Modularized pumping and drainage system of large hydropower station and emergency rescue pumping and drainage control method

Abstract

The invention relates to a modularized water pumping and draining system for a large hydropower station. The system comprises a water pump body module, a counterweight module, a water conveying pipe and connector module and a water pumping and draining control and monitoring device, the water pumping and draining control and monitoring device comprises a container type water pump integrated control station and a remote control monitoring system terminal. The container type water pump integrated control station comprises a water pump power supply cabinet, a control cabinet, a cable winch and a control operation table. The invention further discloses a corresponding emergency rescue pump drainage control method. Through modular design of the water pump body module, the balance weight module, the water conveying pipe and the connector module, rapid transportation and field assembly of the water pumping and draining equipment are achieved, and the response time of emergency rescue is greatly shortened; through intelligent control and data-driven decision making, resource allocation is optimized, emergency rescue efficiency is improved, and rescue cost is reduced.

Description

Technical Field

[0001] The invention belongs to the field of drainage control of hydraulic structures, and in particular relates to a modularized drainage system for a large hydropower station. Background Art

[0002] During the daily operation of the power station in the river basin, emergency rescue conditions may occur. How to quickly pump out water is a problem that needs to be solved. In addition, during the annual maintenance period, the flood discharge and energy dissipation facilities of the hydropower station will be inspected and repaired, and it is also necessary to pump out water from energy dissipation buildings such as the energy dissipation pool and water cushion pond.

[0003] According to the current research status at home and abroad, the pumping and drainage of hydropower station stilling pools and cushion ponds mainly use non-mobile water pumping devices. This non-mobile pumping device has complex pipelines, and each pumping requires frequent transportation and hoisting, which is a large amount of repetitive engineering, high cost, and high safety risks. At the same time, the pumping efficiency and regulation performance are insufficient, and it does not have the functions of remote control of pumping and remote monitoring of pumping and drainage status. Summary of the invention

[0004] The purpose of the present invention is to address the above-mentioned problems and provide a modular drainage system for a large hydropower station, which adopts a lightweight and modular design for water pumps, pumping pipes and joints to facilitate transportation to the installation site and rapid assembly, thereby improving the efficiency of water pump installation and drainage; realizing remote control and fault prediction of water pumps, and meeting the drainage needs of different hydropower stations in multiple scenarios. At the same time, an emergency rescue drainage control method is provided, which can realize the rapid transportation and on-site assembly of drainage equipment and shorten the response time of emergency rescue; the drainage level prediction model is used to predict the drainage level in the future period, providing a scientific basis for emergency rescue decisions; the possibility of water pump failure is predicted by the water pump fault prediction model, potential faults are discovered in advance, and the impact of water pump failure on rescue work is reduced; the emergency rescue drainage plan can be adjusted in time according to the predicted water pump failure, thereby improving the stability and reliability of rescue work.

[0005] The technical solution of the present invention is a modular pumping and drainage system for a large hydropower station, comprising: A water pump body module, comprising a submersible pump and a rectangular parallelepiped frame for supporting the submersible pump, wherein a hook connection device is provided at the top of the rectangular parallelepiped frame, and a plurality of universal wheels with a brake function are provided at the bottom of the rectangular parallelepiped frame; A counterweight module is used to fix the water outlet of the water pipe; the outlet of the counterweight module is provided with a back pressure valve and a flow meter for increasing the flow of the water pump; A water pipe and joint module, including a water pipe and a pipe joint capable of quick lap-joint fastening; The pumping and drainage control and monitoring device is electrically connected to the water pump body module and is used to start and stop the water pump, control the water pump motor, and monitor the water pump status and water pipe flow, electrical parameters of the water pump motor, and water level of the pumping and drainage object.

[0006] Furthermore, the pumping and drainage control and monitoring device includes a container-type integrated water pump control station, which includes a water pump power cabinet, a control cabinet, a cable winch and a control console; the control cabinet includes a PLC controller that is communicatively connected to the control console.

[0007] Preferably, the drainage control and monitoring device also includes a remote control monitoring system terminal communicatively connected to the containerized water pump integrated control station, which is used to remotely control the start and stop of the submersible pump of the water pump body module and monitor drainage parameters and electrical parameters of the submersible pump motor.

