Water-saving irrigation canal and adjusting and control method

Abstract

The invention relates to the technical field of irrigation canals, and discloses a water-saving irrigation canal and an adjusting and controlling method, the water-saving irrigation canal comprises a canal body, the top of the canal body is fixedly connected with a side supporting table, one side of the side supporting table is fixedly connected with a side supporting plate, and one side of the side supporting plate is fixedly connected with a side supporting frame; the liquid can be preliminarily filtered through the front filter screen, so that the condition that the water channel body is blocked by larger impurities is reduced, and the filtering effect is improved; sand in the liquid is filtered, so that impact caused by the sand during irrigation can be reduced, and the irrigation effect is improved; meanwhile, after the adjusting plate is lifted, the flow of water flow in the side supporting frame can be increased, if the adjusting plate is lowered, the flow of the water flow can be reduced through the side communicating holes in the surface of the adjusting plate, the flow is adjusted according to different environments and requirements, and the water saving performance of the canal body can be improved; the stirring blades capable of continuously stirring can be used for mixing liquid in the top mixing barrel, so that the follow-up pouring effect is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of irrigation canals, and in particular relates to a water-saving irrigation canal and a regulating and controlling method thereof. Background Art

[0002] Irrigation canals are artificially dug or constructed water channels, mainly used to guide water from water sources to farmland, orchards or other areas that need irrigation to meet the moisture needs of crop growth. Irrigation canals are an important part of agricultural irrigation systems and are of great significance for improving agricultural production efficiency, promoting crop growth and ensuring food security.

[0003] In the prior art, the authorization publication number "CN112262655A" discloses "an agricultural irrigation canal weed removal device"; it includes a mobile body, a first electric guide rail, a lifting assembly, a weed conveying assembly and a material storage assembly. The lifting assembly includes a first connecting plate, a first cylinder and a first motor. When in use, the first connecting plate is driven by the first electric guide rail to move the conveyor above the canal, and then the material storage box is driven by the second electric guide rail to move above the canal, and then the height of the conveyor descent is determined by the two first cylinders and the first motor according to the depth of the canal, and then the material storage box is lowered below the top of the conveyor by the two second cylinders, and then the conveyor drives the hook to take the weeds out of the canal, and as the conveyor rotates, the weeds will fall into the material storage box, saving a lot of time and improving work efficiency.

[0004] The above-mentioned "a weed removal device for agricultural irrigation canals" still has some disadvantages. For example, in the process of agricultural irrigation, the water flow often carries a lot of silt and debris, which will cause accumulation in the canal, affecting the size of the water flow, making it difficult to control the size of the water flow and causing wear and tear on the canal.

[0005] For this reason, a water-saving irrigation canal and a regulating and controlling method are proposed to solve the above-mentioned problems. Summary of the invention

[0006] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a water-saving irrigation canal and a regulation and control method, which effectively solves the problem that the water flow of the irrigation canals on the market often carries a lot of silt, and the water flow carries debris, and the debris and silt will cause accumulation in the canal, affecting the size of the water flow, making it difficult to control the size of the water flow and causing wear and tear on the canal.

[0007] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a water-saving irrigation canal and a regulation and control method, comprising a canal body, the top of the canal body is fixedly connected to a side support platform, one side of the side support platform is fixedly connected to a side support plate, one side of the side support plate is fixedly connected to a side support frame, the top of the side support frame is provided with a top connecting frame, one side of the side support plate is fixedly connected to a top limit platform, the inner side of the top limit platform is provided with a top threaded hole, the inner side of the top threaded hole is threadedly connected to a threaded rod, the bottom of the threaded rod is provided with an adjustment plate, and the surface of the adjustment plate is provided with a side connecting hole.

[0008] Preferably: the top of the adjustment plate is fixedly connected to a top sealing plate, the top of the top sealing plate is movably connected to a bottom bearing disk, the top of the bottom bearing disk is movably connected to the bottom of the threaded rod, and the top of the threaded rod is fixedly connected to a top adjustment handle.

[0009] Preferably: one side of the side support platform is fixedly connected to the top support platform, one side of the top support platform is fixedly connected to the top support plate, one side of the top support plate is fixedly connected to the bottom connecting frame, one side of the bottom connecting frame is fixedly connected to the side connecting plate, one side of the side connecting plate is fixedly connected to the connecting pipe, one side of the connecting pipe is provided with a water pump body, the bottom of the water pump body is fixedly connected to the top mixing barrel, the bottom of the top mixing barrel is fixedly connected to the filter box, and the inner side of the filter box is fixedly connected to the inner filter screen.

[0010] Preferably, a front filter is fixedly connected to one side of the water channel body, and one side of the front filter and one side of the side connecting plate are arranged parallel to each other along the surface of the water channel body.

[0011] Preferably, a top communicating hole is provided at the top of the filter box, and a bottom square frame is provided at the bottom of the filter box.

[0012] Preferably: a side discharge chute is provided on one side of the filter box, and a storage frame is provided on one side of the filter box.

[0013] Preferably, the top of the bottom connecting frame is fixedly connected to a supporting base frame, and the top of the supporting base frame is arranged at the bottom of the storage frame.

[0014] Preferably: the top of the top mixing barrel is fixedly connected to a top fixing barrel, the top of the top mixing barrel is fixedly connected to a motor body, the output shaft of the motor body is fixedly connected to a transmission rod, the surface of the transmission rod is fixedly connected to a stirring blade, and a top connecting hole is opened on the top of the top mixing barrel.

[0015] Preferably, one side of the side connecting plate is provided with a liquid inlet hole, and the number of the liquid inlet holes is two.

[0016] A method for regulating a water-saving irrigation canal of the present invention comprises the following steps:

[0017] S1. When the liquid flows along the water channel, it will contact the front filter screen and filter the larger debris through the front filter screen to block the larger debris. The filtered liquid will flow to one side of the side connecting plate through the front filter screen and accumulate on one side of the side connecting plate. The water pump body is started and the liquid on one side of the side connecting plate is driven by the water pump body to be pumped into the top mixing barrel along the connecting pipe. Through the opening of the top connecting hole, the liquid in the top mixing barrel will flow into the filter box with gravity;

[0018] S2. The liquid entering the filter box will contact the inner filter screen and be filtered through the inner filter screen with a smaller gap, blocking and filtering the smaller sand and gravel particles in the liquid. The inner filter screen is in an inclined state. After the sand and gravel particles stay on the surface of the inner filter screen, they will fall and roll along the inclined angle. The sand and gravel filtered on the surface of the inner filter screen will slide into the surface of the storage frame along the side discharge trough, and the sand and gravel will be uniformly stored through the storage frame on the surface of the supporting bottom frame. The filtered liquid in the filter box will flow downward along the bottom square frame and enter the water channel body to flow;

[0019] S3. The operator first holds the top adjustment handle and rotates it. The rotating top adjustment handle drives the threaded rod to rotate. The threaded rod is connected to the top threaded hole through a thread. During the threaded rotation of the threaded rod, the height of the top sealing plate is adjusted. The bottom end of the threaded rod is movably connected to the top sealing plate through a bottom bearing disk. The connection through the bottom bearing disk maintains the stability of the top sealing plate when the bottom bearing disk rotates. By adjusting the height of the top sealing plate, the height of the adjustment plate on the inner side of the top connecting frame is adjusted.

[0020] S4. Different microorganisms or liquid fertilizers can be added into the top mixing barrel through the top fixed barrel. The added liquid will enter the top mixing barrel along the top fixed barrel. At this time, the motor body is started to drive the transmission rod to rotate along the inner side of the top connecting hole. The rotating transmission rod will drive the stirring blade to stir along the inner side of the top mixing barrel.

[0021] A water-saving irrigation canal control method of the present invention optimizes the canal control based on a combination of deep reinforcement learning and quantum self-supervised contrastive learning algorithm. The specific process is as follows:

[0022] Step 1: System Initialization

[0023] Quantum Adaptive Hybrid Graph Neural Network Modeling

[0024] Digital modeling: In the canal system, in order to automate water flow regulation, the physical components are digitally modeled and converted into nodes in the graph neural network; in this model, each node represents the threaded rod, top sealing plate, water pump physical equipment and the physical characteristics of the equipment and their interactive relationship, which provides a mathematical basis for subsequent system optimization and regulation;

[0025] Threaded rod node: The key physical characteristics of the threaded rod include the rotation angle θ(t) and the rotation speed ω(t); these variables are monitored in real time by a rotation sensor mounted on the threaded rod;

[0026] Top sealing plate node: The key characteristics of the top sealing plate include its height h(t) and the water pressure p(t) it bears. The height of the top sealing plate directly affects the water flow f(t), while the water pressure p(t) reflects the pressure of the water flow on the top sealing plate.

[0027] The water pressure p(t) depends on the height of the top sealing plate and the water flow velocity, reflecting the force of the water flow on the system. These variables are calculated through a quantum adaptive hybrid graph neural network, and the interaction between nodes and edges simulates the dynamic relationship between the top sealing plate and the threaded rod.

