Precision feeding method and device for deep-sea mining based on intelligent monitoring and control

Abstract

This application proposes a method and device for precise feeding of deep-sea mining based on intelligent monitoring and control, including the following steps: multiple feed ports are provided at the bottom of the storage tank, each feed port is connected to a screw feeder, and each screw feeder has a different pitch, and the outlets of each screw feeder are connected to the same main feed pipeline; an Actor network is set up to predict action parameters based on the state parameters at the current moment; the action parameters are executed and the actual reward value and the state parameters at the next moment are obtained based on the execution results of the action parameters. When the ore concentration at the outlet of the main feed pipeline in the state parameters at the next moment meets the preset standard, the prediction of the action parameters is stopped. This solution precisely controls the feeding accuracy by providing screw feeders with different pitch sizes, installing multiple sensors at the outlet and inlet of each screw feeder, and based on the sensor data.

Description

Technical Field

[0001] The present application relates to the field of ore transportation technology, and in particular to a method and device for precise feeding of deep-sea mining based on intelligent monitoring and control. Background Art

[0002] In the field of deep-sea mining, precise feeding and slurry mixing technology plays a key role in the efficient and stable operation of the entire mining operation. However, the technology currently faces many shortcomings.

[0003] First, from the perspective of feeding accuracy, precise flow control is an important link to ensure the smooth progress of deep-sea mining. The solid content in the slurry must be strictly controlled within ±3% of the target value. If the feeding accuracy cannot be guaranteed and the solid content fluctuates too much, it will easily lead to blockage of the riser, which will not only reduce mining efficiency but also cause equipment failure, increase maintenance costs and safety risks. However, existing feeding technology is difficult to achieve such high precision requirements in flow control and is difficult to meet the growing demand of deep-sea mining.

[0004] In terms of equipment operation stability, the deep-sea environment is complex and harsh, and the feeding equipment faces severe tests. The strong corrosiveness of seawater will accelerate the corrosion damage of the equipment, shorten the equipment service life, and increase maintenance costs. At the same time, the geological conditions of the seabed are complex, and the properties of the ore (such as particle size distribution, concentration, etc.) vary greatly. This requires the feeding equipment to be able to adjust the operating parameters in time according to the changes in the ore characteristics to ensure stable feeding. However, traditional feeding equipment lacks an effective control mechanism when dealing with these complex changes, and cannot ensure the stable operation of the equipment under different working conditions. Summary of the Invention

[0005] The embodiments of the present application provide a method and device for precise feeding of deep-sea mining based on intelligent monitoring and control. The method and device are configured by setting screw feeders with different pitch sizes, installing multiple sensors at the outlet and inlet of each screw feeder, and precisely controlling the feeding accuracy based on the sensor data.

[0006] In a first aspect, an embodiment of the present application provides a method for precise feeding of deep-sea mining based on intelligent monitoring and control, the method comprising:

[0007] A plurality of feed ports are provided at the bottom of the storage tank, each of which is connected to a screw feeder, and each screw feeder has a different pitch. The discharge ports of the respective screw feeders are connected to the same main feed pipeline, and a main and auxiliary feed valve group is provided at the discharge port of each screw feeder. The main feed pipeline is connected to a water supply pump, and the main and auxiliary feed valve groups are composed of a main feed valve and an auxiliary feed valve connected in parallel, and the flux of the main feed valve is greater than that of the auxiliary feed valve;

[0008] Set up an Actor network and obtain the current state parameters. The Actor network predicts the action parameters based on the current state parameters. The state parameters include the speed of each screw feeder, the ore concentration at the discharge port, the opening of the main and auxiliary feed valve groups, the flow rate of the water supply pump, and the ore concentration at the outlet of the main feed pipeline. The action parameters include the speed of each screw feeder and the opening of the main and auxiliary feed valve groups.

[0009] Execute the action parameters, and obtain the real reward value and the state parameters at the next moment based on the execution results of the action parameters, set a Critic network, the Critic network performs a value evaluation on the action parameters based on the state parameters at the current moment to obtain a predicted reward value, update the parameters of the Actor network based on the predicted reward value, calculate the difference between the predicted reward value and the real reward value to obtain a reward difference, update the parameters of the Critic network with the reward difference, use the Actor network and the Critic network after the parameter update to predict the action parameters at the next moment, and stop predicting the action parameters when the ore concentration at the outlet of the main feed pipeline in the state parameters at the next moment meets the preset standard.

[0010] In a second aspect, an embodiment of the present application provides a maintenance detection method based on intelligent monitoring and control, comprising:

[0011] Obtain the corrosion status parameters of the corrosion maintenance parts of each screw feeder at the current moment;

[0012] Based on the optimization objective, a current network and a target network with the same initial parameters are constructed. The current corrosion state is input into the current network to obtain the Q value of the current network. Based on the current network Q value, a maintenance action is selected and executed using the ε-greedy strategy, where the Q value is the weight of each maintenance action under the current corrosion state. The maintenance actions include full maintenance, partial maintenance, and no maintenance.

