Tunnel drainage system with multi-dimensional monitoring and early warning functions and control method thereof

Abstract

The invention discloses a tunnel drainage system with a multi-dimensional monitoring and early warning function and a control method of the tunnel drainage system. The system comprises a water collecting tank, a wastewater pump group, a water level sensor, a water sample collection pipeline, a monitoring module, a data collection module, a data iteration module, a water leakage safety monitoring model and an alarm module. The monitoring module is integrated with a pH detector, a turbidity detector, a conductivity detector and a COD (Chemical Oxygen Demand) detector, and is used for acquiring a water sample through a collection pipeline and carrying out multi-dimensional water quality detection. The method comprises the following steps: collecting water sample and water level information; constructing a fusion data matrix; the safety monitoring model is optimized through iterative training, and leakage water source classification and a water level prediction trend are output; and performing abnormity judgment based on the model output and a preset threshold value, triggering early warning, and controlling the water pump to start and stop. According to the method, the water level and the multi-dimensional water quality data are fused, leakage source identification and trend prediction are realized by using the iterative model, the problems of response lag and single monitoring of an existing system are solved, and the intelligent level and the early warning capability of tunnel drainage management are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of tunnel drainage and monitoring technology, specifically relating to a tunnel drainage system and its control method with multi-dimensional monitoring and early warning functions. Background Technology

[0002] Currently, existing tunnel drainage systems mainly rely on wastewater pumps and level sensors to control pump start / stop and alarms based on changes in the water level in the collection tank. Specifically, a single pump house typically has two submersible sewage pumps. The drainage volume of these pumps includes tunnel structural leakage water, fire-fighting water, and tunnel flushing water. The collection tank has five water level nodes: low alarm level, pump stop level, first pump level, second pump level, and high alarm level. The control logic is as follows: when the water level is too low, an alarm is triggered and the pumps stop; when the water level reaches the pump stop level, both pumps stop; when the water level rises to the first pump level, the first pump starts; when the water level reaches the second pump level, both pumps run simultaneously; and when the water level is too high, an alarm is triggered.

[0003] While existing tunnel drainage systems can achieve basic drainage functions, they have significant limitations: First, relying solely on water level information makes it impossible to distinguish the source of leakage (such as groundwater or damaged water supply pipes), making it difficult to determine the cause of leakage and hindering targeted treatment. Second, the response is delayed, unable to handle emergencies such as a large influx of rainwater or a burst underground pipe network causing a rapid rise in water volume; alarms are only triggered when the water level is already high, making early warning and preventative intervention difficult. Third, monitoring methods are limited, ignoring differences in the physicochemical indicators of leakage water (such as pH, turbidity, conductivity, and COD), making it impossible to predict abnormal water level trends in advance based on water quality changes. Fourth, they are highly dependent on manual inspection; multiple wastewater pumping stations along subway lines require on-site inspections, resulting in a large workload, low efficiency, and susceptibility to errors due to subjective judgment, making automated and intelligent monitoring and early warning impossible. With the increasing complexity of tunnel networks and the intricate integration of municipal pipe networks, existing systems are insufficient to meet the demands for efficient and accurate leakage control.

[0004] Therefore, there is an urgent need for a tunnel drainage system that can monitor wastewater indicators from multiple dimensions, integrate historical data to build models, and achieve early warning, in order to solve the key problems of existing technologies in wastewater source identification, emergency warning, and intelligent management. Summary of the Invention

[0005] To address the above technical issues, this invention proposes a tunnel drainage system and its control method with multi-dimensional monitoring and early warning functions. The aim is to improve the safety and efficiency of the drainage system by identifying the source of tunnel leakage, providing early warning, and automating control through multi-dimensional water quality monitoring, data fusion modeling, and intelligent iterative training.