[0028] Water pump node: The working state of the water pump and the water flow rate v(t) are the key control parameters; the working state of the water pump is represented by the binary state of the node, while the water flow rate is provided by real-time data;

[0029] The quantum adaptive hybrid graph neural network simulates the relationship between the pump state and the water flow change to predict the flow adjustment effect in real time;

[0030] Sensor nodes: The sensor nodes monitor the real-time data of water flow f(t), water pressure p(t), and temperature T(t); these data are transmitted to other key nodes, threaded rods, and top sealing plates through the connections between nodes to adjust their operations;

[0031] Sensor modeling: Each sensor node inputs real-time data into the deep reinforcement learning model and quantum self-supervised contrastive learning model, and adjusts the system control strategy by dynamically updating the node status;

[0032] The relationship between nodes and edges in graph neural networks

[0033] An important feature of quantum adaptive hybrid graph neural networks is that the edges between nodes represent the dynamic interactions between physical devices;

[0034] The edge of the threaded rod and the top sealing plate: The rotation of the threaded rod directly determines the height of the top sealing plate, which is determined by the edge weight w ij To express; the influence of the rotation speed ω(t) of the threaded rod on the height change Δh(t) of the top sealing plate is expressed by a linear relationship, and the weight of the edge is:

[0035] Δh(t)=w ij ω(t)

[0036] where w ij represents the coupling coefficient between the threaded rod and the top sealing plate, which determines the effect of each rotation of the threaded rod on the height of the top sealing plate, ω(t) is the rotation speed of the threaded rod, and Δh(t) is the height change of the top sealing plate;

[0037] By updating these weights, the system is able to accurately simulate the complex physical interactions in water flow regulation;

[0038] The edge between the water pump and the water flow sensor: The on / off state of the water pump affects the water flow, which is represented by the weight of the edge. The water flow f(t) depends on the working state of the water pump and the flow rate v(t):

[0039] f(t)=w ij v(t)

[0040] where w ij represents the efficiency coefficient of the water pump, describing the direct impact of the water pump state on the water flow rate, v(t) is the water pump flow rate, and f(t) is the water flow rate;

[0041] Edges between sensor nodes and control nodes: Sensor data is transmitted to control nodes through edges and used to adjust their operations; if the sensor detects that the water pressure is too high, the system adjusts the rotation speed ω(t) of the threaded rod through the weight of the edge to reduce the water flow;

[0042] Quantum computing accelerates parallel computing

[0043] When dealing with complex dynamic interactions in the canal system, quantum computing is introduced. The quantum adaptive hybrid graph neural network uses the superposition and entanglement states of quantum bits to process interactions between multiple nodes at the same time, significantly accelerating the computing process.

[0044] Quantum state superposition: In quantum computing, qubits are in a superposition state, which means that multiple states are calculated at the same time. In the process of adjusting the threaded rod and the top sealing plate, the system can simultaneously calculate multiple combinations of rotation speed ω(t) and height h(t) to find the optimal adjustment strategy;

[0045] Quantum state entanglement: Through quantum entanglement, the complex interactions between multiple physical components are modeled and optimized simultaneously; the linkage effect between the threaded rod and the top sealing plate is optimized and evaluated in multiple combinations in multiple operating states through parallel processing of quantum computing;

[0046] In the complex water flow regulation process, the quantum adaptive hybrid graph neural network simultaneously optimizes the switch state of the water pump, the rotation speed of the threaded rod, and the height of the top sealing plate, and adjusts the system operating parameters in real time to ensure that the system operates in the optimal state;

[0047] Sensor placement and data acquisition: In order to ensure the adjustment accuracy of the system, the placement of sensors at key locations is essential;

[0048] Sensor placement

[0049] Water flow sensor f(t): arranged at the inlet and outlet of the canal to monitor the change of water flow in real time;

[0050] Water pressure sensor p(t): arranged near the top sealing plate, monitoring the pressure of water flow on the top sealing plate, used to control the rotation angle of the threaded rod;

[0051] Threaded rod rotation sensor ω(t): monitors the rotation speed of the threaded rod and helps adjust the height of the top sealing plate;

[0052] Temperature and impurity sensors: monitor the ambient temperature and impurity content in the water channel to ensure that the system can operate stably in different environments;

[0053] Data collection and preprocessing

[0054] All data collected by sensors will be used as input to the deep reinforcement model to adjust the regulation strategy in real time:

[0055] Dynamic sampling frequency: When the water flow fluctuates violently, the system will increase the data collection frequency to ensure that important changes can be captured;

[0056] Data filtering and cleaning: Before data is input into the deep reinforcement model, the system will perform preprocessing to remove noise and abnormal data to ensure the accuracy of the adjustment;

[0057] Cache mechanism: Sensor data is transmitted to the central control system in real time, and a certain amount of historical data is cached for the quantum self-supervised comparative learning model to learn and use, thereby optimizing future adjustment decisions;

[0058] Step 2: Deep reinforcement learning decision system

[0059] In this step, the system models the water flow regulation problem as a Markov decision process through a deep reinforcement learning model, so that the system can adaptively learn the regulation strategy to ensure that the water flow reaches the target value while optimizing energy consumption and response time; The following is a detailed expansion of each part, and the application of each variable in the present invention is explained in combination with specific formulas;

[0060] Deep reinforcement learning model construction

[0061] State space construction: In the deep reinforcement learning model, the state space s(t) represents the overall state of the system at time t, covering all key variables that affect the system performance; these variables are data that the system monitors in real time through sensors and inputs into the deep reinforcement learning model; the construction of the state space is the basis of model decision-making, and the main variables are as follows:

[0062] Water flow rate f(t): The core variable of the system, indicating the current water flow in the canal; it is obtained in real time through the water flow sensor and compared with the target value f target For comparison; water flow is the key target of system regulation. The deep reinforcement learning model decides whether to adjust the speed ω(t) of the threaded rod and the height h(t) of the top sealing plate based on the real-time feedback of the water flow;

[0063] s(t)=f(t)

[0064] Among them, s(t) is the state space, f(t) is the water flow;

[0065] Height of top sealing plate h(t): The height of top sealing plate directly controls the size of water flow; the rotation angle of threaded rod will affect the height of top sealing plate, so this variable determines the effect of system regulation; the real-time height of top sealing plate h(t) is also an important part of state space:

[0066] h(t)=h0+k·θ(t)

[0067] Where h0 is the initial height, k is the proportional constant between the rotation of the threaded rod and the height change of the top sealing plate, h(t) is the height of the top sealing plate at time t, and θ(t) is the rotation angle of the threaded rod;

[0068] Rotation speed of the threaded rod ω(t): The rotation speed of the threaded rod directly affects the movement speed of the top sealing plate, so it is a key variable for controlling the accuracy of water flow regulation. By monitoring ω(t) in real time, deep reinforcement learning can accurately adjust the height of the top sealing plate;

[0069]

[0070] Where θ(t) is the rotation angle of the threaded rod, v(t) is the angular velocity of the threaded rod, represents the time derivative of the rotation angle of the threaded rod;

[0071] In summary, the state space s(t) is expressed as:

[0072] s(t)=[f(t),h(t),ω(t)]

[0073] Action space construction: The action space a(t) is the adjustment operation performed by the system, which mainly includes the following actions:

[0074] Top sealing plate height adjustment Δh(t): By adjusting the rotation speed ω(t) of the threaded rod, the height h(t) of the top sealing plate changes, thereby controlling the water flow; this is the main means for the system to regulate water flow;

[0075] Δh(t)=k·Δθ(t)

[0076] Among them, Δθ(t) is the angular change of the threaded rod rotation, Δh(t) is the height adjustment of the top sealing plate, and k is the proportional constant of the threaded rod rotation and the top sealing plate height change.

[0077] Adjustment of the rotation angle of the threaded rod: The rotation angle of the threaded rod determines the height of the top sealing plate, so the rotation speed ω(t) of the threaded rod is the key control variable; the rotation angle of the threaded rod affects the moving speed of the top sealing plate, thereby affecting the response speed of the water flow regulation;

[0078] Reward function design: The reward function R(s(t), a(t)) is the core of the deep reinforcement learning model and is used to measure whether the adjustment action currently performed by the system is effective. A reasonable reward function is designed to guide the system to achieve the optimization goal:

[0079] Difference between water flow and target water flow: The reward function encourages the water flow f(t) to be close to the target water flow f target ; By using the form of square error, the system can effectively reduce the deviation between the water flow and the target value:

[0080] R(s(t),a(t))=-(f(t)-f target ) 2

[0081] Among them, R(s(t),a(t)) is the reward function, f(t) is the water flow, and f target is the target water flow rate; when the water flow rate f(t) is closer to f target , the smaller the negative value of this part, the higher the reward;

[0082] Energy consumption penalty: In order to reduce the energy consumption of the system, a penalty term of the rotation speed ω(t) of the threaded rod is introduced into the reward function; the faster the rotation speed of the threaded rod, the higher the energy consumption, so a penalty term proportional to the speed is designed:

[0083] R(s(t),a(t))=-(f(t)-f target ) 2 -c·v(t)

[0084] Among them, c is the energy consumption penalty coefficient. By adjusting the value of c, the system finds a balance between adjustment accuracy and energy consumption;

[0085] Response time penalty: In some scenarios, the system needs to quickly adjust the water flow, so the response time is also one of the factors affecting the quality of regulation. A time penalty term is introduced to reward quick response operations:

[0086] R(s(t),a(t))=-(f(t)-f target ) 2 -c·v(t)-λ·T(t)

[0087] Where, T(t) is the response time, λ is the weight coefficient of the time penalty term;

[0088] Action strategy of reinforcement learning: Based on the state space s(t) and the reward function R(s(t), a(t)), the deep reinforcement learning model can learn the optimal action strategy through feedback:

[0089] When the water flow rate f(t) is significantly less than the target value f target , the system will increase the rotation speed ω(t) of the threaded rod, thereby increasing the height of the top sealing plate and increasing the water flow; when the water flow f(t) is too large, the system will slow down the speed ω(t) of the threaded rod, reduce the height of the top sealing plate, and reduce the water flow;

[0090] Through multiple feedback and adjustments, the deep reinforcement learning model gradually converges to the optimal strategy and can efficiently control water flow;

[0091] Reinforcement learning training and execution: Deep reinforcement learning training process: Deep reinforcement learning training gradually learns the optimal strategy through "trial and feedback"; the specific steps are as follows:

[0092] Initialization: Initialize the system state variables s(t), water flow, top sealing plate height, threaded rod speed, and set the initial strategy;