[0008] Preferably, the drainage system also includes a tracked pump truck that can be quickly moved and deployed. A tracked walking mechanism is provided at the bottom of the tracked pump truck for carrying the water pump body module. The movement of the tracked pump truck is controlled by remote control.

[0009] Preferably, the drainage system also includes an emergency pump truck control end cart, which includes a chassis and a cable winch module, a water pump start and stop control module and a tracked pump truck control module arranged on the chassis. A plurality of universal wheels are arranged at the bottom of the chassis. The tracked pump truck control module is used to control the tracked pump truck and is communicatively connected with the tracked pump truck.

[0010] Preferably, the pumping and drainage control and monitoring device also includes a first prediction model for predicting the pumping and drainage level. The first prediction model is constructed using a bidirectional gated recurrent unit BiGRU. The input of the first prediction model is the pumping and drainage object, water pump operation data, current pumping and drainage level and meteorological condition data, and the output of the first prediction model is the predicted pumping and drainage level.

[0011] Furthermore, the first prediction model introduces an attention mechanism into BiGRU, and the calculation process of the prediction output with the attention mechanism is: Calculate the hidden state vector Similarity score with the query vector ; In the formula Indicates i A hidden state vector, formed by the concatenation of the forward hidden state and the backward hidden state; Q The query vector representing the attention mechanism; Calculate the hidden state vector Similarity score The index value of ;

[0012] In the formula e Represents the base of the natural exponential function; Hidden State The attention weight ; Prediction Output ; In the formula Represents the output of BiGRU with attention mechanism.

[0013] Preferably, the pumping and drainage control and monitoring device also includes a second prediction model for predicting water pump failure. The second prediction model is constructed using an autoencoder. The input of the second prediction model is water pump operating parameters, water pump and water pipe vibration signal characteristic data, predicted pumping and drainage levels and meteorological conditions data. The output of the second prediction model is the possibility of water pump failure, that is, the probability of water pump failure.

[0014] Pumping and drainage monitoring system, including: On-site control unit: used to control the start and stop of water pumps and motors; obtain the real-time water level of the pumping object and the operating parameters of the water pump and motor through sensors; Water pump operation monitoring unit: used to monitor the start and stop status of the water pump, pumping and drainage flow, water pipe outlet pressure and water level of the pumping and drainage object; Pump motor monitoring unit: used to monitor the speed, voltage, current and temperature of the water pump motor; Pumping and drainage level prediction module: using the first prediction model to predict the pumping and drainage level of the pumping and drainage object in the future period; Water pump failure prediction module: using the second prediction model to predict the possibility of water pump failure; Alarm unit: issues fault alarm signals for detected or predicted pump and motor faults; Host computer unit: used for configuration monitoring of pumping and drainage equipment; issuing start and stop control commands for water pumps and motors; remotely controlling crawler pump trucks.

[0015] The above-mentioned emergency rescue pumping and drainage control method of the modular pumping and drainage system of a large hydropower station includes the following steps: Step 1: Transport the water pump body module, counterweight module, water pipe and joint module to the rescue site; Step 2: Use a crawler pump truck to deploy the water pump body module to the emergency rescue drainage site; Step 3: Connect the water pump body module, water pipe and joint module to each other, and electrically connect the water pump body module to the water pump power cabinet or emergency generator; Step 4: According to the emergency rescue plan, a start command of the water pump body module is issued through the control console or the remote control monitoring system terminal or the crawler pump truck control module to start the submersible pump; Step 5: Monitor the drainage level and pump flow rate as well as the pump operating parameters; Step 5.1: Obtain water level signal data of the pumping object through a liquid level meter set at the bottom of the pumping object; Step 5.2: Collect the pumping and drainage flow data of the water pump through the flow meter; Step 5.3: Obtain the water flow rate of the water pump and the electrical parameters of the water pump motor through the control console or the crawler pump truck control module; Step 5.4: Use the vibration displacement sensors installed on the water pump and the water pipe to obtain vibration signal data of the water pump and the water pipe; Step 6: Acquire real-time meteorological condition data, and use the meteorological condition data, the pump operating parameters obtained in step 5, and the current pumping and drainage level as inputs of the first prediction model to obtain the pumping and drainage level in the future period; Step 7: Based on the pumping and drainage level in the future period obtained in step 6 and the emergency rescue plan, determine whether the emergency rescue task can be completed on time. If not, control the increase of the water pumping flow rate or execute steps 1-4 to add a water pump body module and the corresponding water pipe and joint module; Step 8: Use the meteorological condition data and the pumping and drainage levels and pump operating parameters of the future time period obtained in step 6 as inputs of the second prediction model, use the second prediction model to predict the possibility of pump failure, that is, the probability of pump failure, and adjust the emergency rescue pumping and drainage plan according to the possibility of pump failure to ensure that the emergency rescue task is completed on time.