[0013] At each moment, the corrosion state of the current moment is input into the target network to obtain the Q value of the target network. The target value is calculated based on the Q value of the current network and the Q value of the target network. The loss function is defined according to the target value, and the parameters of the current network are updated based on the result of the loss function. The parameters of the current network are synchronized to the target network at fixed intervals.

[0014] In a third aspect, the embodiments of the present application provide a deep-sea mining precision feeding device based on intelligent monitoring and control, comprising:

[0015] A setting module is used to set multiple feed ports at the bottom of the storage tank, each feed port is connected to a screw feeder, and the pitch of each screw feeder is different. The discharge ports of each screw feeder are respectively connected to the same main feed pipeline, and a main and auxiliary feed valve group is set at the discharge port of each screw feeder. A water supply pump is connected to the main feed pipeline. The main and auxiliary feed valve groups are composed of a main feed valve and an auxiliary feed valve connected in parallel, and the flux of the main feed valve is greater than that of the auxiliary feed valve;

[0016] The prediction module is used to set up the Actor network and obtain the current state parameters. The Actor network predicts the action parameters based on the current state parameters. The state parameters include the speed of each screw feeder, the ore concentration at the discharge port, the opening of the main and auxiliary feed valve groups, the flow rate of the water supply pump, and the ore concentration at the outlet of the main feed pipeline. The action parameters include the speed of each screw feeder and the opening of the main and auxiliary feed valve groups.

[0017] An iterative module is used to execute action parameters, obtain a real reward value and a state parameter at the next moment based on the execution result of the action parameters, set a Critic network, and the Critic network evaluates the value of the action parameters based on the state parameters at the current moment to obtain a predicted reward value. The parameters of the Actor network are updated based on the predicted reward value, and the difference between the predicted reward value and the real reward value is calculated to obtain a reward difference. The parameters of the Critic network are updated with the reward difference. The Actor network and the Critic network after the parameter update are used to predict the action parameters at the next moment. When the ore concentration at the outlet of the main feed pipeline in the state parameters at the next moment meets the preset standard, the prediction of the action parameters is stopped.

[0018] In a fourth aspect, an embodiment of the present application provides a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute a deep-sea mining precision feeding method based on intelligent monitoring and control or a maintenance and detection method based on intelligent monitoring and control.

[0019] In a fifth aspect, an embodiment of the present application provides a readable storage medium, in which a computer program is stored. The computer program includes a program code for controlling a process to execute a process, and the process includes a deep-sea mining precision feeding method based on intelligent monitoring and control or a maintenance and detection method based on intelligent monitoring and control.

[0020] The main contributions and innovations of the present invention are as follows:

[0021] The embodiment of the present application can convey ores of different particle size and gradation by using multiple screw feeders in parallel with different pitch parameters. Through the unique spiral phase design, vibrator configuration, real-time monitoring and dynamic speed regulation, the slurry particles are uniformly conveyed, the solid content fluctuation is controlled within the target value ±3%, the uniformity of solid feeding in the lifting pipe is guaranteed, and blockage is effectively prevented; the spiral shaft of the embodiment of the present application is arranged with torque and vibration sensors to monitor the conveying status in real time. Once the motor torque rises suddenly or the vibrator amplitude increases above the threshold, the protection mechanism and the reverse rotation mechanism are immediately activated, which can effectively prevent and solve the blockage problem, reduce downtime, ensure the continuous and stable feeding process, and improve production efficiency; the embodiment of the present application achieves slurry flow balance and stable system operation by training dual network optimization decision-making, improves the intelligence and accuracy of feeding control, and through the formulation of anti-corrosion maintenance strategy for the feeding mechanism, comprehensively considers multiple factors to define the reward function, rationally plans maintenance measures, minimizes equipment loss, optimizes maintenance intervals, and reduces unplanned downtime.

[0022] The details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0024] Figure 1 This is a flow chart of a method for precise feeding of deep-sea mining based on intelligent monitoring and control according to an embodiment of the present application;

[0025] Figure 2 is a structural diagram of a deep-sea mining device according to an embodiment of the present application;

[0026] Among them, 1. Storage tank, 2. Screw feeder, 3. Main feed pipeline, 4. Main and auxiliary feed valve groups, 401. Main feed valve group, 402. Auxiliary feed valve group, 5. Water supply pump.

[0027] Figure 3 is a flow chart of a maintenance detection method based on intelligent monitoring and control according to an embodiment of the present application;

[0028] Figure 4 This is a structural block diagram of a deep-sea mining precision feeding device based on intelligent monitoring and control according to an embodiment of the present application;

[0029] Figure 5 Schematic diagram of the hardware structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0030] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The implementations described in the following exemplary embodiments are not intended to represent all implementations consistent with one or more embodiments of this specification. Rather, they are merely examples of apparatuses and methods consistent with certain aspects of one or more embodiments of this specification, as detailed in the appended claims.

[0031] It should be noted that in other embodiments, the steps of the corresponding method are not necessarily performed in the order shown and described in this specification. In some other embodiments, the method may include more or fewer steps than those described in this specification. In addition, a single step described in this specification may be broken down into multiple steps for description in other embodiments, and multiple steps described in this specification may be combined into a single step for description in other embodiments.