[0093] State update: Whenever the sensor detects a change in water flow or top seal position, the system captures the current state s(t);

[0094] Action execution: According to the action a(t) selected by the deep reinforcement learning model, the adjustment is performed to adjust the rotation angle θ(t) of the threaded rod;

[0095] Reward feedback: Through sensor feedback, the system calculates the reward value R(s(t), a(t)) to measure water flow and energy consumption;

[0096] Strategy update: Update the action strategy according to the feedback value and gradually optimize the system's adjustment strategy;

[0097] Real-time adjustment and model update: Deep reinforcement learning models can dynamically adjust according to environmental changes:

[0098] Step 3: Quantum self-supervised contrastive learning adaptive adjustment

[0099] Quantum self-supervised contrastive learning data processing: Quantum self-supervised contrastive learning combines quantum computing and self-supervised learning to process multi-dimensional sensor data in the canal system, quickly extract key features and predict optimal operations; quantum self-supervised contrastive learning accelerates the processing of complex data through quantum computing, enabling the system to efficiently respond to dynamic changes in water flow regulation;

[0100] Historical data pattern extraction: In addition to the real-time data obtained in step 1, quantum self-supervised comparative learning also relies on historical data; the system will extract the best adjustment strategy under similar water flow conditions from the historical database; these data include operation modes under different water flow, pressure, and temperature conditions; when the water flow f(t) in the historical data is close to the current flow under a certain state, the system will refer to the best combination of adjusting the threaded rod speed v(t) and the top sealing plate height h(t) in history, and make adjustments based on the current environmental conditions;

[0101] Feature screening and dimensionality reduction: The multidimensional data collected by sensors contains a large amount of irrelevant information; quantum self-supervised contrastive learning uses quantum computing technology to reduce the dimensionality of the data and extract the most representative features from it. The feature dimensionality reduction formula is:

[0102]

[0103] Among them, X is the original data matrix, Q is the dimension reduction matrix generated by the quantum algorithm, is the data after dimensionality reduction;

[0104] Through the quantum feature extraction algorithm, the system can quickly screen out the key variables that affect the regulation effect and identify the sensitivity of water flow changes to the speed of the threaded rod;

[0105] Quantum-accelerated feature learning: The parallel nature of quantum computing can significantly increase the speed of feature learning, especially when processing high-dimensional data. Quantum self-supervised contrastive learning accelerates feature extraction and pattern matching through quantum computing.

[0106] Quantum computing parallel processing: The superposition and entanglement characteristics of quantum computing enable quantum self-supervised contrastive learning to process multiple data states simultaneously, significantly shortening the time for feature learning:

[0107] Quantum superposition allows the system to analyze multiple flow states simultaneously and calculate the similarity of these states with historical data; this speeds up the pattern matching process and quickly finds the most similar historical regulation strategy; through quantum superposition, the system compares multiple historical states simultaneously;

[0108] Quantum feature learning: Through quantum computing, quantum self-supervised comparative learning analyzes the multi-dimensional data of the water flow system in a very short time and quickly finds key features; when the water flow fluctuates greatly, the system processes the water flow h(t), threaded rod speed v(t) and top sealing plate height h(t) in parallel to determine which variable has the greatest impact on the current regulation; feature learning formula:

[0109]

[0110] Among them, w i is the weight, indicating that the variable x i Contribution to the loss function L;

[0111] By calculating the impact of each variable on the water flow regulation effect, quantum self-supervised contrastive learning identifies which features should be prioritized;

[0112] Fast pattern matching: The quantum self-supervised comparative learning system can quickly match the current water flow state with similar patterns in historical data through the parallelism of quantum computing:

[0113] As the current water flow rate h(t) increases, the quantum self-supervised comparative learning system finds the best adjustment plan for the threaded rod speed v(t) and the top sealing plate height h(t) in history by comparing similar situations of rapid increase in water flow in history. The quantum algorithm speeds up the pattern matching, allowing the system to give the best adjustment plan in an instant.

[0114] Contrastive learning model adjustment: Another important function of quantum self-supervised contrastive learning is to adjust model parameters through contrastive learning so that the system can adaptively respond to dynamically changing water flow regulation needs;

[0115] Adaptive adjustment of the model: By comparing the current water flow state with similar states in historical data in real time, quantum self-supervised comparative learning dynamically adjusts the regulation strategy;

[0116] When the current water flow h(t) increases, the quantum self-supervised comparative learning system finds the historically optimal adjustment plan for the threaded rod speed v(t) and the top sealing plate height h(t) by comparing similar situations in which the water flow rate increased rapidly in history;

[0117] v new (t) = v opt (t history )

[0118] Among them, v opt (t history ) is the optimal threaded rod speed under similar water flow conditions in history, v new (t) is the latest selected threaded rod speed

[0119] Real-time comparison and historical pattern matching: The system compares the current water flow status with the patterns in historical data, finds similar water flow regulation strategies in history, and applies them in real time; the specific process is as follows:

[0120] The data f(t), h(t), and v(t) captured by the current sensor are input into the quantum computing model to quickly calculate its relationship with the historical state f history (t) similarity;

[0121]

[0122] Where f(t) is the flow rate captured by the sensor, f history (t) is the optimal flow rate under similar historical conditions, Similarity(f(t),f history (t)) represents f(t) and f history (t), the similarity value is used to judge the matching degree between the current state and the historical pattern; if the similarity is high, the system will apply the best adjustment parameters in history; ||||: represents the norm of the data, indicating the size and strength of the data; the norm is used to normalize the strength of the data to ensure that the calculation of similarity is not affected by the absolute value of the data;

[0123] Prediction based on contrastive learning: The quantum self-supervised contrastive learning system can not only compare the current state with historical data, but also make predictions based on historical trends;

[0124] When the current water flow rate f(t) increases rapidly, the system predicts that the water flow rate will increase further in the future and adjusts the top sealing plate height h(t) in advance to avoid water flow overload;

[0125] h predicted (t) = h current (t)+α·Δf(t)

[0126] Among them, h predicted (t) is the expected top sealing plate height, h current (t) is the measured top sealing plate height, α is an empirical parameter, which represents the influence coefficient of water flow change on the top sealing plate height, and Δf(t) is the change value of water flow;

[0127] Dynamic adjustment and feedback mechanism: Through contrast learning and feedback mechanism, the quantum self-supervised contrast learning system can adaptively adjust the threaded rod speed and top sealing plate height to achieve the optimal adjustment effect;

[0128] Real-time dynamic adjustment: When the system detects that the current water flow rate f(t) is equal to the target value f target When there is a gap, quantum self-supervised comparative learning will adjust the threaded rod rotation speed v(t) and the top sealing plate height h(t) according to the feedback to ensure that the water flow rate gradually approaches the target value;

[0129] v(t+1)=v(t)+Δv(t)

[0130] Among them, v(t) is the rotation speed of the threaded rod at time t, Δv(t) is the speed increment of the threaded rod adjusted according to contrast learning, and v(t+1) is the rotation speed of the threaded rod at time t+1;

[0131] Model iteration and update: The quantum self-supervised comparative learning system will update the model based on the feedback data after each adjustment. If the water flow rate still does not reach the target value after adjustment, the system will further optimize the adjustment parameters and iteratively improve the adjustment effect. After each adjustment, the system will update its understanding of the water flow regulation effect and be more accurate in future adjustments.

[0132] In order to further improve the present invention, step 4 is also included: multi-objective quantum evolutionary search and hyper-dimensional distributed hybrid optimization; in the canal regulation system, the control of the system not only involves the single goal of water flow regulation, but also needs to comprehensively consider multiple mutually constrained goals such as energy consumption and response time; in order to achieve efficient, accurate and low-energy regulation of the system, multi-objective quantum evolutionary search and hyper-dimensional distributed hybrid optimization are introduced, and combined with quantum computing technology, global and local optimization under complex systems is achieved; the process of each step, the derivation of the formula and its integration with the patent content are described in detail below;

[0133] Multi-objective optimization refers to the system finding the global optimal solution while meeting multiple different regulation objectives. The multi-objective optimization of the canal regulation system includes the following three key objectives: water flow f(t), regulation energy consumption E(t), and system response time T(t). Each objective will affect the overall performance of the system, and the optimal effect must be achieved through balanced optimization.

[0134] Water flow rate: Make the water flow rate f(t) as close to the target value f as possible target ;

[0135] R f (t)=-|f(t)-f target |

[0136] Among them, f(t) is the real-time water flow, f target is the target water flow rate, R f (t) is the reward value of water flow deviation;

[0137] The main regulation task of the system is to minimize |f(t)-f target |To achieve precise water flow regulation; precise control of water flow is related to the practical application of irrigation and canal water delivery;

[0138] Adjustment energy consumption: Minimize the energy consumption during adjustment;

[0139] The energy consumption is mainly proportional to the rotation speed v(t) of the threaded rod; the energy consumption E(t) is expressed as:

[0140] E(t)=c·v 2 (t)

[0141] Where c is the energy consumption coefficient, v(t) is the rotation speed of the threaded rod; the energy consumption target of the system is to minimize E(t) to reduce the energy loss caused by frequent adjustments;

[0142] System response time T(t): shorten the system response time

[0143] The response time T(t) represents the time from when the external conditions change to when the water flow rate is adjusted to the target value. The system hopes to complete the adjustment in the shortest time to ensure that the water flow rate reaches the target value quickly. The objective function of response time optimization is expressed as:

[0144] R T (t)=-T(t)

[0145] Among them, R T (t) is the reward value of response time;

[0146] The system minimizes the response time in each adjustment by quantifying the time response target;

[0147] Quantum evolutionary search and global optimization: Multi-objective quantum evolutionary search is a global optimization algorithm implemented through quantum computing. It searches for the optimal solution of multi-dimensional objectives through the superposition and evolution of quantum states. The task of multi-objective quantum evolutionary search in the canal regulation system is to comprehensively consider the three objectives of water flow, energy consumption, and response time, and explore the global optimal regulation strategy.

