A natural potential method-based underground concrete structure leakage water detection system

Abstract

The application relates to a natural potential method-based underground concrete structure leakage water detection system, which comprises a data acquisition device and a data analysis system; the data acquisition device comprises a main control module and a signal acquisition module and a data uploading module connected with the main control module; the signal acquisition module is connected with the surface of the underground concrete structure and is used for collecting natural potential signal digital quantity; the main control module is used for reading the natural potential signal digital quantity and transmitting the natural potential signal digital quantity to a local data storage module, a display module and the data uploading module; the data analysis system is in communication connection with the data uploading module and receives the natural potential signal digital quantity; the data analysis system is used for processing the natural potential signal digital quantity and performing leakage monitoring on the processed natural potential data through an A-LSTM algorithm. Compared with other conventional underground concrete structure leakage detection methods, the application has the advantages of low cost, simple method and high use value.

Description

TECHNICAL FIELD

[0001] The application relates to a natural potential method-based underground concrete structure water leakage detection system and belongs to the technical field of underground building concrete structure water leakage data monitoring. BACKGROUND

[0002] There are various conventional underground concrete structure leakage detection methods, two of which are most commonly used, one is a manual listening tool for listening to water leakage sound, and the other is a high-density electric method using an artificial external electric field. However, the above methods have the following problems: extremely dependent on manual detection, unable to automatically sample, greatly affected by environmental factors, and low collection accuracy. Therefore, the above monitoring methods often require monitoring personnel to take samples on site and further analyze through experience, which is time-consuming and laborious, has low automation degree, and has high labor cost. SUMMARY

[0003] The application provides a system for water leakage detection based on a natural potential method to solve the problems in the prior art.

[0004] To achieve the above purpose, the technical scheme provided by the application is as follows: a natural potential method-based underground concrete structure water leakage detection system, characterized by comprising a data acquisition device and a data analysis system.

[0005] The data acquisition device comprises a main control module, a signal acquisition module connected with the main control module, and a data uploading module.

[0006] The signal acquisition module is connected with the surface of the underground concrete structure and is used for acquiring a natural potential signal digital quantity, the main control module is used for reading the natural potential signal digital quantity and transmitting the natural potential signal digital quantity to a local data storage module, a display module and the data uploading module.

[0007] The data analysis system is in communication connection with the data uploading module and receives the natural potential signal digital quantity, and the data analysis system is used for processing the natural potential signal digital quantity and performing water leakage monitoring on the processed natural potential data through an A-LSTM algorithm.

[0008] The further design of the above technical scheme is as follows: the signal acquisition module comprises two analog-to-digital conversion chips, the common ends of the two analog-to-digital conversion chips are connected with each other as a reference end, each analog-to-digital conversion chip is provided with eight signal acquisition ends, the reference end and the acquisition end are connected with a sensing electrode piece through a single-core shielding long wire, the sensing electrode piece is pasted on the surface of the underground concrete structure and is used for acquiring a natural potential signal analog quantity.

[0009] The signal acquisition module is connected with an analog-digital conversion chip after the first-stage RC passive filtering of the natural potential signal analog quantity, and the analog-digital conversion chip converts the natural potential signal analog quantity into a natural potential signal digital quantity.

[0010] The main control module comprises a power supply circuit, a clock circuit and a main control microprocessor chip; the power supply circuit is connected with the clock circuit and the main control microprocessor chip, the clock circuit is connected with the main control microprocessor chip and used for sending a clock signal to the main control microprocessor chip; the main control microprocessor chip reads the natural potential signal digital quantity through a serial peripheral interface (SPI) and transmits the natural potential signal digital quantity to the data uploading module through a universal asynchronous receiver-transmitter (UART) interface.

[0011] The system further comprises a local data storage module and a display module connected with the main control module respectively; the main control microprocessor chip drives the display module through a variable static memory controller and reads and writes the local data storage module through a secure digital input / output (SDIO) interface.

[0012] The data analysis system filters and reduces the dimension of the natural potential signal digital quantity through a local weighted regression algorithm; a loss function after the weighted processing in the local weighted regression algorithm is

[0013]

[0014] In the formula, J(θ) is the final output result, w i is a given weight matrix according to the distance x (i) , h θ (x (i) ) is a predicted value, and y (i) is an actual value.

