate the influence of environmental temperature and pressure fluctuations in liquid level measurement and improve the liquid level measurement accuracy.

[0048] Specifically, the pressure compensation algorithm is based on the real-time data of the temperature and air pressure sensors in the manhole and dynamically corrects the liquid level measurement value through the state space model.

[0049] It should be further noted that step S4 is implemented through the Extended Kalman Filter (EKF), which specifically includes: 1. State equation: Wherein, is the state vector (liquid level height , temperature , air pressure , are the process / observation noises ( ), and the state transition function includes a thermal expansion correction: Parameter C (thermal expansion coefficient of stainless steel material), kPa (empirical value).

[0050] S5: Distinguish the normal vibration of the manhole cover from abnormal movement through a machine learning classifier to reduce the false alarm rate.

[0051] Specifically, the machine learning classifier adopts the random forest algorithm, and the training data includes a normal vibration signal feature library and an abnormal movement event sample library.

[0052] Specifically, the step S5 includes: Algorithm: Random Forest (RF) classifier: Feature engineering: Extract the time domain (mean, peak value) and time-frequency domain (wavelet decomposition energy) features of acceleration; Training data: Laboratory simulation of normal vehicle rolling (1000 samples) and abnormal movement (preferably including earthquake / human prying in the present invention, 200 samples); Decision threshold: Classification score > 0.8 is determined as abnormal (optimized through the ROC curve).

[0053] It should be further noted that the S5 includes: Random forest classifier: 1. Feature extraction: Time domain features: RMS acceleration, peak-to-peak value ( ; Frequency domain features: Proportion of energy in the first 5 frequency bands of FFT; Wavelet features: Mean of the 3-layer detail coefficients of db4 wavelet decomposition.

[0054] 2. Model configuration: Number of trees: 100, maximum depth: 20, minimum sample split: 20; Training data: 200 groups of abnormal samples generated by laboratory knocking tests, 1000 groups of on-site vehicle rolling data; Threshold selection: Determine the classification score threshold of 0.8 through the Precision-Recall Curve. Specifically, through the confusion matrix analysis, the threshold of 0.8 can maximize the F1-score, and the maximum value is 0.92.

[0055] S6: Use low-power wide-area communication technology to transmit the monitoring data to the remote monitoring platform, and enable the edge computing node to perform data caching and forwarding when the signal is blocked.

[0056] Specifically, the edge computing node deploys a lightweight data processing algorithm, temporarily stores the monitoring data during communication interruption, and resumes transmission through the breakpoint resumption mechanism.

[0057] It should be further noted that S6 includes the following specific solutions: LoRaWAN communication optimization: Transmission power 14dBm, frequency 868MHz, spreading factor SF12 (the farthest coverage is 5km in an open environment); data frame structure: 8-byte header information + sensor data (20 bytes) + 2-byte CRC check; Retransmission mechanism: Automatically retransmit 3 times when ACK is not received, with exponential backoff intervals (initial interval 1s, maximum 5s).

[0058] Edge computing strategy: The cache queue adopts a circular buffer (capacity 100), and preferentially stores high-priority data (preferably including gas concentration exceeding the standard in the present invention); breakpoint resumption: Send according to the timestamp sorting to ensure continuity.

[0059] S7: Build a multi-source data fusion model at the monitoring platform end, and conduct comprehensive risk assessment and early warning by combining gas concentration, liquid level change and manhole cover status.

[0060] Specifically, the comprehensive risk assessment adopts a weighted scoring mechanism, generates a risk level by combining gas toxicity, liquid level rising rate and manhole cover displacement amplitude, and triggers a hierarchical early warning.

[0061] Specifically, use the following formula to calculate the weighted score: Among them, is the weighted score, is the concentration of toxic gas, is the liquid level rising rate, is the impact energy of the manhole cover, t is the event duration (in minutes), and the denominator of the formula is the time decay factor, which is used to suppress short-term false alarms, and triggers a hierarchical early warning according to the weighted score and the preset hierarchical early warning threshold list.

[0062] It is understandable that the preset hierarchical warning threshold list and other related thresholds or weights can be specifically set according to the actual needs of the users of the present invention, as long as they are applicable to the solution proposed by the present invention, and the present invention does not limit them here.