[0016] Compared with the prior art, the beneficial effects of the present invention include: 1) The present invention modularizes the water pump body, water pipe and joint, cable winch, and operating console, and integrates the water pump power cabinet, control cabinet, and control console into a container, thereby realizing the rapid transportation, installation, and operation of the pumping and drainage device of the hydropower station, and improving the transportation, installation, and pumping and drainage efficiency of the pumping and drainage device of the hydropower station.

[0017] 2) The present invention provides a tracked pump truck and an emergency pump truck control end cart, which can realize emergency drainage of hydropower stations and are deployed using tracked pump trucks, thereby improving mobility and flexibility in complex terrain and emergency situations.

[0018] 3) The present invention uses AI models to predict water pump failures based on water pump operating parameters and meteorological conditions data, making it easier to identify potential water pump failures as early as possible and take corresponding measures, effectively avoiding damage to the pump body due to electrical failures and preventing economic losses caused by untimely water pumping and drainage in hydropower stations.

[0019] 4) The present invention uses the AI ​​model to predict the pumping and drainage level at the pump site according to the pump operating parameters, current pumping and drainage level and meteorological conditions data, so as to facilitate real-time adjustment of the pumping and drainage plan of the hydropower station and ensure timely completion of the pumping and drainage task of the hydropower station.

[0020] 5) The drainage level prediction model of the present invention adopts a bidirectional gated recurrent unit with an attention mechanism, which can effectively capture the time dependence and dynamic changes in the water pump operating parameters and meteorological environment data, identify the nonlinear relationship between the water pump on-site drainage level and the water pump parameters, meteorological conditions and other data, and improve the accuracy of the water pump on-site drainage level prediction results; the attention mechanism makes the drainage level prediction model pay more attention to the key features in the input data sequence, further improving the drainage level prediction accuracy.

[0021] 6) The present invention uses the drainage level calculated by the drainage level prediction model as the input of the water pump fault prediction model, which increases the input data of the water pump fault prediction model. The predicted drainage level of the drainage level prediction model is real-time and high-precision, which can improve the accuracy of the water pump fault prediction results.

[0022] 7) The emergency rescue pumping control method provided by the present invention realizes the rapid transportation and on-site assembly of equipment through modular design of the water pump body module, counterweight module, water pipe and joint module, greatly shortening the response time of emergency rescue; it provides a variety of water pump control modes, making the water pump control operation more flexible and convenient, and the most appropriate control mode can be selected according to actual conditions; it monitors the pumping and drainage level, water pump flow and water pump operating parameters in real time to ensure accurate control and timely adjustment of the pumping process.

[0023] 8) The emergency rescue and drainage control method of the present invention collects and analyzes data such as real-time meteorological conditions, water pump operating parameters, and current drainage levels, and uses the first prediction model to predict the drainage levels in future time periods, providing a scientific basis for emergency rescue decisions; according to the prediction results and the emergency rescue plan, the water pumping flow rate of the water pump can be adjusted in time or a water pump module can be added to ensure that the rescue task is completed on time. The possibility of water pump failure is predicted by the second prediction model, and potential failures are discovered in advance, reducing the impact of water pump failures on rescue work; the emergency rescue drainage plan is adjusted according to the possibility of water pump failure, which improves the stability and reliability of the rescue work. Through intelligent control and data-driven decision-making, resource allocation is optimized, emergency rescue efficiency is improved, and rescue costs are reduced.