[0006] A tunnel drainage system with multi-dimensional monitoring and early warning functions includes: A water collection tank is used to collect leaking water from the tunnel. A water level sensor is installed in the water collection tank to monitor in real time whether the water level has reached one of the five preset water levels: ultra-low, pump stop, pump 1, pump 2, and ultra-high. A water sample collection tube is connected to the bottom of the water collection tank for extracting water samples; A regulating valve is installed on the water sample collection tube to control the flow of the water sample; An electric pump, connected to the water sample collection tube, is used to provide power for extracting water samples; The monitoring module receives water samples through the water sample collection tube, performs multi-dimensional water quality index detection on the water samples, and transmits the data to the data acquisition module. The monitoring module includes at least a pH meter, a turbidity meter, a conductivity meter, and a chemical oxygen demand meter. The data acquisition module is communicatively connected to the water level sensor and the monitoring module, and is used to collect time-series water level information and water quality index data, and construct a fused data matrix; The data iteration module is used to iteratively train the fused data matrix; The water leakage safety monitoring model is optimized through training by the data iteration module and is used to output water leakage source classification and water level prediction trend based on the input fusion data matrix. The water leakage alarm module is communicatively connected to the water leakage safety monitoring model and is used to determine whether to trigger an alarm based on the model output results and preset thresholds. The wastewater pump set is communicatively connected to the leakage alarm module and is used to start and stop according to the alarm command to drain the accumulated water in the collection tank.

[0007] A wastewater pipe is connected to the monitoring module, and a drain valve is installed at its end to control the drainage flow rate.

[0008] Furthermore, the leakage safety monitoring model employs machine learning algorithms, such as neural networks or decision trees; the data iteration module uses gradient descent algorithm for iterative training.

[0009] Furthermore, the preset thresholds of the water leakage alarm module include a water level threshold and a water quality threshold; wherein, the water quality threshold includes at least a pH threshold, a turbidity threshold, a conductivity threshold, and a chemical oxygen demand threshold.

[0010] Furthermore, the electric pump is an electric pump that supports timed automatic sampling; each detector in the monitoring module is a high-precision online sensor that supports continuous monitoring.

[0011] Furthermore, the system also includes a historical data storage unit connected to the data acquisition module, used to store long-term multi-dimensional monitoring data to form a time-series dataset; the data acquisition module uses cloud platform storage to support data sharing and integration of multiple pump stations, facilitating full-line analysis.

[0012] The present invention also relates to a control method for a tunnel drainage system, comprising the following steps: S1: Start the electric pump, draw water samples from the water collection tank through the water sample collection tube, and transport them to the monitoring module for multi-dimensional water quality index detection through the regulating valve. At the same time, collect water level information through the water level sensor. S2: Collect real-time monitoring data through the data acquisition module and integrate historical data to construct a fused data matrix; S3: The data iteration module is used to iteratively train the fused data matrix to optimize the leakage safety monitoring model, and the model outputs the leakage source classification and water level prediction trend. S4: The leakage alarm module performs anomaly judgment based on the model output and preset threshold. If an anomaly is judged, an early warning is triggered and the start and stop of the wastewater pump group are controlled.

[0013] Furthermore, in step S1, sampling supports both manual and automatic modes, and automatically increases the sampling frequency when an abnormal rate of water level rise is detected.

[0014] Furthermore, in step S4: when the water quality index is abnormal but the water level is normal, it is judged as a potential leakage risk and a low-level warning is issued; when the water quality is abnormal and the water level is predicted to rise, it is judged as an emergency, two wastewater pump sets are started and a high-level alarm is issued.

[0015] Furthermore, in step S4: when the conductivity is greater than 1000 μS / cm or the chemical oxygen demand is greater than 50 mg / L, it is determined that the water supply pipeline is damaged and a targeted maintenance alarm is triggered; when the turbidity is greater than 20 NTU and the pH value is less than 7, it is determined that groundwater has seeped in and a geological alarm is issued.

[0016] Furthermore, in step S4: when the water quality indicators are normal but the water level rises sharply, it is determined by combining historical data to be a sudden rainwater inrush, and the emergency drainage mode is activated; when the water quality indicators are abnormal but the water level is low, it is determined to be an early leakage risk, and early intervention is carried out.

[0017] Compared with the prior art, the beneficial effects of the present invention are: This invention provides a tunnel drainage system and its control method with multi-dimensional monitoring and early warning functions for leakage. The drainage system integrates a water level sensor and a multi-dimensional water quality monitoring module (pH, turbidity, conductivity, COD) to achieve coordinated control of leakage source identification (such as groundwater or pipe damage) and water level trend prediction, effectively improving the safety and efficiency of tunnel drainage. The system employs a data fusion matrix and iterative training model, possessing advantages such as high efficiency, intelligence, modular design, and flexible operation. It can dynamically switch early warning modes based on real-time data, flexibly responding to sudden leakage and abnormal water level issues. Combined with data acquisition and alarm modules, it achieves full-line monitoring of multiple pump stations, enabling early prediction of water level rises through water quality changes, significantly reducing the risk of overflow. Furthermore, this invention adopts automated sampling and remote monitoring design, greatly reducing manual maintenance workload and improving the overall economy, adaptability, and intelligence level of the system. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of a tunnel drainage system with multi-dimensional monitoring and early warning functions for water leakage. Figure 2 This is a flowchart of the control method for a tunnel drainage system according to an embodiment of the present invention.