[0148] Quantum state superposition and evolution

[0149] Quantum state superposition: In multi-objective quantum evolutionary search, each adjustment scheme, the top sealing plate height h(t) and the threaded rod speed v(t) are represented as the state of quantum bits; quantum state superposition allows multiple adjustment schemes to be represented simultaneously, and multiple combinations can be evaluated simultaneously through parallel calculation. The quantum state formula is:

[0150]

[0151] Where N is the number of states, |ψ(t)> is the superposition of the regulated states, and s i (t) is each adjustment state, s i (t)> is the superposition state of each regulated state; α i is the probability amplitude of the corresponding state; through quantum state superposition, the system calculates multiple adjustment schemes in parallel, greatly improving the search efficiency;

[0152] Quantum state evolution: As the evolutionary search proceeds, the quantum state continues to evolve in each iteration, gradually tending towards the global optimal solution; after each iteration, the system collapses the quantum state and selects the optimal adjustment parameters; quantum state evolution formula:

[0153] |ψ′(t)>=U(t)|ψ(t)>

[0154] Among them, |ψ′(t)> is the superposition state of the adjusted state after adjustment, |ψ(t)> is the superposition state of the adjusted state, and U(t) is the unit matrix of quantum state evolution, which adjusts the quantum state according to the fitness function; through iterative evolution, the system gradually converges to the global optimal solution;

[0155] Global optimization process: The global search process of multi-objective quantum evolutionary search is as follows:

[0156] Initialize quantum state: At the beginning of the search, the system generates multiple adjustment parameter combinations such as the top sealing plate height h(t) and the threaded rod speed v(t); these combinations are represented by the superposition state of quantum bits, and the system explores multiple adjustment schemes simultaneously;

[0157] Fitness function evaluation: The system evaluates each regulation scheme through the fitness function; the fitness function comprehensively considers water flow error, energy consumption and response time:

[0158] F(s(t))=w1·|f(t)-f target |+w2·E(t)+w3·T(t)

[0159] Among them, F(s(t)) is the fitness function of state s(t), w1, w2, w3 are the weight coefficients of each target, which determines the importance of each target in the global optimization;

[0160] Quantum state update and evolution: According to the evaluation results of the fitness function, the system updates the quantum state, retains the adjustment schemes with excellent performance, and eliminates the schemes with poor performance; through the quantum state evolution formula, the system gradually approaches the global optimal solution;

[0161] Collapse quantum state to determine the optimal solution: When the set number of iterations is reached, the system collapses the quantum state and selects the optimal combination of adjustment parameters; these parameters ensure the balance of water flow, energy consumption and response time;

[0162] Hyperdimensional Distributed Hybrid Optimization

[0163] Hyperdimensional distributed hybrid optimization is an optimization method that combines distributed computing and hyperdimensional search. It is suitable for complex multi-objective systems. Hyperdimensional distributed hybrid optimization assigns each target to a different computing node and optimizes the parameters of each dimension in parallel, thereby speeding up the overall optimization speed.

[0164] Distributed optimization process: Hyperdimensional distributed hybrid optimization decomposes the multi-objective optimization task into multiple independent computing tasks and performs the optimization of each dimension in parallel;

[0165] Node decomposition and task allocation: multiple target water flow errors, energy consumption and response time of the system are allocated to different computing nodes; each node is responsible for optimizing one target independently, reducing the complexity of the overall calculation;

[0166] For each node i, the optimization objective is expressed as:

[0167] F i (s(t)) = w i ·G(s(t))

[0168] Among them, G(s(t)) is the objective function that the node is responsible for, and w i is the weight of the corresponding target; F i (s(t)) represents the fitness function of node i for state s(t);

[0169] Parallel optimization execution: Each computing node performs its optimization task independently. Through real-time sensor data, the system continuously adjusts the parameters responsible for each node, finds their own local optimal solution, and returns it to the central control system.

[0170] Hyperdimensional search and parameter adjustment: Hyperdimensional distributed hybrid optimization’s hyperdimensional search technology allows efficient search for optimal solutions in multidimensional space;

[0171] Multi-dimensional search space construction: Hyper-dimensional distributed hybrid optimization constructs the system's adjustment parameters into a multi-dimensional search space; each dimension represents an adjustment parameter, and the system seeks the global optimal solution by optimizing the values ​​of each dimension; search space formula:

[0172]

[0173] The system searches in S to achieve the optimal balance of all objectives;

[0174] Among them, S: is the set of search space, which represents all combinations of system adjustment parameters; by searching in S, the system can find the adjustment parameter value that optimizes the water flow; (v(t), h(t), f(t)) is a vector in the search space, which represents the value of the three adjustment parameters of the system at a certain time t; the system balances water flow, energy consumption and response time by optimizing these three parameters; v(t): represents the rotation speed of the threaded rod, and its unit is revolutions per minute; this is a key parameter for the system to adjust the water flow. The faster the speed, the more sensitive the water flow adjustment; h(t): represents the height of the top sealing plate, and its unit is meter; by adjusting the height of the top sealing plate, the system can directly control the water flow in the canal; f(t): represents the current water flow, and its unit is liters per minute; this is the core goal of the regulation. The system needs to optimize v(t) and h(t) to make the water flow reach the target value; This symbol indicates that (v(t), h(t), f(t)) belongs to the three-dimensional real space That is, these three parameters can take any real value; it defines that the search space is a three-dimensional vector space; v min : Minimum threaded rod speed; v max : Maximum threaded rod speed; h min : The minimum height of the top sealing plate; h max : The highest height of the top sealing plate;

[0175] Dynamic adjustment and feedback optimization: During the hyper-dimensional search process, the system performs real-time feedback optimization based on sensor data and adjusts the adjustment parameters to ensure that each target reaches the optimal state;

[0176] Step 5: Feedback and optimization iteration

[0177] Feedback and optimization iteration are crucial links in the entire water flow regulation system, which ensures that the system can continuously learn and optimize its operation to cope with the changing environmental needs. Through real-time feedback from sensors, the system can adaptively adjust the regulation strategy and continuously improve the operating performance through quantum computing and optimization algorithms.

[0178] Real-time feedback mechanism

[0179] Feedback data and adjustment strategies: Real-time feedback is an important part of the regulation system. The system collects real-time data through a variety of sensors arranged in the canal. These data directly reflect the current operating status of the system and the actual effect of the regulation.

[0180] Feedback data collection: The data provided by the sensor include water flow f(t), top sealing plate height h(t), threaded rod rotation speed v(t) and energy consumption data E(t); these data will be transmitted to the system's central control module in real time to analyze the difference between the current operation effect and the expected target;

[0181] State evaluation and feedback analysis: The system first analyzes the feedback data provided by the sensor to evaluate whether the current water flow rate f(t) is close to the target water flow rate f target ; If the current water flow is too large or too small, the system will record the difference and feed it back to the deep reinforcement learning model as input for decision-making on the next action;

[0182] The system also analyzes the energy consumption E(t). If the energy consumption is too high, it means that there is unnecessary energy waste in the operation of the system. It is necessary to optimize the rotation speed of the threaded rod and the height adjustment strategy of the top sealing plate.

[0183] Action strategy adjustment: The deep reinforcement learning model will readjust the action strategy based on the current feedback data. If the water flow rate f(t) deviates from the target value, the system will select a new action a(t) based on the deep reinforcement learning model to adjust the top sealing plate height Δh(t) and the threaded rod rotation speed. If the energy consumption is high, the system will choose to slow down the threaded rod rotation speed v(t) to reduce energy consumption, and continue to optimize the water flow rate by slightly adjusting the top sealing plate height.

[0184] Global optimization and readjustment: The real-time feedback system can ensure the effectiveness of local optimization, and local adjustments are not enough to achieve global optimization; therefore, the system needs to periodically perform global optimization through multi-objective quantum evolutionary search and hyper-dimensional distributed hybrid optimization;

[0185] Periodic global evaluation: At the end of each adjustment cycle, the system will evaluate the current overall performance; the core of the evaluation is to check whether the goals such as water flow, energy consumption and response time have reached the global optimization state;

[0186] The system calculates the deviation between the current water flow and the target flow, the difference between the current energy consumption and the minimum energy consumption, and the comparison between the system response speed and the set value;

[0187] Global optimization recalculation: If the expected effect is not achieved during the local adjustment process, the water flow is still not within the ideal range, and the energy consumption is too high, the system will start global optimization; the multi-objective quantum evolutionary search will restart the quantum evolutionary search algorithm to find the global optimal solution in a larger range:

[0188] Quantum evolutionary search: The system uses the parallel characteristics of quantum computing to explore new combinations of adjustment schemes, including different threaded rod speeds, top sealing plate heights, etc., to avoid falling into local optimality;

[0189] Hyper-dimensional distributed optimization: By optimizing different objectives in parallel, we can ensure the balance between the objectives and find the global optimal combination of adjustment parameters.

[0190] Readjustment and feedback: When the global optimal solution is found, the system will adjust the operating parameters of the threaded rod and the top sealing plate, and monitor the new adjustment effect through a real-time feedback mechanism; if the new round of adjustment is effective, the system will continue to maintain the current operating state; if the adjustment effect still does not meet expectations, it will continue to iterate and optimize;

[0191] Final adjustments and optimization completed

[0192] System convergence and automation: After multiple rounds of feedback, adjustment and optimization, the system will gradually converge to an optimal adjustment strategy. At this stage, the system adjustment process will become highly automated, and the operating parameters of the threaded rod and the top sealing plate will also stabilize at an optimal state.