[0063] In a preferred embodiment of the present invention, the triggering of the hierarchical warning according to the weighted score and the preset hierarchical warning threshold list includes: If 0.8 ≥ R > 0.6, trigger a level-three yellow warning; If 0.95 ≥ R > 0.8, trigger a level-two orange warning; If R > 0.95, trigger a level-one red warning; It should be noted here that the level-one red warning is the highest-level warning, and the above settings are the preferred values of the present invention, which can better ensure the effect of the hierarchical warning.

[0064] It should be further noted that the present invention also sets thresholds for the concentration of toxic gases, the rising rate of the liquid level, and the impact energy of the manhole cover respectively, and then issues separate warnings for the concentration of toxic gases, the rising rate of the liquid level, and the impact energy of the manhole cover. Specifically, the threshold for the concentration of toxic gases is 50 ppm, the threshold for the rising rate of the liquid level is 0.1 m / h, and the threshold for the impact energy of the manhole cover is 50 J. The above settings are the preferred values of the present invention, which can better ensure the effect of the separate warning and further improve the intelligence, reliability, and usability of the present invention.

[0065] It is understandable that the beneficial effects of this embodiment include: 1. Precise fusion of multi-source data: Multi-sensor fusion eliminates time deviation through Kalman filtering to achieve data synchronization, and the CNN-LSTM model effectively identifies gas components.

[0066] 2. Reduction of liquid level measurement error: The pressure compensation algorithm dynamically corrects the sound velocity error and improves the stability of liquid level measurement.

[0067] 3. Decrease in false alarm rate of manhole cover: The random forest classifier extracts time-frequency features, reduces the false alarm rate of manhole cover status monitoring, and supports the detection of micro-impacts.

[0068] 4. Improvement of communication reliability: LoRaWAN communication ensures data transmission, and the edge computing node cache ensures that data is not lost and the transmission is restored when the communication is interrupted.

[0069] 5. Enhancement of dynamic risk assessment ability: The weighted scoring model introduces a time decay factor, improves the accuracy of hierarchical warning, and reduces the missed alarm rate.

[0070] Please refer to Figure 2, the present invention provides another embodiment, which provides a comprehensive monitoring system for gas, liquid level, and manhole cover movement in an underground manhole. The comprehensive monitoring system for gas, liquid level, and manhole cover movement in the underground manhole includes: An acquisition module 100, configured to collect gas concentration, liquid level height, and manhole cover displacement data of the underground manhole through multi-sensor fusion technology; A control module 200, configured to suppress noise and eliminate outliers from sensor data by using an adaptive filtering algorithm to improve data accuracy; to identify gas components based on a deep learning model, distinguish target gases from interfering gases, and calculate the concentration of each component; to eliminate the influence of environmental temperature and pressure fluctuations in liquid level measurement by using a pressure compensation algorithm to improve liquid level measurement accuracy; to distinguish normal vibration of the manhole cover from abnormal movement by using a machine learning classifier to reduce false alarm rates; to transmit monitoring data to a remote monitoring platform by using a low-power wide-area communication technology, and enable an edge computing node to perform data caching and forwarding when the signal is blocked; to construct a multi-source data fusion model at the monitoring platform end, and perform comprehensive risk assessment and early warning by combining gas concentration, liquid level changes, and manhole cover status.

[0071] It can be understood that the beneficial effects of the present invention include: 1. Precise fusion of multi-source data: Multi-sensor fusion eliminates time deviation through Kalman filtering to achieve data synchronization, and the CNN-LSTM model effectively identifies gas components.

[0072] 2. Reduction of liquid level measurement error: The pressure compensation algorithm dynamically corrects the sound speed error to improve the stability of liquid level measurement.

[0073] 3. Decrease in false alarm rate of manhole cover: The random forest classifier extracts time-frequency features to reduce the false alarm rate of manhole cover status monitoring and supports micro-impact detection.

[0074] 4. Improvement in communication reliability: LoRaWAN communication ensures data transmission, and the edge computing node cache ensures that data is not lost and the transmission is restored when the communication is interrupted.