[0024] 9) The drainage monitoring system of the present invention realizes the monitoring of the start and stop status of the water pump, the drainage flow rate and the operating status of the pump motor; the drainage level prediction model is used to realize the prediction of the drainage level of the drainage object; the water pump fault prediction model is used to realize the accurate prediction of the water pump fault; and the alarm signal is issued in time for the water pump fault. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0026] Figure 1 It is a schematic diagram of the connection between the water pump body module, the water delivery pipe and the counterweight module according to an embodiment of the present invention.

[0027] Figure 2 It is a structural diagram of a water pump body module according to an embodiment of the present invention.

[0028] Figure 3 Schematic diagram of a containerized water pump integrated control station according to an embodiment of the present invention.

[0029] Figure 4 Schematic diagram of a crawler pump truck according to an embodiment of the present invention.

[0030] Figure 5 It is a connection diagram of the crawler pump truck and the emergency pump truck control end trolley according to an embodiment of the present invention.

[0031] Figure 6 Schematic diagram of the structure of the first prediction model of an embodiment of the present invention.

[0032] Figure 7 It is a schematic diagram of signal communication of a modular pumping and drainage system of a large hydropower station according to an embodiment of the present invention.

[0033] Figure 8 It is a schematic diagram of an on-site control unit of a control cabinet according to an embodiment of the present invention. DETAILED DESCRIPTION

[0034] Embodiment 1 like Figure 1-3 As shown, the modular pumping and drainage system of a large hydropower station includes a water pump body module 1, a counterweight module 2, a water pipe and joint module, and a pumping and drainage control and monitoring device.

[0035] The water pump body module 1 includes a submersible pump 101 and a rectangular frame 102 for supporting the submersible pump, a hook connection device 103 is provided on the top of the rectangular frame, a plurality of universal wheels 104 with braking function are provided on the bottom of the rectangular frame, and an underwater camera 105 is provided on the top of the rectangular frame.

[0036] The counterweight module 2 is used to fix the water outlet of the water pipe; the outlet of the counterweight module is provided with a back pressure valve 201 and a flow meter 202 for increasing the flow of the water pump; The water pipe and connector module includes a water hose 3 and a pipe connector capable of quick lap-fastening; The pumping and drainage control and monitoring device is electrically connected to the water pump body module and is used to start and stop the water pump, control the water pump motor, and monitor the water pump status and water pipe flow, electrical parameters of the water pump motor, and water level of the pumping and drainage object.

[0037] The pumping and drainage control and monitoring device comprises a container-type integrated water pump control station 4 and a remote control monitoring system terminal connected to the control station in communication.

[0038] The containerized integrated water pump control station 4 comprises a water pump power cabinet 401, a control cabinet 402, a cable winch 5 and a control console 403 arranged in a container 404; the control cabinet 402 comprises a PLC controller which is communicatively connected with the control console.

[0039] The remote control monitoring system terminal is used to remotely control the start and stop of the submersible pump of the pump body module, and monitor the drainage parameters and electrical parameters of the submersible pump motor.

[0040] In an embodiment, the pumping and drainage control and monitoring device also includes a first prediction model for predicting the pumping and drainage level, namely, a pumping and drainage level prediction model. The first prediction model adopts a bidirectional gated recurrent unit BiGRU with an attention mechanism.

[0041] like Figure 6 As shown in the figure, the bidirectional gated recurrent unit BiGRU with attention mechanism includes an input layer, a BiGRU layer, an attention unit and an output layer. The attention unit includes a Softmax layer, a normalization layer and a Σ summation layer.

[0042] The calculation process of the predicted output of BiGRU with attention mechanism is: Calculate the hidden state vector Similarity score with the query vector ; In the formula Indicates i A hidden state vector, formed by the concatenation of the forward hidden state and the backward hidden state; Q The query vector representing the attention mechanism is a trainable parameter vector, which is randomly initialized when the BiGRU model is initialized and optimized through back-propagation.