[0193] Automated adjustment: Once the system converges to the optimal solution, the rotation of the threaded rod and the height adjustment of the top sealing plate will be automatically performed based on the learned optimal strategy; the system no longer requires frequent manual intervention, but relies on deep reinforcement learning models and the best strategy from historical feedback data for fully automatic adjustment;

[0194] Intelligent adjustment and control: The deep reinforcement learning model combines the features learned by quantum self-supervised comparative learning to adaptively adjust operations according to changes in the external environment. When the external water flow state changes significantly, the system quickly adjusts the operation strategy of the top sealing plate and the threaded rod based on similar historical situations to achieve real-time response and dynamic adjustment.

[0195] Long-term stable operation

[0196] Continuous monitoring and intelligent optimization: Even if the system has converged to the optimal state, sensors will continue to monitor water flow, energy consumption and other environmental factors; if external conditions change, water pressure in the canal rises, and temperature changes, the system can adjust its operation in time through sensor feedback to ensure that the water flow regulation process continues to operate stably under the new environmental conditions;

[0197] Long-term adaptive learning: The system not only adjusts based on current operating data, but also continuously optimizes the adjustment strategy based on long-term operating data; the deep reinforcement learning model learns from new feedback data and updates its strategy, thereby improving the stability and efficiency of future adjustments;

[0198] Maintenance and adjustment prompts: During long-term operation, the system can also identify potential operational anomalies through continuous self-monitoring; when the sensor detects that the water flow regulation effect is gradually deteriorating or the rotation speed of the threaded rod is abnormal, the system will issue an early warning, indicating that maintenance or further adjustment of the equipment is required.

[0199] Compared with the prior art, the present invention has the following beneficial effects:

[0200] 1) During the operation of the water-saving irrigation canal, the water pump body is started, and the liquid on one side of the connecting plate on the driving side of the water pump body is pumped into the top mixing barrel along the connecting pipe, and through the opening of the top connecting hole, the liquid in the top mixing barrel will flow into the filter box with gravity, and the liquid can be initially filtered through the front filter screen, reducing the blockage of large debris in the canal body and improving the filtering effect;

[0201] 2) During the operation of the water-saving irrigation canal, the inner filter screen is in an inclined state. After the sand and gravel particles stay on the surface of the inner filter screen, they will fall and roll along the inclined angle. The sand and gravel filtered on the surface of the inner filter screen will slide into the surface of the storage frame along the side discharge trough, and the sand and gravel will be uniformly stored by the storage frame on the surface of the supporting bottom frame. The filtered liquid in the filter box will flow downward along the bottom square frame and flow into the canal body. By filtering the sand and gravel in the liquid, the impact caused by the sand and gravel during irrigation can be reduced, thereby improving the irrigation effect.

[0202] 3) During the operation of the water-saving irrigation canal, the height of the adjusting plate inside the top connecting frame can be adjusted by adjusting the height of the top sealing plate. By raising the adjusting plate, the flow rate of water in the side supporting frame can be increased. If the adjusting plate is lowered, the flow rate of water can be reduced through the side connecting holes on the surface of the adjusting plate. The flow rate can be adjusted according to different environments and needs, which can improve the water-saving performance of the canal body.

[0203] 4) During the operation of the water-saving irrigation canal, the motor body is started and the transmission rod is driven by the motor body to rotate along the inner side of the top connecting hole. The rotating transmission rod drives the stirring blade to stir along the inner side of the top mixing barrel. The continuously stirring stirring blade mixes the liquid in the top mixing barrel, thereby improving the effect of subsequent irrigation.

[0204] 5) Improve regulation accuracy and response speed: The quantum adaptive hybrid graph neural network can dynamically capture the complex dependencies between physical components in the canal system (such as threaded rods, top sealing plates, sensors, etc.) by modeling them as graph neural network nodes. Through the parallelism and adaptive characteristics of quantum computing, the system can analyze and optimize operating parameters such as water flow, threaded rod rotation speed, and top sealing plate height in real time, making the regulation process more accurate and efficient. The combination of deep reinforcement learning and quantum self-supervised comparative learning enables the system to respond quickly under complex and changeable water flow conditions. Through the self-supervised learning model, the best regulation plan can be quickly extracted from historical data, reducing the trial time in the regulation process and greatly improving the response speed.

[0205] 6) Strengthen adaptive capabilities and achieve dynamic regulation: The quantum self-supervised contrastive learning (QSSCL) model enables the system to have adaptive learning capabilities. Compared with the water flow regulation method based on fixed rules in the prior art, the system of the present invention can dynamically analyze the current water flow state and automatically predict the optimal regulation parameters in combination with historical data, so that the system can adapt to changes in the environment and ensure that the optimal regulation effect can be achieved under different water flow and environmental conditions. The system can automatically adjust the regulation strategy according to real-time sensor feedback. When the external water flow conditions change drastically, the system can update the model in time to ensure that the regulation effect is not affected by the outside world.

[0206] 7) Improve the automation and intelligence level of the system: By introducing deep reinforcement learning, the system can automatically learn from external feedback, continuously optimize the operation strategy, reduce manual intervention, and achieve a high degree of automation and intelligence in the system's adjustment process. The system can continuously adjust the operating parameters according to the feedback from the sensor, and accelerate the feature extraction and pattern recognition process through quantum computing, so that it can automatically make the best decision in a dynamic and complex environment. This makes the present invention more robust and adaptive than traditional technologies when dealing with complex adjustment tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0207] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0208] Figure 1 It is a schematic diagram of the overall appearance structure of the present invention;

[0209] Figure 2 It is a schematic diagram of the side support platform structure of the present invention;

[0210] Figure 3 It is a schematic diagram of the side support platform structure of the present invention;

[0211] Figure 4 It is a schematic diagram of the top support plate structure of the present invention;

[0212] Figure 5 It is a schematic diagram of the structure of the inner filter of the present invention;

[0213] Figure 6 It is a schematic diagram of the side discharge chute structure of the present invention;

[0214] Figure 7 It is a schematic diagram of the top mixing barrel structure of the present invention;

[0215] Figure 8 A flowchart of an intelligent algorithm introduced into step S3 of the present invention.

[0216] In the figure: 1, water channel body; 201, side support platform; 202, side support plate; 203, side support frame; 204, top connecting frame; 205, top limit platform; 206, top threaded hole; 207, top sealing plate; 208, adjustment plate; 209, side connecting hole; 2010, threaded rod; 2011, top adjustment handle; 2012, bottom bearing plate; 301, top support platform; 302, top support plate; 303, bottom connecting frame; 304, support bottom frame ; 305, front filter; 306, side connecting plate; 307, liquid inlet; 308, connecting pipe; 309, water pump body; 3010, top mixing cylinder; 3011, filter box; 3012, top connecting hole; 3013, inner filter; 3014, bottom square frame; 3015, side discharge trough; 3016, storage frame; 401, top fixing cylinder; 402, top connecting hole; 403, motor body; 404, transmission rod; 405, stirring blade. DETAILED DESCRIPTION

[0217] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0218] In this embodiment, Figure 1-Figure 7 Given, the present invention provides the following technical solutions:

[0219] A water-saving irrigation canal comprises a canal body 1, a side support platform 201 is fixedly connected to the top of the canal body 1, a side support plate 202 is fixedly connected to one side of the side support platform 201, a side support frame 203 is fixedly connected to one side of the side support plate 202, a top connecting frame 204 is provided on the top of the side support frame 203, a top limiting platform 205 is fixedly connected to one side of the side support plate 202, a top threaded hole 206 is provided on the inner side of the top limiting platform 205, a threaded rod 2010 is threadedly connected to the inner side of the top threaded hole 206, an adjusting plate 208 is provided at the bottom of the threaded rod 2010, and a side connecting hole 209 is provided on the surface of the adjusting plate 208.

[0220] It should be noted that the flow rate of water can be reduced through the side connecting holes 209 on the surface of the regulating plate 208 . The flow rate can be adjusted according to different environments and needs, thereby improving the water-saving performance of the water channel body 1 .

[0221] In this embodiment, the top of the adjustment plate 208 is fixedly connected to the top sealing plate 207, the top of the top sealing plate 207 is movably connected to the bottom bearing disk 2012, the top of the bottom bearing disk 2012 is movably connected to the bottom of the threaded rod 2010, and the top of the threaded rod 2010 is fixedly connected to the top adjustment handle 2011.

[0222] It should be noted that the operator first holds the top adjustment handle 2011 and then rotates it. The rotating top adjustment handle 2011 will drive the threaded rod 2010 to rotate. Through the threaded connection between the threaded rod 2010 and the top threaded hole 206, the height of the top sealing plate 207 will be adjusted during the threaded rotation of the threaded rod 2010.

[0223] In this embodiment, one side of the side support platform 201 is fixedly connected to the top support platform 301, one side of the top support platform 301 is fixedly connected to the top support plate 302, one side of the top support plate 302 is fixedly connected to the bottom connecting frame 303, one side of the bottom connecting frame 303 is fixedly connected to the side connecting plate 306, one side of the side connecting plate 306 is fixedly connected to the connecting pipe 308, one side of the connecting pipe 308 is provided with a water pump body 309, the bottom of the water pump body 309 is fixedly connected to the top mixing barrel 3010, the bottom of the top mixing barrel 3010 is fixedly connected to the filter box 3011, and the inner side of the filter box 3011 is fixedly connected to the inner filter net 3013.

[0224] It should be noted that the water pump body 309 is started and drives the liquid on one side of the side connecting plate 306 to be pumped into the top mixing barrel 3010 along the connecting pipe 308. Through the opening of the top connecting hole 3012, the liquid in the top mixing barrel 3010 will flow into the filter box 3011 by gravity.

[0225] In this embodiment, a front filter 305 is fixedly connected to one side of the water channel body 1 , and one side of the front filter 305 and one side of the side connecting plate 306 are arranged parallel to each other along the surface of the water channel body 1 .