[0015] The leakage monitoring of the processed natural potential data through the A-LSTM algorithm comprises the following steps: obtaining a water seepage start time through the MAD average absolute deviation method from the filtered natural potential digital quantity; obtaining abnormal natural potential data in a window of a sliding window detection algorithm through the sliding window detection algorithm from the reduced natural potential digital quantity; the determination standard of the abnormal natural potential data is that, compared with a normal situation time period, if an absolute value of a difference between a certain natural potential data and an average value of natural potential data in the normal situation time period is greater than 50% of the average value of the normal situation time period, or a variation coefficient of natural potential data in a certain time period is greater than 15%, or an absolute value of a difference between an average value of the natural potential data in the certain time period and the average value of the normal situation time period is greater than 20% of the average value of the normal situation time period; the natural potential data or the natural potential data in the certain time period is abnormal natural potential data;

[0016] The abnormal natural electric potential data in the water seepage starting time after window is input into an A-LSTM network model for classification, and the potential field variation law before and after seepage is determined based on the potential field potential function theory combined with back analysis, so as to early warn.

[0017] The local data storage module comprises an SD card memory.

[0018] The display module comprises an LCD screen, which is used for displaying the natural electric potential signal digital quantity.

[0019] The data uploading module comprises a 4G module, and the natural electric potential signal digital quantity is uploaded to the remote cloud in real time through the 4G module, and the data analysis system obtains the natural electric potential signal digital quantity from the remote cloud in real time.

[0020] The present application has the advantages that:

[0021] The present application proposes to detect water seepage based on the natural electric potential method, to collect the natural electric potential data of the area to be detected through a metal electrode sheet sensor, that is, to collect the direct current voltage signal generated due to the electrochemical effect and the electrokinetic effect, so as to effectively calculate the potential distribution and current density distribution of the underground space and to quantitatively determine the abnormal distribution characteristics.

[0022] The present application can realize real-time monitoring of the natural electric potential of the sampling area, and through data acquisition and processing, it can monitor in real time whether there is water leakage, analyze the water seepage trend, and if the water seepage trend is significantly expanded, find the leakage point for processing. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 is a system block diagram of the present application;

[0024] Figure 2 is a comparison diagram of different channel test results in the data analysis system;

[0025] Figure 3 is a whole structure diagram of the A-LSTM network model in the data analysis system;

[0026] Figure 4 is a schematic diagram of the present application installed on the underground concrete wall surface for application verification. DETAILED DESCRIPTION

[0027] The present application will be described in detail below in combination with the drawings and specific embodiments.

[0028] EMBODIMENT

[0029] As Figure 1 shown, the underground concrete structure leakage water detection system based on the natural potential method of the embodiment comprises a data acquisition device and a data analysis system, the data acquisition device comprises a main control module and a signal acquisition module, a local data storage module, a display module and a data uploading module connected with the main control module respectively; the data analysis system refers to the host computer program for processing and analyzing the data collected by the data acquisition device.

[0030] In the circuit of the signal acquisition module, two 24-bit high-precision analog-to-digital conversion chips ADS1256 are adopted, each chip contains 8 signal acquisition ends, and the common ends of the two chips are connected together as a reference end, so that 16 signal sampling electrode channels and 1 reference electrode channel can be formed; one end of each channel uses conductive glue to paste a copper electrode piece with a diameter of about 5 centimeters on the surface of the measured concrete, as Figure 4 shown, and the comparison chart of test results of different channels is as Figure 1 shown; in the circuit on the circuit board, the natural potential signal analog quantity is first subjected to one-stage RC passive filtering, and then connected to the above high-precision analog-to-digital conversion chip, and subjected to frequency sampling set by the main control module, so as to convert the natural potential signal analog quantity into natural potential signal digital quantity.

[0031] The main control module comprises a power supply circuit, a clock circuit and a main control microprocessor chip STM32F407ZGT6. The power supply circuit is externally connected with a 12V direct current voltage, and converts the power supply into 5V and 3.3V direct current voltages respectively for use of each module. The clock circuit is linked with the main control microprocessor chip for transmitting clock signals required for the work of the main control microprocessor chip. The main control microprocessor chip sets the sampling frequency according to the program, reads the natural potential signal digital quantity provided by the signal acquisition module through the serial peripheral interface SPI, and drives the display module through the variable static storage controller FSMC; reads and writes the local data storage module through the secure digital input / output interface SDIO; and transmits the natural potential signal digital quantity to the data uploading module through the universal asynchronous receive / transmit interface.