[0043] Calculate the hidden state vector Similarity score The index value of ;

[0044] In the formula e Represents the base of the natural exponential function; Hidden State The attention weight ; Prediction Output ; In the formula Represents the output of BiGRU with attention mechanism.

[0045] In an embodiment, the pumping and drainage control and monitoring device also includes a second prediction model for predicting water pump failure. An autoencoder is used to construct the second prediction model, namely, a water pump failure prediction model. The input of the second prediction model is water pump operating parameters, characteristic data of vibration signals of the water pump and water pipe, predicted pumping and drainage levels and meteorological conditions data. The output of the second prediction model is the possibility of water pump failure, namely, the probability of water pump failure.

[0046] In the embodiment, an integrated control console is used in the containerized water pump integrated control station, which is fixedly installed in the container. The information displayed on the control console screen is reasonably distributed, and an integrated control method is adopted to facilitate information sharing and efficient communication between operators. It can quickly process various data information of the pumping and drainage equipment; the power supply, communication and other transmissions should be stable and reliable; the data information transmission of each part does not occupy or interfere with each other.

[0047] In the embodiment, the containerized water pump integrated control station is connected to the server and the host computer in the monitoring room, such as Figure 7 The containerized pump integrated control station is equipped with a local control unit, such as Figure 8 shown.

[0048] The control mode of the control console in the container combined with tablet computers and mobile terminals has two control modes: manual control and automatic control. In the automatic control mode, the user-friendly human-machine interface is used to monitor the equipment operation data in real time, to ensure that each device can operate independently and reliably, to achieve centralized control, and to automatically adjust the key parameters of the equipment; to achieve the detection, storage and monitoring of the operation parameters of the pumping and drainage equipment; to display, record, alarm and query fault information; to detect the electrical data of the pump motor and to automatically stop the pump; thus achieving automatic control of the entire pumping and drainage system.

[0049] The main functions of the automatic control system of the drainage control and monitoring device of the embodiment include: ① Monitoring and display of water pump operation status data, including: monitoring and display of flow rate, inlet and outlet pressure, flow rate and other data.

[0050] ② Monitoring and display of pump motor operating status data, including: voltage, current, temperature, operating time and other parameters when the motor is running, and the start / stop working status of the motor.

[0051] ③Electric energy monitoring data of the drainage power supply system, including: various electrical parameters, motor overvoltage, overload, temperature rise, phase loss, vibration protection; and other necessary protection functions.

[0052] It has the function of setting the upper and lower limits of the alarm, and alarms will be issued when the upper and lower limits are exceeded. It mainly includes: (1) display, record and query of operating faults; (2) when a fault occurs, the equipment operation record curve is used to analyze and judge the fault, identify the fault type, and issue an alarm, including: motor overtemperature, overload, and short circuit fault alarms; (3) It has sound and light alarm functions on site.

[0053] Recording and reporting function: Recording the historical data of equipment operation and forming reports, mainly including: (1) operation log recording function; (2) operating parameter and electrical parameter recording, automatic generation of trend curve; (3) fault and alarm recording function; (4) report generation and downloading function.

[0054] Underwater communication transmission: A reliable waterproof power communication composite cable mode is adopted. 4G / 5G wireless communication is adopted for water remote monitoring, and the direct connection interface of the communication cable is retained. The communication protocols are compatible with each other. If the 4G / 5G signal is not good, the signal is transmitted through the communication cable.

[0055] Embodiment 2 The difference between the large-scale hydropower station modular pumping and drainage system of the first embodiment and the large-scale hydropower station modular pumping and drainage system of the second embodiment is that the large-scale hydropower station modular pumping and drainage system of the second embodiment also includes a crawler pump truck 6 and an emergency pump truck control end trolley 7 connected to the crawler pump truck 6 and the ... Figure 4 and Figure 5 shown.

[0056] like Figure 4 As shown, a crawler walking mechanism 601 is provided at the bottom of the crawler pump truck 6 for carrying the water pump body module, and the movement of the crawler pump truck is controlled by remote control.