[0019] Figure labeling: 1-Collection tank; 2-Water level sensor; 3-Water sample collection tube; 4-Regulating valve; 5-Electric pump; 6-pH meter; 7-Turbidity meter; 8-Conductivity meter; 9-Chemical oxygen demand meter; 10-Drain valve; 11-Water sample discharge tube; 12-Data acquisition module; 13-Data iteration module; 14-Leakage safety monitoring model; 15-Leakage alarm module; 16-Wastewater pump set. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0021] Example A tunnel drainage system with multi-dimensional monitoring and early warning functions for water leakage (such as...) Figure 1 As shown, it includes a water collection tank 1, a water level sensor 2, a water sample collection tube 3, a regulating valve 4, an electric pump 5, a monitoring module, a wastewater pipe (water sample discharge pipe) 11, a data acquisition module 12, a data iteration module 13, a leakage safety monitoring model 14, a leakage alarm module 15, a wastewater pump set 16, etc.

[0022] The water collection tank 1 is used to collect seepage water from the tunnel. It is equipped with a water level sensor 2 to monitor the water level in real time, indicating whether it has reached one of five preset levels: ultra-low, pump stop, pump 1, pump 2, and ultra-high. A water sample collection pipe 3 connects to the bottom of the water collection tank 1. The pipe is equipped with a regulating valve 4 and an electric pump 5 that provides the extraction power. This electric pump 5 supports automatic timed sampling. The water sample collected through this pipe is transported to the monitoring module. This module integrates a pH meter 6, a turbidity meter 7, a conductivity meter 8, and a chemical oxygen demand (COD) meter 9, all of which are high-precision online sensors. These are used to measure the water's pH (normal range 7~7.5), suspended particulate matter content (normal <5 NTU), dissolved ion concentration (normal <500 μS / cm), and the degree of organic and reducing inorganic pollution (normal <20 mg / L), respectively, thereby achieving multi-dimensional indicator detection and transmitting the data to the data acquisition module 12. The data acquisition module 12 is responsible for collecting the aforementioned real-time water level, water quality, and historical data, constructing a fused data matrix, and storing it on a cloud platform to achieve data sharing and full-line analysis across multiple pump stations. The data iteration module 13 uses a gradient descent algorithm to iteratively train the fused data matrix to optimize the leakage safety monitoring model 14 based on a neural network or decision tree. This model takes the fused data matrix as input and outputs a leakage source classification (e.g., low pH and high turbidity indicate groundwater infiltration, high conductivity and high COD indicate water supply pipeline damage) and a water level prediction trend. The leakage alarm module 15, based on the model output and combined with preset water level thresholds (ultra-low, pump stop, pump 1, pump 2, ultra-high) and water quality thresholds (e.g., pH 7.5, turbidity < 5 NTU, conductivity < 500 μS / cm, COD < 20 mg / L), determines whether to trigger an alarm and controls the start and stop of the wastewater pump group 16 accordingly. Wastewater pump set 16 is used to drain water accumulated in collection tank 1, and a drain valve 10 is provided at the end of the wastewater pipe 11 connected to the outlet of the monitoring module to control the drainage flow. In addition, the historical data storage unit of data acquisition module 12 is used to store multi-dimensional monitoring data for a long time to form a time series dataset, which supports the continuous optimization of the model.

[0023] In this embodiment, the system has multiple monitoring modes: in normal mode, it samples every 15 minutes to monitor water quality changes; in emergency mode, it automatically increases the sampling frequency when the water level rise rate is abnormal. It predicts water level trends and provides early warnings based on early changes in water quality indicators (such as a sharp increase in turbidity indicating rainwater intrusion). For example, when pH < 6.5 and turbidity > 15 NTU, it determines groundwater infiltration and issues a source alarm; when conductivity > 800 μS / cm and COD > 30 mg / L, it determines pipeline damage and triggers a maintenance notification. Simultaneously, combined with water level information, it starts one pump in advance to prevent overflow when the predicted water level is about to reach the level of the second pump. This system can be deployed in multiple wastewater pumping stations distributed along the subway line. Each pumping station is equipped with the aforementioned drainage system, and data is integrated through the subway line monitoring network to achieve unified monitoring and remote viewing of leakage water throughout the entire line. Real-time data and alarm information are displayed via mobile devices or a control center.