[0032] The local data storage module is connected with the main control module, and the local data storage module comprises an SD card memory. The local data storage module receives the natural potential signal digital quantity sent by the main control module through the universal asynchronous receive / transmit interface, and saves the data on the SD card memory according to the FATFS file format. The maximum capacity of the SD card memory can reach 128G, which can meet the data storage needs of the continuous work of the device for more than one year.

[0033] The display module is connected with the main control module, the display module includes an LCD screen, the display module receives the natural potential signal digital quantity sent by the main control module, and displays corresponding numerical values and curves on the LCD screen.

[0034] The data uploading module is connected with the main control module, the data uploading module includes a 4G module, the data uploading module receives the natural potential signal digital quantity sent by the main control module through a universal asynchronous receiving and transmitting interface, and uploads the natural potential signal digital quantity to a remote cloud in real time through the 4G module.

[0035] The data analysis system communicates with the data acquisition device through the Internet, the data analysis software receives the natural potential signal digital quantity transmitted to the cloud by the data uploading module in real time through the Internet, analyzes the natural potential data through a private protocol, performs lowess (local weighted regression) filtering and PCA (principal component analysis) dimension reduction on the natural potential data, and draws a natural potential signal waveform and a three-dimensional cloud chart, and then uses an A-LSTM algorithm (a long short-term memory neural network algorithm with an attention mechanism) to analyze the processed natural potential signal digital quantity. The data analysis system is developed based on MATLAB, and the lowess algorithm is used to filter noise in the signal, so that the subsequent analysis result is more accurate. The loss function after weighting processing in the lowess algorithm is as follows

[0036]

[0037] In the formula, J(θ) is the final output result, w i is a weight matrix given according to the distance x (i) , h θ (x (i) ) is a predicted value, and y (i) is an actual value. The PCA principal component analysis is used to reduce the dimension of the filtered 16-channel natural potential data for subsequent data processing. The A-LSTM algorithm is used to perform leakage monitoring on the processed natural potential data. The specific steps are as follows:

[0038] First, the data analysis system obtains the filtered natural potential data, uses the PCA principal component analysis algorithm to reduce the dimension of the data, retains channel data with a contribution degree accumulation of more than 90%, and uses a sliding window detection algorithm for detection.

[0039] Then, the filtered natural potential data is analyzed by the MAD average absolute deviation algorithm to obtain a water leakage start time, and the dimension-reduced data is analyzed by the sliding window detection algorithm to obtain abnormal natural potential data in the window.

[0040] The determination criterion of the abnormal spontaneous potential data is that, compared with the normal situation time period, the absolute value of the difference between a spontaneous potential data and the average value of the spontaneous potential data in the normal situation time period (determined according to the seepage field characteristics of the site, and no specific design value is recommended to be 24 hours) is greater than 50% of the average value of the normal situation time period, and the spontaneous potential data is the abnormal spontaneous potential data; or the coefficient of variation of the spontaneous potential data in a time period is greater than 15%, or the absolute value of the difference between the average value of the spontaneous potential data in a time period (determined according to the seepage field characteristics of the site, and no specific design value is recommended to be 24 hours) and the average value of the normal situation time period is greater than 20% of the average value of the normal situation time period, and the data in the time period is the abnormal spontaneous potential data;

[0041] Finally, the abnormal spontaneous potential data in the window after the water seepage start time is sent into an A-LSTM network algorithm for analysis and classification; the A-LSTM network algorithm uses a pre-trained model under supervised learning, and the network structure of the model is as shown in Figure 3 The LSTM (Long Short-Term Memory) layer, Attention (Attention) layer, Fully Connected (Fully Connected) layer and SoftMax function are used to obtain the final classification result of the abnormal spontaneous potential data in the window after the water seepage start time, and based on the potential field potential function theory and the back analysis, the change rule of the potential field before and after the seepage is determined, the spatial distribution of the potential anomaly is predicted, the specific position of the water seepage is obtained, and thus the early warning is provided, which provides a basis for further quickly finding the disease position and repairing.

[0042] The technical scheme of the present application is not limited to the above-mentioned embodiments, and any technical scheme obtained by equivalent replacement falls within the scope of the present application.