[0057] like Figure 5 As shown, the emergency pump truck control end trolley 7 includes a chassis and a cable winch 5 arranged on the chassis, a water pump start and stop control module 701 and a tracked pump truck control module. A plurality of universal wheels are arranged at the bottom of the chassis. The tracked pump truck control module is used to control the tracked pump truck and is communicatively connected with the tracked pump truck; the water pump start and stop control module 701 is electrically connected to the emergency generator 8.

[0058] The emergency rescue drainage control method includes the following steps: Step 1: Transport the water pump body module, counterweight module, water pipe and joint module to the rescue site; Step 2: Use a crawler pump truck to deploy the water pump body module to the emergency rescue drainage site; Step 3: Connect the water pump body module, water pipe and joint module to each other, and electrically connect the water pump body module to the water pump power cabinet or emergency generator; Step 4: According to the emergency rescue plan, a start command of the water pump body module is issued through the control console or the remote control monitoring system terminal or the crawler pump truck control module to start the submersible pump; Step 5: Monitor the drainage level and pump flow rate as well as the pump operating parameters; Step 5.1: Obtain water level signal data of the pumping object through a liquid level meter set at the bottom of the pumping object; Step 5.2: Collect the pumping and drainage flow data of the water pump through the flow meter; Step 5.3: Obtain the water flow rate of the water pump and the electrical parameters of the water pump motor through the control console or the crawler pump truck control module; Step 5.4: Use the vibration displacement sensors installed on the water pump and the water pipe to obtain vibration signal data of the water pump and the water pipe; Step 6: Acquire real-time meteorological condition data, and use the meteorological condition data, the pump operating parameters obtained in step 5, and the current pumping and drainage level as inputs of the first prediction model to obtain the pumping and drainage level in the future period; Step 7: Based on the pumping and drainage level in the future period obtained in step 6 and the emergency rescue plan, determine whether the emergency rescue task can be completed on time. If not, control the increase of the water pumping flow rate or execute steps 1-4 to add a water pump body module and the corresponding water pipe and joint module; Step 8: Use the meteorological condition data and the pumping and drainage levels and pump operating parameters of the future time period obtained in step 6 as inputs of the second prediction model, use the second prediction model to predict the possibility of pump failure, that is, the probability of pump failure, and adjust the emergency rescue pumping and drainage plan according to the possibility of pump failure to ensure that the emergency rescue task is completed on time.

Claims

1. Modular pumping and drainage system for large hydropower stations, characterized by: include: A water pump body module, comprising a submersible pump and a rectangular parallelepiped frame for supporting the submersible pump, wherein a hook connection device is provided at the top of the rectangular parallelepiped frame, and a plurality of universal wheels with a brake function are provided at the bottom of the rectangular parallelepiped frame; A counterweight module is used to fix the water outlet of the water pipe; the outlet of the counterweight module is provided with a back pressure valve and a flow meter for increasing the flow of the water pump; A water pipe and joint module, including a water pipe and a pipe joint capable of quick lap-joint fastening; The pumping and drainage control and monitoring device is electrically connected to the water pump body module and is used to start and stop the water pump, control the water pump motor, and monitor the water pump status and water pipe flow, electrical parameters of the water pump motor, and water level of the pumping and drainage object.

2. The modular pumping and drainage system of a large hydropower station according to claim 1 is characterized in that: The drainage control and monitoring device includes a container-type integrated water pump control station, which includes a water pump power cabinet, a control cabinet, a cable winch and a control console; the control cabinet includes a PLC controller that is communicatively connected to the control console.

3. The modular pumping and drainage system for a large hydropower station according to claim 2 is characterized in that: The drainage control and monitoring device also includes a remote control monitoring system terminal that is communicatively connected to the containerized water pump integrated control station, and is used to remotely control the start and stop of the submersible pump of the water pump body module, and monitor drainage parameters and electrical parameters of the submersible pump motor.

4. The modular pumping and drainage system for a large hydropower station according to claim 2 or 3, characterized in that: The drainage system also includes a tracked pump truck that can be quickly moved and deployed. A tracked walking mechanism is provided at the bottom of the tracked pump truck for carrying a water pump body module. The movement of the tracked pump truck is controlled by remote control.