[0075] 5. Enhancement of dynamic risk assessment ability: The weighted scoring model introduces a time decay factor to improve the accuracy of hierarchical early warning and reduce the missed alarm rate.

[0076] In a preferred embodiment, the present application further provides an electronic device, and the electronic device includes: A memory; and a processor, wherein computer-readable instructions are stored on the memory, and when the computer-readable instructions are executed by the processor, the comprehensive monitoring method for the gas, liquid level, and manhole cover movement in the underground manhole is implemented. This computer device can be broadly a server, a terminal, or any other electronic device with necessary computing and / or processing capabilities. In one embodiment, the computer device may include a processor, a memory, a network interface, a communication interface, etc. connected through a system bus. The processor of the computer device can be used to provide necessary computing, processing, and / or control capabilities. The memory of the computer device may include a non-volatile storage medium and an internal memory. An operating system, a computer program, etc. may be stored in or on the non-volatile storage medium. The internal memory can provide an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface and the communication interface of the computer device can be used to connect and communicate with external devices through a network. The computer program, when executed by the processor, performs the steps of the method of the present invention.

[0077] The present invention can be implemented as a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method of the embodiments of the present invention are caused to be executed. In one embodiment, the computer program is distributed on a plurality of network-coupled computer devices or processors, so that the computer program is stored, accessed, and executed in a distributed manner by one or more computer devices or processors. A single method step / operation, or two or more method steps / operations, can be executed by a single computer device or processor or by two or more computer devices or processors. One or more method steps / operations can be executed by one or more computer devices or processors, and one or more other method steps / operations can be executed by one or more other computer devices or processors. One or more computer devices or processors can execute a single method step / operation, or execute two or more method steps / operations.

[0078] Those of ordinary skill in the art can understand that the method steps of the present invention can be implemented by a computer program to instruct relevant hardware. The present invention preferably includes a computer device or a processor. The computer program can be stored in a non-transitory computer-readable storage medium. When the computer program is executed, the steps of the present invention are caused to be executed. Depending on the circumstances, any reference to a memory, storage, database, or other medium herein may include non-volatile and / or volatile memory. Examples of non-volatile memory include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, magnetic tape, floppy disk, magneto-optical data storage device, optical data storage device, hard disk, solid state disk, etc. Examples of volatile memory include random access memory (RAM), external cache memory, etc.

[0079] It should be noted here that the beneficial effects of the present invention include: 1. Precise multi-source data fusion: Multi-sensor fusion eliminates time deviation through Kalman filtering to achieve data synchronization, and the CNN-LSTM model effectively identifies gas components.

[0080] 2. Reduction of liquid level measurement error: The pressure compensation algorithm dynamically corrects the sound speed error to improve the stability of liquid level measurement.

[0081] 3. Decrease in false alarm rate of manhole covers: The random forest classifier extracts time-frequency features to reduce the false alarm rate of manhole cover status monitoring and supports the detection of minor impacts.

[0082] 4. Improvement of communication reliability: LoRaWAN communication ensures data transmission, and the edge computing node cache ensures that data is not lost and the transmission is restored in case of communication interruption.

[0083] 5. Enhancement of dynamic risk assessment ability: The weighted scoring model introduces a time decay factor to improve the accuracy of hierarchical early warning and reduce the missed alarm rate.

[0084] The above-described technical features can be combined arbitrarily. Although not all possible combinations of these technical features are described, any combination of these technical features should be considered to be covered by this specification as long as such a combination is not contradictory.

[0085] The specific embodiments of the present invention described above do not constitute a limitation on the protection scope of the present invention. Any other corresponding changes and deformations made according to the technical concept of the present invention should be included within the protection scope of the claims of the present invention.

