Intelligent self-adaptive silt filtering system and method for pumped storage power station

Abstract

The invention discloses an intelligent self-adaptive silt filtering system and method for a pumped storage power station, and belongs to the technical field of pumped storage power stations. The system comprises a multi-parameter water quality monitoring module, an intelligent self-adaptive filtering module, a hydraulic cyclone enhanced separation module, a self-cleaning and backwashing module and a central control and data processing module. According to the method, river sediment characteristics are monitored in real time, filtering parameters are dynamically adjusted through an artificial intelligence algorithm, efficient sediment interception is achieved by combining hydraulic cyclone separation and multi-stage filtering, the problem of filtering medium blockage is solved through a self-cleaning mechanism, and meanwhile the self-adaptive learning capacity is achieved so as to optimize system performance. The system can significantly improve the filtering efficiency, reduce energy consumption and operation and maintenance cost, adapt to complex water quality working conditions, and guarantee safe and stable operation of key equipment of the pumped storage power station.

Description

Technical Field

[0001] The present invention relates to the technical field of pumped storage power stations, and in particular to an adaptive sediment filtering system and method for a pumped storage power station. Background Art

[0002] During the operation of a pumped storage power station, sediment in the river water can cause severe wear and tear on key components such as turbine blades and pump impellers, leading to reduced equipment efficiency, shortened lifespan, and even safety accidents. Currently, river sediment treatment technology has the following shortcomings:

[0003] Traditional filtration technologies (such as sedimentation basins and sand barriers) occupy large areas, have low filtration efficiency, are limited in intercepting fine sediment with a particle size of less than 0.1mm, and cannot adjust filtration parameters in real time according to changes in water quality.

[0004] The lack of intelligent monitoring and feedback mechanisms makes it difficult to monitor parameters such as sediment concentration and particle size distribution in real time, resulting in insufficient filtration capacity at high sediment concentrations and energy waste at low sediment concentrations.

[0005] The filter medium is easily clogged and requires frequent shutdown for cleaning or replacement, affecting the continuous operation of the power station;

[0006] It has poor adaptability to special working conditions (such as a sudden increase in sediment content during flood season) and cannot quickly adjust the filtration strategy.

[0007] Therefore, there is an urgent need for a new filtration system and method that can monitor water quality in real time, intelligently adjust parameters, efficiently intercept sediment, and have strong anti-clogging capabilities. Summary of the Invention

[0008] The purpose of the present invention is to propose an intelligent adaptive sediment filtration system and method for pumped storage power stations to solve the problems of low filtration efficiency, poor adaptability, and easy clogging in the existing technology, achieve precise interception of sediment of different particle sizes, and at the same time have self-cleaning and adaptive learning capabilities to ensure safe and efficient operation of the power station.

[0009] In order to achieve the above object, the present invention provides the following technical solutions:

[0010] An intelligent adaptive sediment filtration system for a pumped storage power station, comprising the following modules:

[0011] Multi-parameter water quality monitoring module, used to collect river sediment characteristics and water quality parameters in real time, including:

[0012] Ultrasonic particle size analyzer: Based on the principle of multi-frequency ultrasonic scattering, it measures the particle size distribution of sediment particles in real time (range 0.01μm-10mm);

[0013] Laser turbidity meter: measures water turbidity by laser scattering and calculates sediment concentration (range 0-10000 NTU);

[0014] Pressure sensor array: arranged at the inlet and outlet of the filtration system and key locations to monitor changes in water flow pressure and determine the clogging status of the filter medium;

[0015] Water chemical analyzer: detects water pH, hardness and other chemical parameters to assist in determining sediment sedimentation characteristics.

[0016] Intelligent adaptive filtering module, used to dynamically adjust filtering parameters based on monitoring data to achieve accurate filtering, including:

[0017] Multi-stage composite filtration unit, including:

[0018] Primary filter layer: uses a large-pore (1-5mm) stainless steel wedge-shaped filter to intercept large particles of sediment and debris;

[0019] Intermediate filtration layer: composed of micro-nano fiber filter membrane with adjustable pore size, the pore size can be dynamically adjusted between 10μm-100μm;

[0020] Advanced filtration layer: uses nano-ceramic membrane (pore size less than 0.1μm) to intercept colloidal particles and very fine sediment;

[0021] Dynamic pore size adjustment device: Based on a shape memory alloy (SMA) actuator, it automatically adjusts the pore size of the micro-nano fiber filter membrane according to the sediment particle size distribution;

[0022] Variable porosity filter medium: Made of intelligent hydrogel material, the porosity automatically changes with water pressure and sediment concentration (range 20%-80%).