5. The modular pumping and drainage system for a large hydropower station according to claim 4 is characterized in that: The drainage system also includes an emergency pump truck control end cart, which includes a chassis and a cable winch module, a water pump start and stop control module and a tracked pump truck control module arranged on the chassis. A plurality of universal wheels are arranged at the bottom of the chassis. The tracked pump truck control module is used to control the tracked pump truck and is communicatively connected with the tracked pump truck.

6. The modular pumping and drainage system for large hydropower stations according to claim 1, 2, 3 or 5, characterized in that: The pumping and drainage control and monitoring device also includes a first prediction model for predicting the pumping and drainage level. The first prediction model is constructed using a bidirectional gated recurrent unit BiGRU. The input of the first prediction model is the pumping and drainage object, water pump operation data, current pumping and drainage level and meteorological condition data. The output of the first prediction model is the predicted pumping and drainage level.

7. The modular pumping and drainage system for a large hydropower station according to claim 6 is characterized in that: The first prediction model introduces the attention mechanism in BiGRU. The calculation process of the prediction output with the attention mechanism is: Calculate the hidden state vector Similarity score with the query vector ； In the formula Indicates i A hidden state vector, formed by the concatenation of the forward hidden state and the backward hidden state; Q The query vector representing the attention mechanism; Calculate the hidden state vector Similarity score The index value of ; ； In the formula e Represents the base of the natural exponential function; Hidden State The attention weight ； Prediction Output ； In the formula Represents the output of BiGRU with attention mechanism.

8. The modular pumping and drainage system for a large hydropower station according to claim 6 is characterized in that: The pumping and drainage control and monitoring device also includes a second prediction model for predicting water pump failure. The second prediction model is constructed using an autoencoder. The input of the second prediction model is water pump operating parameters, water pump and water pipe vibration signal characteristic data, predicted pumping and drainage level and meteorological conditions data. The output of the second prediction model is the possibility of water pump failure, that is, the probability of water pump failure.

9. The emergency rescue pumping and drainage control method of the modular pumping and drainage system of a large hydropower station according to claim 7 or 8, characterized in that: The following steps are involved: Step 1: Transport the water pump body module, counterweight module, water pipe and joint module to the rescue site; Step 2: Use a crawler pump truck to deploy the water pump body module to the emergency rescue drainage site; Step 3: Connect the water pump body module, water pipe and joint module to each other, and electrically connect the water pump body module to the water pump power cabinet or emergency generator; Step 4: According to the emergency rescue plan, a start command of the water pump body module is issued through the control console or the remote control monitoring system terminal or the crawler pump truck control module to start the submersible pump; Step 5: Monitor the drainage level and pump flow rate as well as the pump operating parameters; Step 5.1: Obtain water level signal data of the pumping object through a liquid level meter set at the bottom of the pumping object; Step 5.2: Collect the pumping and drainage flow data of the water pump through the flow meter; Step 5.3: Obtain the water flow rate of the water pump and the electrical parameters of the water pump motor through the control console or the crawler pump truck control module; Step 5.4: Use the vibration displacement sensors installed on the water pump and the water pipe to obtain vibration signal data of the water pump and the water pipe; Step 6: Acquire real-time meteorological condition data, and use the meteorological condition data, the pump operating parameters obtained in step 5, and the current pumping and drainage level as inputs of the first prediction model to obtain the pumping and drainage level in the future period; Step 7: Based on the pumping and drainage levels in the future period obtained in step 6 and the emergency rescue plan, determine whether the emergency rescue task can be completed on time. If not, control the increase of the water pump flow rate or execute steps 1-4 to add a water pump body module and the corresponding water pipe and joint module.

10. The emergency rescue and drainage control method according to claim 9, characterized in that: The emergency rescue pumping and drainage control method also includes step 8: using the meteorological condition data and the pumping and drainage level in the future time period obtained in step 6 and the water pump operating parameters as inputs of the second prediction model, using the second prediction model to predict the possibility of water pump failure, that is, the probability of water pump failure, and adjusting the emergency rescue pumping and drainage plan according to the possibility of water pump failure to ensure that the emergency rescue task is completed on time.

Images