[0226] It should be noted that when the liquid flows along the water channel body 1, it will first come into contact with the front filter 305, and filter out larger debris through the front filter 305 to block the larger debris. The filtered liquid will flow through the front filter 305 to one side of the side connecting plate 306 and accumulate on one side of the side connecting plate 306.

[0227] In this embodiment, a top communicating hole 3012 is provided at the top of the filter box 3011 , and a bottom square frame 3014 is provided at the bottom of the filter box 3011 .

[0228] It should be noted that, through the opening of the top communicating hole 3012 , the liquid in the top mixing barrel 3010 will flow into the filter box 3011 under gravity, and the liquid can be preliminarily filtered through the front filter screen 305 .

[0229] In this embodiment, a side discharge chute 3015 is provided on one side of the filter box 3011 , and a storage frame 3016 is provided on one side of the filter box 3011 .

[0230] It should be noted that after the sand and gravel particles stay on the surface of the inner filter 3013, they will fall and roll along the tilt angle, and the sand and gravel filtered on the surface of the inner filter 3013 will slide into the surface of the storage frame 3016 along the side discharge chute 3015.

[0231] In this embodiment, the top of the bottom connecting frame 303 is fixedly connected to the supporting bottom frame 304 , and the top of the supporting bottom frame 304 is arranged at the bottom of the storage frame 3016 .

[0232] It should be noted that the sand and gravel are uniformly stored through the storage frame 3016 on the surface of the support frame 304, and the filtered liquid in the filter box 3011 will flow downward along the bottom square frame 3014 and flow into the canal body 1. By filtering the sand and gravel in the liquid, the impact caused by the sand and gravel during irrigation can be reduced, thereby improving the irrigation effect.

[0233] In this embodiment, the top of the top mixing barrel 3010 is fixedly connected to the top fixing barrel 401, the top of the top mixing barrel 3010 is fixedly connected to the motor body 403, the output shaft of the motor body 403 is fixedly connected to the transmission rod 404, the surface of the transmission rod 404 is fixedly connected to the stirring blade 405, and the top of the top mixing barrel 3010 is opened with a top connecting hole 402.

[0234] It should be noted that the rotating transmission rod 404 will drive the stirring blade 405 to stir along the inner side of the top mixing barrel 3010, and the continuously stirring stirring blade 405 will mix the liquid in the top mixing barrel 3010, thereby improving the effect of subsequent pouring.

[0235] In this embodiment, one side of the side connecting plate 306 is provided with two liquid inlet holes 307 .

[0236] It should be noted that the water pump body 309 is started and drives the liquid on one side of the side connecting plate 306 to be pumped into the top mixing barrel 3010 along the connecting pipe 308. Through the opening of the top connecting hole 3012, the liquid in the top mixing barrel 3010 will flow into the filter box 3011 by gravity.

[0237] The present invention provides a water-saving irrigation canal and a regulating method, which are used to further illustrate the working process or principle of the water-saving irrigation canal and the optimization method provided in the above-mentioned embodiment 1, and the details are as follows:

[0238] A water-saving irrigation canal regulation method comprises the following steps:

[0239] S1. When the liquid flows along the water channel body 1, it will first contact the front filter 305, and filter the larger debris through the front filter 305, and block the larger debris. The filtered liquid will flow through the front filter 305 to one side of the side connecting plate 306, and accumulate on one side of the side connecting plate 306. The water pump body 309 is started, and the liquid on one side of the side connecting plate 306 is driven by the water pump body 309 to be pumped into the top mixing barrel 3010 along the connecting pipe 308, and through the opening of the top connecting hole 3012, the liquid in the top mixing barrel 3010 will flow into the filter box 3011 with gravity, and the liquid can be preliminarily filtered by the front filter 305, reducing the blockage of larger debris in the water channel body 1, thereby improving the filtering effect;

[0240] S2. The liquid entering the filter box 3011 will contact the inner filter screen 3013 and be filtered through the inner filter screen 3013 with a smaller gap, blocking and filtering the smaller sand and gravel particles in the liquid. The inner filter screen 3013 is in an inclined state. After the sand and gravel particles stay on the surface of the inner filter screen 3013, they will fall and roll along the inclined angle. The sand and gravel filtered on the surface of the inner filter screen 3013 will slide along the side discharge trough 3015 into the surface of the storage frame 3016, and the sand and gravel will be uniformly stored through the storage frame 3016 on the surface of the support frame 304. The filtered liquid in the filter box 3011 will flow downward along the bottom square frame 3014 and flow into the water channel body 1. By filtering the sand and gravel in the liquid, the impact caused by the sand and gravel during irrigation can be reduced, thereby improving the irrigation effect.

[0241] S3. During the process of liquid flowing along the water channel body 1, according to the different required flow rates, the operator first holds the top adjustment handle 2011 and rotates it. The rotating top adjustment handle 2011 drives the threaded rod 2010 to rotate. The threaded connection between the threaded rod 2010 and the top threaded hole 206 is used. During the threaded rotation of the threaded rod 2010, the height of the top sealing plate 207 is adjusted, and the bottom end of the threaded rod 2010 is movably connected to the top sealing plate 207 through the bottom bearing plate 2012. The connection of 2012 will maintain the stability of the top sealing plate 207 when the bottom bearing plate 2012 rotates. By adjusting the height of the top sealing plate 207, the height of the adjusting plate 208 inside the top connecting frame 204 can be adjusted. By raising the adjusting plate 208, the flow rate of the water flow in the side supporting frame 203 can be increased. If the adjusting plate 208 is lowered, the flow rate of the water flow can be reduced through the side connecting holes 209 on the surface of the adjusting plate 208. The flow rate can be adjusted according to different environments and needs, which can improve the water-saving performance of the water channel body 1.

[0242] S4. At the same time, different microorganisms or liquid fertilizers can be added to the top mixing barrel 3010 through the top fixed barrel 401. The added liquid will enter the top mixing barrel 3010 along the top fixed barrel 401. At this time, by starting the motor body 403, the transmission rod 404 is driven by the motor body 403 to rotate along the inner side of the top connecting hole 402. The rotating transmission rod 404 will drive the stirring piece 405 to stir along the inner side of the top mixing barrel 3010. The continuously stirring stirring piece 405 will mix the liquid in the top mixing barrel 3010, thereby improving the effect of subsequent irrigation.

[0243] like Figure 8 , the specific numerical case analysis of the control algorithm of the present invention is as follows:

[0244] Feedback data and adjustment strategies

[0245] Initial conditions:

[0246] Target water flow rate f target =00L / min

[0247] Current water flow rate f(t) = 90L / min

[0248] Top sealing plate height h(t)=1.5m

[0249] Threaded rod rotation speed v(t) = 20RPMv

[0250] Energy consumption coefficient c = 0.05kW / RPM2

[0251] According to the formula, the system first calculates the water flow error:

[0252] |f(t)-f target ∣=∣90-100∣=10L / min

[0253] Reward function calculation:

[0254] R(s(t),a(t))=-(90-100)2-0.05·202=-100-20=-120

[0255] Based on this negative reward value, the system determines that the current adjustment effect is not good and needs to be corrected by adjusting the threaded rod speed or the top sealing plate height.

[0256] Global Optimization and Rescaling

[0257] After feedback adjustment, the system still did not achieve the expected effect, so global optimization was performed. The optimization goals are as follows:

[0258] Minimize water flow error |f(t)-f target ∣

[0259] Minimize energy consumption E(t) = 0.05 v(t)2

[0260] Shorten the response time T(t)=5seconds

[0261] The multi-objective optimization function is:

[0262] F(s(t))=0.5·|f(t)-f target ∣+0.3·E(t)+0.2·T(t)

[0263] Substituting the initial conditions into:

[0264] F(s(t))=0.5·10+0.3·(0.05·202)+0.2·5=5+6+1=12

[0265] The system readjusts the threaded rod speed and top sealing plate height through global search and eventually finds the best solution so that the water flow rate is close to the target value with the lowest energy consumption.

[0266] Final adjustments and optimization completed

[0267] System Convergence and Automation

[0268] After multiple feedback and global optimization, the system converges to the following adjustment parameters:

[0269] Threaded rod speed v(t) = 15 RPM

[0270] Top sealing plate height h(t)=1.8m

[0271] At this point, the water flow rate is close to the target value f(t) = 100 L / min, and the system energy consumption is reduced to:

[0272] E(t)=0.05·152=11.25kW

[0273] Long-term stable operation

[0274] During long-term operation, the system continues to monitor external conditions, such as water flow fluctuations and temperature changes. If it detects that the water flow rate has increased to f(t) = 110L / min, the system will automatically adjust the threaded rod speed and top sealing plate height:

[0275] The new threaded rod speed v(t) = 18RPM, the top sealing plate height is adjusted to h(t) = 1.6m, the water flow rate is restored to f(t) = 100L / min, the system response time T(t) = 4seconds, and the energy consumption remains at a low level: E(t) = 0.05·182 = 16.2kW

[0276] It should be noted that the water pump body 309 and the motor body 403 in the present invention are both existing technologies, and the corresponding models can be selected according to actual needs. The internal structure and operating principle of the above parts are also common knowledge of technicians in this field, so no further elaboration will be made on this.

[0277] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A water-saving irrigation canal, comprising a canal body (1), characterized in that: The top of the water channel body (1) is fixedly connected to a side support platform (201), one side of the side support platform (201) is fixedly connected to a side support plate (202), one side of the side support plate (202) is fixedly connected to a side support frame (203), the top of the side support frame (203) is provided with a top connecting frame (204), one side of the side support plate (202) is fixedly connected to a top limiting platform (205), the inner side of the top limiting platform (205) is provided with a top threaded hole (206), the inner side of the top threaded hole (206) is threadedly connected to a threaded rod (2010), the bottom of the threaded rod (2010) is provided with an adjustment plate (208), and the surface of the adjustment plate (208) is provided with a side connecting hole (209).