Claims

1. A comprehensive monitoring method for gas, liquid level and manhole cover movement in an underground manhole, characterized in that: The method comprises the following steps: S1: Collect gas concentration, liquid level and manhole cover displacement data of underground manholes through multi-sensor fusion technology; S2: Adaptive filtering algorithm is used to suppress noise and remove outliers from sensor data to improve data accuracy; S3: Identify gas components based on deep learning models, distinguish target gases from interfering gases, and calculate the concentration of each component; S4: Use pressure compensation algorithm to eliminate the influence of ambient temperature and pressure fluctuations in liquid level measurement and improve the accuracy of liquid level measurement; S5: Use machine learning classifiers to distinguish between normal vibration and abnormal movement of manhole covers, reducing false alarm rates; S6: Use low-power wide-area communication technology to transmit monitoring data to the remote monitoring platform, and enable edge computing nodes to cache and forward data when the signal is blocked; S7: Build a multi-source data fusion model on the monitoring platform to conduct comprehensive risk assessment and early warning based on gas concentration, liquid level changes and manhole cover status.

2. The comprehensive monitoring method for gas, liquid level and manhole cover movement in an underground manhole according to claim 1 is characterized in that: The multi-sensor fusion technology includes a gas sensor array, an ultrasonic level meter and an inertial measurement unit, and realizes data synchronization and joint estimation through Kalman filtering.

3. The comprehensive monitoring method for gas, liquid level and manhole cover movement in an underground manhole according to claim 2, characterized in that: The inertial measurement unit is fixed to the edge of the manhole cover by gluing or magnetic attraction, and is used to monitor the three-axis acceleration data and angular velocity data of the manhole cover.

4. The comprehensive monitoring method for gas, liquid level and manhole cover movement in an underground manhole according to claim 1, characterized in that: The deep learning model adopts a structure that combines a convolutional neural network with a long short-term memory network. The input layer receives the original signal from the sensor, and the output layer is the gas component classification result and concentration value.

5. The comprehensive monitoring method for gas, liquid level and manhole cover movement in an underground manhole according to claim 1, characterized in that: The pressure compensation algorithm is based on the real-time data of the temperature and air pressure sensors in the manhole, and dynamically corrects the liquid level measurement value through the state space model.

6. The comprehensive monitoring method for gas, liquid level and manhole cover movement in an underground manhole according to claim 1, characterized in that: The machine learning classifier adopts a random forest algorithm, and the training data includes a normal vibration signal feature library and an abnormal movement event sample library.

7. The comprehensive monitoring method for gas, liquid level and manhole cover movement in an underground manhole according to claim 1, characterized in that: The edge computing node deploys a lightweight data processing algorithm to temporarily store monitoring data when communication is interrupted and resume transmission through a breakpoint resume mechanism.

8. The comprehensive monitoring method for gas, liquid level and manhole cover movement in an underground manhole according to claim 1, characterized in that: The comprehensive risk assessment adopts a weighted scoring mechanism, combining gas toxicity, liquid level rise rate and manhole cover displacement amplitude to generate a risk level and trigger a graded warning.

9. The comprehensive monitoring method for gas, liquid level and manhole cover movement in an underground manhole according to claim 8, characterized in that: The weighted score is calculated using the following formula: in, is the weighted score, is the concentration of toxic gases, is the liquid level rising rate, is the impact energy of the manhole cover, t is the event duration, and the denominator of the formula is the time attenuation factor. The time attenuation factor is used to suppress short-term false alarms and trigger graded warnings according to the weighted score and the preset graded warning threshold list.

10. A comprehensive monitoring system for gas, liquid level and manhole cover movement in underground manholes, characterized in that: include: The acquisition module is used to collect the gas concentration, liquid level and manhole cover displacement data of underground manholes through multi-sensor fusion technology; The control module is used to suppress noise and remove outliers from sensor data using an adaptive filtering algorithm to improve data accuracy; to identify gas components based on a deep learning model, distinguish target gas from interfering gas, and calculate the concentration of each component; to use a pressure compensation algorithm to eliminate the effects of ambient temperature and pressure fluctuations in liquid level measurement to improve liquid level measurement accuracy; to use a machine learning classifier to distinguish between normal vibration and abnormal movement of manhole covers to reduce false alarm rates; to use low-power wide-area communication technology to transmit monitoring data to a remote monitoring platform, and to enable edge computing nodes to cache and forward data when signals are blocked; It is used to build a multi-source data fusion model on the monitoring platform, and conduct comprehensive risk assessment and early warning based on gas concentration, liquid level changes and manhole cover status.

Images