[0023] The hydrocyclone enhanced separation module is used to pre-separate sediment through centrifugal force to improve filtration efficiency, including:

[0024] Multi-stage hydrocyclone: ​​It adopts a tapered cylinder structure (diameter decreases from 300mm to 100mm step by step). It generates centrifugal force through high-speed rotating water flow to separate the sediment particles into the sand discharge pipe.

[0025] Intelligent flow guide device: Based on real-time monitoring data, the electric actuator dynamically adjusts the cyclone inlet flow rate (5-15m / s) and flow guide angle (0°-60°) to optimize separation efficiency.

[0026] Self-cleaning and backwashing module to prevent filter media clogging and ensure continuous system operation, including:

[0027] Pulse backwash system: adopts high-pressure water pulse technology, the backwash pressure can be dynamically adjusted between 0.5-2MPa, the pulse frequency is 1-10Hz, and the pulse width is 50-500ms;

[0028] Ultrasonic-assisted cleaning device: A piezoelectric ceramic transducer is placed on the surface of the filter medium to assist in removing attached sediment through 20-80kHz high-frequency vibration (power density 0.5-2W / cm²);

[0029] Electrochemical anti-scaling device: uses a titanium-based coated anode and applies a weak current of 1-5mA to prevent calcium and magnesium ions from depositing on the surface of the filter medium to form scale.

[0030] The central control and data processing module is used to integrate and control various modules and implement data analysis and optimization, including:

[0031] Industrial-grade PLC controller: Adopting Siemens S7-1500 series, with a processing speed of no less than 1000 points / second and integrated monitoring, analysis and control functions;

[0032] Artificial intelligence algorithm module: Developed based on the TensorFlow framework, it uses deep neural networks (DNN) and fuzzy logic control algorithms to establish a mapping relationship model between sediment characteristics and filtration parameters;

[0033] Big data storage and analysis system: uses MySQL database (storage capacity not less than 10TB) to store historical monitoring data and operating parameters, supporting trend analysis and predictive maintenance.

[0034] The intelligent adaptive sediment filtration method for a pumped storage power station of the present invention comprises the following steps:

[0035] Real-time monitoring and data analysis

[0036] The multi-parameter water quality monitoring module collects parameters such as sediment particle size distribution, concentration, pH value, and pressure at a set frequency (once every 10 seconds for the ultrasonic particle size analyzer, real-time for the laser turbidity analyzer, and once every 30 seconds for the water quality chemical analyzer);

[0037] The data is transmitted to the central control and data processing module. After preprocessing (denoising, normalization) and feature extraction (wavelet transform, Fourier transform), it is input into the DNN model to predict the changing trend of sediment characteristics in the next 10-30 minutes.

[0038] Smart filtering parameter adjustment

[0039] The central control and data processing module calculates the optimal filtering parameters using fuzzy logic control algorithms based on real-time data and prediction results;

[0040] Control the dynamic aperture adjustment device (SMA driver) to adjust the aperture of the micro-nano fiber filter membrane, and use PID control to ensure the adjustment accuracy;

[0041] Adjust the hydrocyclone inlet flow rate and diversion angle to optimize centrifugal separation efficiency;

[0042] When the sediment concentration exceeds 5000NTU or the proportion of sediment with a particle size less than 0.1μm exceeds 30%, the advanced filtration layer (nano-ceramic membrane) is automatically activated.

[0043] Self-cleaning and backwash control

[0044] Real-time monitoring of the inlet and outlet pressure difference of the filtration system. When it exceeds the set threshold (0.05MPa for the primary filtration layer, 0.1MPa for the intermediate layer, and 0.2MPa for the advanced layer), it is judged as blocked and the backwash procedure is started;

[0045] According to the degree of blockage and sediment characteristics, the pressure, frequency and duration of the pulse backwash system are adjusted through the fuzzy control algorithm, and the ultrasonic auxiliary cleaning device is started at the same time (cleaning time 30-120 seconds);

[0046] During the backwash process, the pressure difference is monitored in real time and cleaning is stopped when it returns to the normal range.