2. A water-saving irrigation canal according to claim 1, characterized in that: The top of the adjustment plate (208) is fixedly connected to a top sealing plate (207), the top of the top sealing plate (207) is movably connected to a bottom bearing disk (2012), the top of the bottom bearing disk (2012) is movably connected to the bottom of the threaded rod (2010), and the top of the threaded rod (2010) is fixedly connected to a top adjustment handle (2011).

3. A water-saving irrigation canal according to claim 2, characterized in that: One side of the side support platform (201) is fixedly connected to a top support platform (301), one side of the top support platform (301) is fixedly connected to a top support plate (302), one side of the top support plate (302) is fixedly connected to a bottom connecting frame (303), one side of the bottom connecting frame (303) is fixedly connected to a side connecting plate (306), one side of the side connecting plate (306) is fixedly connected to a connecting pipe (308), one side of the connecting pipe (308) is provided with a water pump body (309), the bottom of the water pump body (309) is fixedly connected to a top mixing barrel (3010), the bottom of the top mixing barrel (3010) is fixedly connected to a filter box (3011), and the inner side of the filter box (3011) is fixedly connected to an inner filter screen (3013).

4. A water-saving irrigation canal according to claim 3, characterized in that: A front filter (305) is fixedly connected to one side of the water channel body (1), and one side of the front filter (305) and one side of the side connecting plate (306) are arranged in parallel along the surface of the water channel body (1).

5. A water-saving irrigation canal according to claim 4, characterized in that: A top communicating hole (3012) is provided at the top of the filter box (3011), and a bottom square frame (3014) is provided at the bottom of the filter box (3011).

6. A water-saving irrigation canal according to claim 5, characterized in that: A side discharge trough (3015) is provided on one side of the filter box (3011), and a storage frame (3016) is provided on one side of the filter box (3011).

7. A water-saving irrigation canal according to claim 6, characterized in that: The top of the bottom connecting frame (303) is fixedly connected to a supporting base frame (304), and the top of the supporting base frame (304) is arranged at the bottom of the storage frame (3016).

8. A water-saving irrigation canal according to claim 7, characterized in that: The top of the top mixing cylinder (3010) is fixedly connected to a top fixing cylinder (401), the top of the top mixing cylinder (3010) is fixedly connected to a motor body (403), the output shaft of the motor body (403) is fixedly connected to a transmission rod (404), the surface of the transmission rod (404) is fixedly connected to an agitating blade (405), and a top connecting hole (402) is opened at the top of the top mixing cylinder (3010).

9. A method for regulating a water-saving irrigation canal, applied to a water-saving irrigation canal as claimed in any one of claims 1 to 8, characterized in that: The steps include: S1, when the liquid flows along the water channel body (1), it will contact the front filter (305) and filter and block larger debris through the front filter (305). The filtered liquid will flow to one side of the side connecting plate (306) through the front filter (305) and accumulate on one side of the side connecting plate (306). The liquid will be started by the water pump body (309) and driven by the water pump body (309) to be pumped into the top mixing barrel (3010) along the connecting pipe (308). The liquid in the top mixing barrel (3010) will flow into the filter box (3011) by gravity through the opening of the top connecting hole (3012); S2, the liquid entering the filter box (3011) will contact the inner filter screen (3013) and be filtered through the inner filter screen (3013) with a smaller gap, blocking and filtering the smaller sand and gravel particles in the liquid, and the inner filter screen (3013) is in an inclined state, and the sand and gravel particles stay on the surface of the inner filter screen (3013) and fall and roll along the inclined angle, and the sand and gravel filtered on the surface of the inner filter screen (3013) will slide along the side discharge trough (3015) into the surface of the storage frame (3016), and the sand and gravel will be uniformly stored through the storage frame (3016) on the surface of the supporting bottom frame (304), and the filtered liquid in the filter box (3011) will flow downward along the bottom square frame (3014) and enter the water channel body (1) to flow; S3, the operator first holds the top adjustment handle (2011) and rotates it, and the rotating top adjustment handle (2011) drives the threaded rod (2010) to rotate, and the threaded connection between the threaded rod (2010) and the top threaded hole (206) is used to adjust the height of the top sealing plate (207) during the threaded rotation of the threaded rod (2010), and the bottom end of the threaded rod (2010) is movably connected to the top sealing plate (207) through the bottom bearing disk (2012), and the connection through the bottom bearing disk (2012) can maintain the stability of the top sealing plate (207) when the bottom bearing disk (2012) rotates, and by adjusting the height of the top sealing plate (207), the height of the adjustment plate (208) on the inner side of the top connecting frame (204) is adjusted; S4. Different microorganisms or liquid fertilizers can be added into the top mixing barrel (3010) through the top fixed barrel (401). The added liquid will enter the top mixing barrel (3010) along the top fixed barrel (401). At this time, the motor body (403) is started to drive the transmission rod (404) to rotate along the inner side of the top connecting hole (402). The rotating transmission rod (404) will drive the stirring piece (405) to stir along the inner side of the top mixing barrel (3010).