[0024] The control method based on the above tunnel drainage system (process as follows) Figure 2 (As shown), including the following steps: S1: Start the electric pump 5 and draw water samples from the collection tank 1 through the water sample collection tube 3. The samples then enter the monitoring module through the regulating valve 4 for pH, turbidity, conductivity and COD detection. At the same time, the water level sensor 2 collects water level information. The electric pump 5 supports manual or automatic mode for timed operation to ensure the freshness of the water samples. The monitoring module outputs multi-dimensional data in real time, and the sampling frequency increases to once every 15 minutes in case of emergencies. S2: Real-time monitoring data is collected through data acquisition module 12, and historical data is integrated to construct a fusion data matrix containing time, water level and water quality indicators, forming a multidimensional dataset; S3: The data iteration module 13 is used to iteratively train the fused data matrix to optimize the leakage safety monitoring model 14. The model outputs the leakage source and water level prediction trend. This training adopts machine learning methods and updates the model weekly to improve the prediction accuracy. S4: The leakage alarm module 15 makes an anomaly judgment based on the model output and the preset threshold. If an anomaly is found, an early warning is triggered and the wastewater pump group 16 is controlled to start and stop drainage. The threshold can be customized, including alarms for excessively high water level or abnormal water quality, and drainage is started in advance when the water level is predicted to rise.

[0025] Specifically, in step S4, when the water quality is abnormal but the water level is normal, a low-level early warning notification for maintenance is issued; when the water quality is abnormal and the water level rises, a high-level alarm is issued and two wastewater pumps are started; when the conductivity is >1000μS / cm or COD is >50mg / L, it is determined to be a pipe rupture and a special alarm is triggered; when the turbidity is >20NTU and pH <7, it is determined to be groundwater infiltration and a geological alarm is issued; in addition, if the water quality indicators are normal but the water level rises rapidly, it is determined to be a sudden rainwater inrush based on historical data, and emergency drainage is initiated; if the water quality is abnormal but the water level is low, it is identified as an early leakage risk, and early intervention is carried out to avoid the accumulation of problems.

[0026] In one embodiment, multiple water level sensors 2 are distributed at different heights in the collection tank 1, and the monitoring modules are evenly distributed throughout the pumping station, supporting zoned monitoring. During system operation, when an abnormal water quality is detected in a certain area, the data iteration module 13 automatically updates the model to achieve adaptive learning.

[0027] This application integrates traditional water level control with multi-dimensional water quality monitoring, and through data fusion and iterative models, it achieves the identification of leakage sources and early warning. The system can operate independently in a single pumping station or be networked for full-line monitoring, and is suitable for subways or municipal tunnels, possessing a high degree of intelligence and adaptability.

[0028] In summary, this invention discloses a tunnel drainage system and its control method with multi-dimensional monitoring and early warning functions for water leakage. It employs a collaborative design of a collection tank 1, a water level sensor 2, monitoring modules (pH meter 6, turbidity meter 7, conductivity meter 8, chemical oxygen demand meter 9), and a data module to construct a fused data matrix. Through iterative training, a water leakage safety monitoring model 14 and an alarm module 15 are formed. The system collects water level and water quality indicators in real time, predicts leakage sources and water level trends, and achieves early warning and intelligent drainage control.

[0029] This technical solution not only overcomes the problems of slow response and limited monitoring in existing systems, but also reduces the need for manual maintenance and improves the safety, economy, and intelligence of tunnel drainage. For complex environments such as subways and municipal tunnels, this invention provides an efficient, accurate, and sustainable solution for leakage control, with promising application prospects and widespread value.

[0030] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the invention (including the claims) is limited to these examples. Within the framework of this invention, technical features of the above embodiments or different embodiments can also be combined, steps can be implemented in any order, and many other variations of different aspects of the invention as described above exist, which are not provided in detail for the sake of brevity. Although the invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features, and these modifications or substitutions do not cause the essence of the corresponding technical solutions to depart from the scope of the technical solutions of the embodiments of this invention.