[0047] System optimization and adaptive learning

[0048] Record the operating parameters (such as pore size, flow rate, backwash pressure) and effect data (such as filtration efficiency, pressure difference change) of each filtration process and store them in the big data system;

[0049] Optimize the DNN model weight parameters through reinforcement learning algorithms every week to improve the system's adaptability to different water qualities;

[0050] Automatically adjust strategies based on power plant operating conditions and water quality patterns (e.g., increase monitoring frequency and filtration accuracy during flood season, and reduce energy consumption during dry season).

[0051] This technology has the following advantages and beneficial effects:

[0052] 1. Improved filtration efficiency: Through real-time monitoring and dynamic parameter adjustment, it can accurately intercept sediments of different particle sizes, and the filtration efficiency is increased by 30%-50% compared with traditional systems;

[0053] 2. Enhanced anti-clogging ability: Pulse backwashing combined with ultrasonic cleaning and intelligent hydrogel media can extend the backwash cycle by 2-3 times and reduce water consumption by 40%;

[0054] 3. Reduced energy consumption: Optimized filtration parameters and intelligent control can reduce system energy consumption by 20%-30%, thus reducing water waste;

[0055] 4. Adaptability to complex working conditions: It can quickly respond to changes in water quality, maintain stable filtration performance under special working conditions such as floods, and ensure the safety of the power station;

[0056] 5. Reduced operation and maintenance costs: Predictive maintenance is achieved based on big data and artificial intelligence, providing early warning of failures and filter media life, reducing operation and maintenance costs;

[0057] 6. Easy to integrate: The system has a compact structure and occupies a small area. It can be integrated with the existing water supply system with low transformation cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 Schematic diagram of the system structure of the present invention;

[0059] Figure 2 It is a control flow chart of the present invention. DETAILED DESCRIPTION

[0060] The present invention is further described below with reference to the embodiments. It should be noted that these are merely examples and illustrations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the claims, they should be deemed to fall within the scope of protection of the present invention.

[0061] Example 1:

[0062] like Figure 1 As shown, the present invention proposes an intelligent adaptive sediment filtration system for a pumped storage power station, comprising the following devices:

[0063] Multi-parameter water quality monitoring module:

[0064] An ultrasonic particle size analyzer is installed on the water inlet pipe, utilizing dual-frequency ultrasonic transmission technology (2 MHz and 10 MHz), collecting data every 10 seconds. A laser turbidity meter utilizes a 90° scattering method with a measurement accuracy of ±0.1 NTU. It is installed 5-10 meters from the ultrasonic particle size analyzer. The pressure sensor array utilizes high-precision strain gauge pressure sensors with a measurement range of 0-1 MPa and an accuracy of ±0.05%. Sensors are located at the inlet and outlet of the filtration system and before and after each filter layer. A water chemical analyzer utilizes an ion-selective electrode method for real-time monitoring of pH and hardness, with a measurement cycle of 30 seconds.

[0065] Intelligent adaptive filtering module:

[0066] The primary filtration layer utilizes a 316L stainless steel wedge-shaped filter with a 2mm wire diameter and a 1mm slit width, capable of intercepting sediment and debris larger than 1mm in size. The secondary filtration layer comprises a shape memory alloy (SMA) actuator and a micro-nano fiber filter membrane. The SMA actuator, made of nickel-titanium alloy, has a response time of less than 1 second and allows for stepless adjustment of the membrane pore size from 10μm to 100μm. The advanced filtration layer utilizes an alumina-based nano-ceramic membrane with an average pore size of 50nm and a porosity of 40%-50%. Its operation is controlled by a pneumatic valve. The variable-porosity filter medium utilizes a polyacrylamide hydrogel material, whose porosity automatically adjusts between 20% and 80% based on water pressure.

[0067] Hydrocyclone enhanced separation module:

[0068] The multi-stage hydrocyclone utilizes a tapered cylinder structure, with the diameter gradually decreasing from 300mm to 100mm, forming a four-stage separation system. The intelligent flow guide device uses an electric actuator to control the guide vane angle, with an adjustment range of 0°-60° and a response time of less than 5 seconds. A PID controller automatically adjusts the cyclone's inlet flow rate based on real-time monitoring of the sediment particle size distribution, with an optimal flow rate range of 5-15m / s.