10. An intelligent control method for a water-saving irrigation canal, characterized in that: Based on the combination of deep reinforcement learning and quantum self-supervised contrastive learning algorithm, the canal is optimized and controlled, including the following steps: Step 1: System Initialization Quantum Adaptive Hybrid Graph Neural Network Modeling Threaded rod (2010) node: The key physical properties of a threaded rod include the rotation angle θ(t) and the rotation speed ω(t); these variables are monitored in real time by a rotation sensor mounted on the threaded rod; Top sealing plate (207) node: The key characteristics of the top sealing plate include its height h(t) and the water pressure p(t) it bears. The height of the top sealing plate directly affects the water flow f(t), while the water pressure p(t) reflects the pressure of the water flow on the top sealing plate. The water pressure p(t) depends on the height of the top sealing plate and the water flow velocity, reflecting the force of the water flow on the system. These variables are calculated through a quantum adaptive hybrid graph neural network, and the interaction between nodes and edges simulates the dynamic relationship between the top sealing plate and the threaded rod. Water pump node: The working state of the water pump and the water flow rate v(t) are the key control parameters; the working state of the water pump is represented by the binary state of the node, while the water flow rate is provided by real-time data; The quantum adaptive hybrid graph neural network simulates the relationship between the pump state and the water flow change to predict the flow adjustment effect in real time; Sensor nodes: The sensor nodes monitor the real-time data of water flow f(t), water pressure p(t), and temperature T(t); these data are transmitted to other key nodes, threaded rods, and top sealing plates through the connections between nodes to adjust their operations; Sensor modeling: Each sensor node inputs real-time data into the deep reinforcement learning model and quantum self-supervised contrastive learning model, and adjusts the system control strategy by dynamically updating the node status; The relationship between nodes and edges in graph neural networks An important feature of quantum adaptive hybrid graph neural networks is that the edges between nodes represent the dynamic interactions between physical devices; The edge of the threaded rod and the top sealing plate: The rotation of the threaded rod directly determines the height of the top sealing plate, which is determined by the edge weight w ij To express; the influence of the rotation speed ω(t) of the threaded rod on the height change Δh(t) of the top sealing plate is expressed by a linear relationship, and the weight of the edge is: Δh(t)=w ij ·ω(t) where w ij represents the coupling coefficient between the threaded rod and the top sealing plate, which determines the effect of each rotation of the threaded rod on the height of the top sealing plate, ω(t) is the rotation speed of the threaded rod, and Δh(t) is the height change of the top sealing plate; By updating these weights, the system is able to accurately simulate the complex physical interactions in water flow regulation; The edge between the water pump and the water flow sensor: The on / off state of the water pump affects the water flow, which is represented by the weight of the edge. The water flow f(t) depends on the working state of the water pump and the flow rate v(t): f(t)=w ij ·v(t) where w ij represents the efficiency coefficient of the water pump, describing the direct impact of the water pump state on the water flow rate, v(t) is the water pump flow rate, and f(t) is the water flow rate; Edges between sensor nodes and control nodes: Sensor data is transmitted to control nodes through edges and used to adjust their operations; if the sensor detects that the water pressure is too high, the system adjusts the rotation speed ω(t) of the threaded rod through the weight of the edge to reduce the water flow; Sensor placement Water flow sensor f(t): arranged at the inlet and outlet of the canal to monitor the change of water flow in real time; Water pressure sensor p(t): arranged near the top sealing plate, monitoring the pressure of water flow on the top sealing plate, used to control the rotation angle of the threaded rod; Threaded rod rotation sensor ω(t): monitors the rotation speed of the threaded rod and helps adjust the height of the top sealing plate; Temperature and impurity sensors: monitor the ambient temperature and impurity content in the water channel to ensure that the system can operate stably in different environments; Data collection and preprocessing All data collected by sensors will be used as input to the deep reinforcement model to adjust the regulation strategy in real time: Dynamic sampling frequency: When the water flow fluctuates violently, the system will increase the data collection frequency to ensure that important changes can be captured; Data filtering and cleaning: Before data is input into the deep reinforcement model, the system will perform preprocessing to remove noise and abnormal data to ensure the accuracy of the adjustment; Cache mechanism: Sensor data is transmitted to the central control system in real time, and a certain amount of historical data is cached for the quantum self-supervised comparative learning model to learn and use, thereby optimizing future adjustment decisions; Step 2: Deep reinforcement learning decision system In this step, the system models the water flow regulation problem as a Markov decision process through a deep reinforcement learning model, so that the system can adaptively learn the regulation strategy to ensure that the water flow reaches the target value while optimizing energy consumption and response time; The following is a detailed expansion of each part, and the application of each variable in the present invention is explained in combination with specific formulas; Deep reinforcement learning model construction State space construction: In the deep reinforcement learning model, the state space s(t) represents the overall state of the system at time t, covering all key variables that affect the system performance; these variables are data that the system monitors in real time through sensors and inputs into the deep reinforcement learning model; the construction of the state space is the basis of model decision-making, and the main variables are as follows: Water flow rate f(t): The core variable of the system, indicating the current water flow in the canal; it is obtained in real time through the water flow sensor and compared with the target value f target For comparison; water flow is the key target of system regulation. The deep reinforcement learning model decides whether to adjust the speed ω(t) of the threaded rod and the height h(t) of the top sealing plate based on the real-time feedback of the water flow; s(t)=f(t) Among them, s(t) is the state space, f(t) is the water flow; Height of top sealing plate h(t): The height of top sealing plate directly controls the size of water flow; the rotation angle of threaded rod will affect the height of top sealing plate, so this variable determines the effect of system regulation; the real-time height of top sealing plate h(t) is also an important part of state space: h(t)=h0+k·θ(t) Where h0 is the initial height, k is the proportional constant between the rotation of the threaded rod and the height change of the top sealing plate, h(t) is the height of the top sealing plate at time t, and θ(t) is the rotation angle of the threaded rod; Rotation speed of the threaded rod ω(t): The rotation speed of the threaded rod directly affects the movement speed of the top sealing plate, so it is a key variable for controlling the accuracy of water flow regulation. By monitoring ω(t) in real time, deep reinforcement learning can accurately adjust the height of the top sealing plate; Where θ(t) is the rotation angle of the threaded rod, and v(t) is the angular velocity of the threaded rod; In summary, the state space s(t) is expressed as: s(t)=[f(t),h(t),ω(t)] Action space construction: The action space a(t) is the adjustment operation performed by the system, which mainly includes the following actions: Top sealing plate height adjustment Δh(t): By adjusting the rotation speed ω(t) of the threaded rod, the height h(t) of the top sealing plate changes, thereby controlling the water flow; this is the main means for the system to regulate water flow; Δh(t)=k·Δθ(t) Among them, Δθ(t) is the angular change of the threaded rod rotation, Δh(t) is the height adjustment of the top sealing plate, and k is the proportional constant between the threaded rod rotation and the top sealing plate height change; Adjustment of the rotation angle of the threaded rod: The rotation angle of the threaded rod determines the height of the top sealing plate, so the rotation speed ω(t) of the threaded rod is the key control variable; the rotation angle of the threaded rod affects the moving speed of the top sealing plate, thereby affecting the response speed of the water flow regulation; Reward function design: The reward function R(s(t), a(t)) is the core of the deep reinforcement learning model and is used to measure whether the adjustment action currently performed by the system is effective. A reasonable reward function is designed to guide the system to achieve the optimization goal: Difference between water flow and target water flow: The reward function encourages the water flow f(t) to be close to the target water flow f target ; By using the form of square error, the system can effectively reduce the deviation between the water flow and the target value: R(s(t),a(t))=-(f(t)-f target ) 2 Among them, R(s(t),a(t)) is the reward function, f(t) is the water flow, and f target is the target water flow rate; when the water flow rate f(t) is closer to f target , the smaller the negative value of this part, the higher the reward; Energy consumption penalty: In order to reduce the energy consumption of the system, a penalty term of the rotation speed ω(t) of the threaded rod is introduced into the reward function; the faster the rotation speed of the threaded rod, the higher the energy consumption, so a penalty term proportional to the speed is designed: R(s(t),a(t))=-(f(t)-f target ) 2 -c·v(t) Among them, c is the energy consumption penalty coefficient. By adjusting the value of c, the system finds a balance between adjustment accuracy and energy consumption; Response time penalty: In some scenarios, the system needs to quickly adjust the water flow, so the response time is also one of the factors affecting the quality of regulation. A time penalty term is introduced to reward quick response operations: R(s(t),a(t))=-(f(t)-f target ) 2 -c·v(t)-λ·T(t) Where, T(t) is the response time, λ is the weight coefficient of the time penalty term; Action strategy of reinforcement learning: Based on the state space s(t) and the reward function R(s(t), a(t)), the deep reinforcement learning model can learn the optimal action strategy through feedback: When the water flow rate f(t) is significantly less than the target value f target , the system will increase the rotation speed ω(t) of the threaded rod, thereby increasing the height of the top sealing plate and increasing the water flow; when the water flow f(t) is too large, the system will slow down the speed ω(t) of the threaded rod, reduce the height of the top sealing plate, and reduce the water flow; Through multiple feedback and adjustments, the deep reinforcement learning model gradually converges to the optimal strategy and can efficiently control water flow; Step 3: Quantum self-supervised contrastive learning adaptive adjustment Historical data pattern extraction: In addition to the real-time data obtained in step 1, quantum self-supervised comparative learning also relies on historical data; the system will extract the best adjustment strategy under similar water flow conditions from the historical database; these data include operation modes under different water flow, pressure, and temperature conditions; when the water flow f(t) in the historical data is close to the current flow under a certain state, the system will refer to the best combination of adjusting the threaded rod speed v(t) and the top sealing plate height h(t) in history, and make adjustments based on the current environmental conditions; Feature screening and dimensionality reduction: The multidimensional data collected by sensors contains a large amount of irrelevant information; quantum self-supervised contrastive learning uses quantum computing technology to reduce the dimensionality of the data and extract the most representative features from it. The feature dimensionality reduction formula is: Among them, X is the original data matrix, Q is the dimension reduction matrix generated by the quantum algorithm, is the data after dimensionality reduction; Through the quantum feature extraction algorithm, the system can quickly screen out the key variables that affect the regulation effect and identify the sensitivity of water flow changes to the speed of the threaded rod; Quantum feature learning: Through quantum computing, quantum self-supervised comparative learning analyzes the multi-dimensional data of the water flow system in a very short time and quickly finds key features; when the water flow fluctuates greatly, the system processes the water flow h(t), threaded rod speed v(t) and top sealing plate height h(t) in parallel to determine which variable has the greatest impact on the current regulation; feature learning formula: Among them, w i is the weight, indicating that the variable x i Contribution to the loss function L; By calculating the impact of each variable on the water flow regulation effect, quantum self-supervised contrastive learning identifies which features should be prioritized; Fast pattern matching: The quantum self-supervised comparative learning system can quickly match the current water flow state with similar patterns in historical data through the parallelism of quantum computing: As the current water flow rate h(t) increases, the quantum self-supervised comparative learning system finds the best adjustment plan for the threaded rod speed v(t) and the top sealing plate height h(t) in history by comparing similar situations of rapid increase in water flow in history. The quantum algorithm speeds up the pattern matching, allowing the system to give the best adjustment plan in an instant. Contrastive learning model adjustment: Another important function of quantum self-supervised contrastive learning is to adjust model parameters through contrastive learning so that the system can adaptively respond to dynamically changing water flow regulation needs; Adaptive adjustment of the model: By comparing the current water flow state with similar states in historical data in real time, quantum self-supervised comparative learning dynamically adjusts the regulation strategy; When the current water flow h(t) increases, the quantum self-supervised comparative learning system finds the historically optimal adjustment plan for the threaded rod speed v(t) and the top sealing plate height h(t) by comparing similar situations in which the water flow rate increased rapidly in history: v new (t)=v opt (t history ) Among them, v opt (t history ) is the optimal threaded rod speed under similar water flow conditions in history, v new (t) is the most recently selected threaded rod speed; Real-time comparison and historical pattern matching: The system compares the current water flow status with the patterns in historical data, finds similar water flow regulation strategies in history, and applies them in real time; the specific process is as follows: The data f(t), h(t), and v(t) captured by the current sensor are input into the quantum computing model to quickly calculate its relationship with the historical state f history (t) similarity; Where f(t) is the flow rate captured by the sensor, f history (t) is the optimal flow rate under similar historical conditions, Similarity(f(t),f history (t)) represents f(t) and f history (t), the similarity value is used to judge the matching degree between the current state and the historical pattern; if the similarity is high, the system will apply the best adjustment parameters in history; ||||: represents the norm of the data, indicating the size and strength of the data; the norm is used to normalize the strength of the data to ensure that the calculation of similarity is not affected by the absolute value of the data; Prediction based on contrastive learning: The quantum self-supervised contrastive learning system can not only compare the current state with historical data, but also make predictions based on historical trends; When the current water flow rate f(t) increases rapidly, the system predicts that the water flow rate will increase further in the future and adjusts the top sealing plate height h(t) in advance to avoid water flow overload; h predicted (t)=h current (t)+α·Δf(t) Among them, h predicted (t) is the expected top sealing plate height, h current (t) is the measured top sealing plate height, α is an empirical parameter, which represents the influence coefficient of water flow change on the top sealing plate height, and Δf(t) is the change value of water flow; Dynamic adjustment and feedback mechanism: Through contrast learning and feedback mechanism, the quantum self-supervised contrast learning system can adaptively adjust the threaded rod speed and top sealing plate height to achieve the optimal adjustment effect; Real-time dynamic adjustment: When the system detects that the current water flow rate f(t) is equal to the target value f target When there is a gap, quantum self-supervised comparative learning will adjust the threaded rod rotation speed v(t) and the top sealing plate height h(t) according to the feedback to ensure that the water flow rate gradually approaches the target value; v(t+1)=v(t)+Δv(t) Among them, v(t) is the rotation speed of the threaded rod at time t, Δv(t) is the speed increment of the threaded rod adjusted according to contrast learning, and v(t+1) is the rotation speed of the threaded rod at time t+1; Model iterative update: The quantum self-supervised comparative learning system will update the model based on the feedback data after each adjustment; if the water flow rate still does not reach the target value after adjustment, the system will further optimize the adjustment parameters and iteratively improve the adjustment effect; after each adjustment, the system will update its understanding of the water flow regulation effect and be more accurate in future adjustments.

Images