Claims

1. A tunnel drainage system with multi-dimensional monitoring and early warning functions, characterized in that, include: Water collection pool (1) is used to collect seepage water from the tunnel; A water level sensor (2) is installed in the water collection tank (1) to monitor in real time whether the water level has reached the five preset water levels: ultra-low, pump stop, first pump, second pump and ultra-high. A water sample collection tube (3) is connected to the bottom of the water collection tank (1) for extracting water samples; A regulating valve (4) is installed on the water sample collection tube (3); An electric pump (5) is connected to the water sample collection tube (3) to provide power for extracting water samples; The monitoring module receives water samples through the water sample collection tube (3), performs multi-dimensional water quality index detection on the water samples, and transmits the data to the data acquisition module (12). The monitoring module includes at least a pH meter (6), a turbidity meter (7), a conductivity meter (8), and a chemical oxygen demand meter (9). The data acquisition module (12) is communicatively connected to the water level sensor (2) and the monitoring module, and is used to collect time series water level information and water quality index data, and construct a fusion data matrix; The data iteration module (13) is used to iteratively train the fused data matrix; The leakage safety monitoring model (14) is optimized through training of the data iteration module (13) and is used to output leakage source classification and water level prediction trend based on the input fusion data matrix. The water leakage alarm module (15) is connected in communication with the water leakage safety monitoring model (14) and is used to determine whether to trigger an alarm based on the model output results and preset thresholds. Wastewater pump set (16) is communicatively connected to the water leakage alarm module (15) and is used to start and stop according to the alarm command to discharge the water in the collection tank (1). Wastewater pipe (11) is connected to the monitoring module, and a drain valve (10) is installed at its end. The drain valve (10) is used to control the drainage flow.

2. The system according to claim 1, characterized in that, The leakage safety monitoring model (14) adopts machine learning algorithms, such as neural networks or decision trees; the data iteration module (13) adopts gradient descent algorithm for iterative training.

3. The system according to claim 1 or 2, characterized in that, The preset thresholds of the water leakage alarm module (15) include water level thresholds and water quality thresholds; wherein, the water quality thresholds include at least pH thresholds, turbidity thresholds, conductivity thresholds and chemical oxygen demand thresholds.

4. The system according to claim 1, characterized in that, The electric pump (5) is an electric pump that supports timed automatic sampling; each detector in the monitoring module is a high-precision online sensor that supports continuous monitoring.

5. The system according to claim 1, characterized in that, It also includes a historical data storage unit connected to the data acquisition module (12) for storing long-term multi-dimensional monitoring data to form a time series dataset; the data acquisition module (12) uses cloud platform storage to support data sharing and integration of multiple pump stations, which facilitates full-line analysis.

6. A control method based on the tunnel drainage system according to any one of claims 1-5, characterized in that, Includes the following steps: S1: Start the electric pump (5), extract water samples from the water collection tank (1) through the water sample collection tube (3), and transport them to the monitoring module for multi-dimensional water quality index detection through the regulating valve (4). At the same time, collect water level information through the water level sensor (2). S2: Real-time monitoring data is collected through the data acquisition module (12), and historical data is integrated to construct a fused data matrix; S3: The data iteration module (13) is used to iteratively train the fused data matrix to optimize the leakage water safety monitoring model (14), and the model outputs the leakage water source classification and water level prediction trend; S4: The leakage alarm module (15) makes an anomaly judgment based on the model output and preset threshold. If the anomaly is judged, an early warning is triggered and the start and stop of the wastewater pump group (16) are controlled.

7. The method according to claim 6, characterized in that, In step S1, sampling supports both manual and automatic modes, and the sampling frequency is automatically increased when an abnormal rate of water level rise is detected.

8. The method according to claim 6, characterized in that, In step S4: when the water quality index is abnormal but the water level is normal, it is judged as a potential leakage risk and a low-level warning is issued; when the water quality is abnormal and the water level is predicted to rise, it is judged as an emergency, and two wastewater pump sets (16) are started and a high-level alarm is issued.

9. The method according to claim 6, characterized in that, In step S4: when the conductivity is greater than 1000 μS / cm or the chemical oxygen demand is greater than 50 mg / L, it is determined that the water supply pipeline is damaged and a targeted maintenance alarm is triggered; when the turbidity is greater than 20 NTU and the pH value is less than 7, it is determined that groundwater has seeped in and a geological alarm is issued.

10. The method according to claim 6, characterized in that, In step S4: when the water quality indicators are normal but the water level rises sharply, it is determined by combining historical data to be a sudden rainwater inrush and the emergency drainage mode is activated; when the water quality indicators are abnormal but the water level is low, it is determined to be an early leakage risk and early intervention is carried out.

Images