[0069] Self-cleaning and backwashing module:

[0070] The pulsed backwash system utilizes a high-pressure water pump and electromagnetic pulse valve. The backwash pressure is adjustable between 0.5 and 2 MPa, with a pulse frequency of 1 to 10 Hz and a pulse width of 50 to 500 ms. The ultrasonic-assisted cleaning device utilizes a piezoelectric ceramic transducer with a frequency range of 20 to 80 kHz and a power density of 0.5 to 2 W / cm², evenly distributed across the filter media surface. The electrochemical anti-scaling device utilizes a titanium-coated anode, applying a weak current of 1 to 5 mA to prevent the deposition of calcium and magnesium ions on the filter media surface.

[0071] Central control and data processing module:

[0072] The system uses a Siemens S7-1500 series PLC as the core controller, equipped with 16GB of memory and a 512GB solid-state drive, with a processing speed of no less than 1,000 points per second. The artificial intelligence algorithm module is developed based on the TensorFlow framework and utilizes a deep neural network (DNN) model. Input parameters include sediment particle size distribution, concentration, pH value, and pressure differential, while output parameters include filter media pore size, backwash pressure, and cyclone inlet flow rate. The big data storage and analysis system utilizes a MySQL database with a storage capacity of no less than 10TB, supporting historical data query, trend analysis, and report generation.

[0073] Example 2,

[0074] The present invention also discloses an intelligent adaptive sediment filtering method for a pumped storage power station, comprising the following steps:

[0075] (1) Real-time monitoring and data analysis;

[0076] After the system is started, the multi-parameter water quality monitoring module synchronously collects data at the set sampling frequency. The collected data first undergoes preprocessing, including noise removal and data normalization. Wavelet and Fourier transforms are used to extract features from the data and obtain characteristic parameters of the sediment particle size distribution. These parameters are then fed into a pretrained DNN model to predict sediment characteristics over the next 10-30 minutes.

[0077] (2) Intelligent filtering parameter adjustment;

[0078] Based on real-time monitoring data and prediction results, the central control and data processing module calculates optimal filtration parameters using a fuzzy logic control algorithm. This controls the SMA actuator to adjust the pore size of the micro-nano fiber filter membrane. The PID control algorithm is used during this adjustment process to ensure accuracy and stability. Simultaneously, the electric actuator adjusts the guide vane angle and inlet flow rate of the hydrocyclone to optimize sediment separation efficiency. The advanced filtration layer (nanoceramic membrane) is automatically activated when the sediment concentration exceeds 5000 NTU or the proportion of sediment particles smaller than 0.1 μm exceeds 30%.

[0079] (3) Self-cleaning and backwash control;

[0080] The pressure differential between the inlet and outlet of the filtration system is monitored in real time. When the pressure differential exceeds a set threshold (0.05 MPa for the primary filtration layer, 0.1 MPa for the intermediate filtration layer, and 0.2 MPa for the advanced filtration layer), the backwash process is initiated. Based on the degree of blockage and sediment characteristics, a fuzzy control algorithm automatically adjusts the pressure, frequency, and duration of the pulsed backwash system. Simultaneously, an ultrasonic-assisted cleaning device is activated, with the cleaning time determined by the degree of blockage, typically 30-120 seconds. During the backwash process, the pressure differential is monitored in real time and stopped when it returns to the normal range.

[0081] (4) System optimization and adaptive learning;

[0082] After each filtration process, operating parameters and performance data are recorded in a database. Historical data is analyzed regularly (weekly), and reinforcement learning algorithms are used to optimize the weight parameters of the DNN model to improve the system's adaptability to varying water quality conditions. Monitoring frequency and filtration strategies are automatically adjusted based on the power plant's operating conditions and water quality fluctuations. For example, monitoring frequency and filtration accuracy are increased during flood season, while monitoring frequency and energy consumption are reduced during dry season.

[0083] The above is an exemplary description of the invention. Obviously, the specific implementation of the present invention is not limited to the above-mentioned method. As long as such non-substantial improvements are made by adopting the method concept and technical solution of the present invention, or the concept and technical solution of the invention are directly applied to other occasions without improvement, they are all within the scope of protection of the present invention.