Claims

1. A system for detecting water leakage in underground concrete structures based on the natural potential method, characterized in that: Includes data acquisition devices and data analysis systems; The data acquisition device includes a main control module, a signal acquisition module connected to the main control module, and a data upload module. The signal acquisition module is connected to the surface of the underground concrete structure and is used to acquire digital quantities of natural potential signals. The main control module is used to read the digital quantities of natural potential signals and transmit them to the data upload module. The data analysis system is communicatively connected to the data upload module and receives digital quantities of natural potential signals. The data analysis system is used to process the digital quantities of natural potential signals and to perform leakage monitoring on the processed natural potential data using the A-LSTM algorithm. Leakage monitoring using processed natural potential data via the A-LSTM algorithm includes the following steps: The filtered natural potential digital value is used to determine the seepage start time using the MAD (mean absolute deviation) method. The reduced natural potential data is processed by a sliding window detection algorithm to obtain abnormal natural potential data within the window used by the algorithm. The criteria for judging abnormal natural potential data are as follows: if the absolute value of the difference between a certain natural potential data and the average value of natural potential data in a normal time period is greater than 50% of the average value of natural potential data in a normal time period, or the coefficient of variation of natural potential data in a certain time period is greater than 15%, or the absolute value of the difference between the average value of natural potential data in a certain time period and the average value of natural potential data in a certain time period is greater than 20% of the average value of natural potential data in a normal time period, then the natural potential data or the natural potential data in that time period is abnormal natural potential data. The abnormal spontaneous potential data within the window after the start of seepage are input into the A-LSTM network model for classification, and the change law of potential field before and after seepage is determined based on the potential field potential function theory combined with back analysis.

2. The underground concrete structure leakage detection system based on the natural potential method according to claim 1, characterized in that: The signal acquisition module includes two analog-to-digital converter chips. The common terminals of the two analog-to-digital converter chips are connected to each other as a reference terminal. Each analog-to-digital converter chip is equipped with eight signal acquisition terminals. The reference terminal and the acquisition terminals are connected to sensing electrode plates through single-core shielded long wires. The sensing electrode plates are pasted on the surface of the underground concrete structure for acquiring analog quantities of natural potential signals.

3. The underground concrete structure leakage detection system based on the natural potential method according to claim 2, characterized in that: The signal acquisition module performs a first-stage RC passive filter on the acquired analog natural potential signal and then connects it to the analog-to-digital converter chip, which converts the analog natural potential signal into a digital natural potential signal.

4. The underground concrete structure leakage detection system based on the natural potential method according to claim 3, characterized in that: The main control module includes a power supply circuit, a clock circuit, and a main control microprocessor chip. The power supply circuit is connected to the clock circuit and the main control microprocessor chip. The clock circuit is connected to the main control microprocessor chip and is used to send clock signals to the main control microprocessor chip. The main control microprocessor chip reads the digital value of the natural potential signal through the serial peripheral interface SPI and transmits the digital value of the natural potential signal to the data upload module through the universal asynchronous transceiver interface.

5. The underground concrete structure leakage detection system based on the natural potential method according to claim 4, characterized in that: It also includes a local data storage module and a display module that are respectively connected to the main control module. The main control microprocessor chip drives the display module through a variable static storage controller and performs read and write communication with the local data storage module through a secure digital input / output interface.

6. The underground concrete structure leakage detection system based on the natural potential method according to claim 5, characterized in that: The data analysis system performs filtering and principal component analysis dimensionality reduction on the digital quantity of the natural potential signal using a local weighted regression algorithm; the loss function after weighting in the local weighted regression algorithm is: ; In the formula: For the final output, By distance Given a weight matrix, For predicted values, This is the actual value.

7. The underground concrete structure leakage detection system based on the natural potential method according to claim 6, characterized in that: The local data storage module includes an SD card storage device.

8. The underground concrete structure leakage detection system based on the natural potential method according to claim 7, characterized in that: The display module includes an LCD screen for displaying digital quantities of natural potential signals.

9. The underground concrete structure leakage detection system based on the natural potential method according to claim 8, characterized in that: The data upload module includes a 4G module, through which the digital quantity of the natural potential signal is uploaded to the remote cloud in real time, and the data analysis system obtains the digital quantity of the natural potential signal from the remote cloud in real time.

Images