Claims

1. An intelligent adaptive sediment filtration system for a pumped storage power station, characterized in that: It includes multi-parameter water quality monitoring module, intelligent adaptive filtration module, hydrocyclone enhanced separation module, self-cleaning and backwash module and central control and data processing module; The multi-parameter water quality monitoring module includes an ultrasonic particle size analyzer, a laser turbidity meter, a pressure sensor array, and a water quality chemical analyzer, which are used to monitor the particle size distribution, concentration, pressure, and chemical properties of river water sediment in real time; The intelligent adaptive filtration module includes a multi-stage composite filtration unit, a dynamic pore size adjustment device and a variable porosity filter medium, which is used to intelligently adjust the filtration parameters according to the monitoring data; The hydrocyclone enhanced separation module includes a multi-stage hydrocyclone and an intelligent flow guide device for separating sediment particles by centrifugal force; The self-cleaning and backwashing module includes a pulse backwashing system, an ultrasonic-assisted cleaning device, and an electrochemical anti-scaling device to prevent clogging of the filter medium; The central control and data processing module includes an industrial-grade PLC controller, an artificial intelligence algorithm module and a big data storage and analysis system, which is used to integrate and control the above modules to achieve data analysis and system optimization.

2. The system according to claim 1, wherein: The multi-stage composite filtration unit comprises: Primary filter layer: stainless steel wedge filter with a pore size of 1-5mm; Intermediate filtration layer: uses a micro-nano fiber filter membrane with a pore size that can be dynamically adjusted between 10μm-100μm; Advanced filtration layer: uses nano-ceramic membrane with a pore size less than 0.1μm.

3. The system according to claim 1, wherein: The dynamic pore size adjustment device is based on a shape memory alloy driver with a response time of less than 1 second, and can automatically adjust the pore size of the micro-nano fiber filter membrane according to the sediment particle size distribution; The variable porosity filter medium uses intelligent hydrogel material, and its porosity can automatically change within the range of 20%-80% according to the water pressure and sediment concentration; The multi-stage hydrocyclone adopts a tapered cylindrical structure, with the diameter gradually decreasing from 300 mm to 100 mm; the intelligent diversion device adjusts the diversion angle and inlet flow rate through an electric actuator.

4. The system according to claim 1, wherein: The backwash pressure of the pulse backwash system can be adjusted between 0.5-2MPa, the pulse frequency is 1-10Hz, and the pulse width is 50-500ms; the frequency range of the ultrasonic assisted cleaning device is 20-80kHz, and the power density is 0.5-2W / cm².

5. The system according to claim 1, wherein: The artificial intelligence algorithm module adopts deep neural network and fuzzy logic control algorithm. The input parameters include sediment particle size distribution, concentration, pH value and pressure difference, and the output parameters include filter medium pore size, backwash pressure and cyclone inlet flow rate.

6. An adaptive sediment filtration method based on the system according to any one of claims 1 to 5, characterized in that: The following steps are involved: Step 1: Real-time monitoring and data analysis: The multi-parameter water quality monitoring module collects parameters and transmits them to the central control module. After preprocessing and feature extraction, the DNN model is used to predict the trend of sediment characteristics. Step 2: Intelligent filtration parameter adjustment: Based on real-time data and prediction results, adjust the pore size of the micro-nano fiber filter membrane, the inlet flow rate of the hydrocyclone, and the diversion angle, and activate the advanced filtration layer when necessary; Step 3: Self-cleaning and backwash control: real-time monitoring of pressure difference, starting the backwash process when it exceeds the threshold, adjusting backwash parameters and combining ultrasonic assisted cleaning; Step 4: System optimization and adaptive learning: record operating data, regularly optimize the model through reinforcement learning algorithms, and adjust monitoring frequency and filtering strategies.

7. The method according to claim 6, characterized in that In step 1, the ultrasonic particle size analyzer samples once every 10 seconds, the laser turbidity analyzer samples in real time, and the water quality chemical analyzer samples once every 30 seconds; the DNN model predicts changes in sediment characteristics in the next 10-30 minutes.

8. The method according to claim 6, characterized in that In step 2, when the sediment concentration exceeds 5000 NTU or the proportion of sediment with a particle size less than 0.1 μm exceeds 30%, the advanced filtration layer is automatically started; the pore size of the micro-nano fiber filter membrane is adjusted using a PID control algorithm.

9. The method according to claim 6, characterized in that In step 3, the pressure difference threshold is set as: 0.05 MPa for the primary filtration layer, 0.1 MPa for the intermediate filtration layer, and 0.2 MPa for the advanced filtration layer; the backwash time is 30-120 seconds, which is dynamically adjusted according to the degree of blockage.

10. The method according to claim 6, characterized in that In step 4, the DNN model weight parameters are optimized weekly through the reinforcement learning algorithm; the monitoring frequency and filtering accuracy are increased during the flood season, and the monitoring frequency and energy consumption are reduced during the dry season.

Images