[0032] Example 1

[0033] The embodiment of the present application provides a method for accurate feeding of deep-sea mining based on intelligent monitoring and control, which is achieved by setting screw feeders with different pitch sizes and installing multiple sensors at the outlet and inlet of each screw feeder, and accurately controlling the feeding accuracy based on the sensor data. Specifically, referring to Figure 1 , the method comprising:

[0034] A plurality of feed ports are provided at the bottom of the storage tank 1, each of which is connected to a screw feeder 2, and each screw feeder 2 has a different pitch. The discharge ports of the respective screw feeders are connected to the same main feed pipeline 3, and a main and auxiliary feed valve group 4 is provided at the discharge port of each screw feeder 2. A water supply pump 5 is connected to the main feed pipeline. The main and auxiliary feed valve group 4 consists of a main feed valve 401 and an auxiliary feed valve 402 connected in parallel, and the flux of the main feed valve 401 is greater than that of the auxiliary feed valve 402;

[0035] Set up an Actor network and obtain the current state parameters. The Actor network predicts the action parameters based on the current state parameters. The state parameters include the speed of each screw feeder, the ore concentration at the discharge port, the opening of the main and auxiliary feed valve groups, the flow rate of the water supply pump, and the ore concentration at the outlet of the main feed pipeline. The action parameters include the speed of each screw feeder and the opening of the main and auxiliary feed valve groups.

[0036] Execute the action parameters, and obtain the real reward value and the state parameters at the next moment based on the execution results of the action parameters, set a Critic network, the Critic network performs a value evaluation on the action parameters based on the state parameters at the current moment to obtain a predicted reward value, update the parameters of the Actor network based on the predicted reward value, calculate the difference between the predicted reward value and the real reward value to obtain a reward difference, update the parameters of the Critic network with the reward difference, use the Actor network and the Critic network after the updated parameters to predict the action parameters at the next moment, and stop predicting the action parameters when the ore concentration at the outlet of the main feed pipeline 3 in the state parameters at the next moment meets the preset standard.

[0037] In some embodiments, the deep sea mining equipment structure of this solution is as follows Figure 2 As shown, the storage tank 1 of the deep-sea mining equipment is used to collect the ore collected by the mining equipment, and the storage tank 1 is placed horizontally. The feed port at the bottom of the storage tank 1 is a conical feed port, and each feed port is connected to a screw feeder 2. The feed port is used to transport the ore in the storage tank to each screw feeder.

[0038] Specifically, this solution provides an overflow hole 102 on the storage tank 1, and the overflow hole 102 is used to discharge excess seawater during the ore stacking process, and an ore detection and stirring device is provided in the storage tank 1. When the detection and stirring device detects that there is a certain proportion of ore in the storage tank 1 that is stationary, it is stirred to activate this part of the ore.

[0039] In some embodiments, the staggered amount of the spiral blades of two adjacent spiral feeders 2 is half the pitch, and the spiral phase difference is 2π / n, which has the advantage of solving the periodic fluctuation problem in the spiral feeder, where n is the total number of spiral feeders.

[0040] Specifically, each screw feeder 2 in this solution is connected to the storage tank through a separate feed port, thereby ensuring that the slurry in the screw feeder will not be fed into another one.

[0041] Specifically, the interior of the screw feeder in this solution is a single-helix structure, and the blade surface of the screw feeder 2 is laser micro-textured or coated with polytetrafluoroethylene (PTFE) to reduce adhesion and prevent fine particles from sticking and causing material jamming. The main spiral variable pitch section is also equipped with a jet system equipped with a high-pressure nozzle (0.5~1MPa) to regularly spray water to remove particles accumulated at the root of the blades to ensure smooth operation of the equipment. This design ensures the efficient and stable operation of the screw feeder 2 under different slurry concentrations and particle distribution changes.

[0042] In some embodiments, the water supply pump 5 is used to supply water to the main material delivery pipeline 3 to ensure the ore concentration at the outlet of the main material delivery pipeline 3.

[0043] Specifically, in order to more accurately adjust the ore concentration at the outlet of the main feed pipeline 3, this solution sets the water output of the water supply pump 5 to a fixed value, thereby ensuring that the ore concentration at the outlet of the main feed pipeline 3 can be adjusted only by adjusting the speed of the spiral discharger 2 and the opening of the main and auxiliary feed valve groups 4.

[0044] In some embodiments, the rotational speed of each screw feeder 2, the opening of the main and auxiliary feed valve groups 4, and the flow rate of the water make-up pump 5 are obtained by capturing the parameters of the background control system, and the ore concentration at the discharge port of each screw feeder 2 is obtained by installing a flow sensor and a concentration and particle size counter at the discharge port of each screw feeder 2. The ore concentration at the outlet of the main feed pipeline 3 is then calculated from the ore concentration at the discharge port of each screw feeder 2 and the flow rate of the water make-up pump 5.

[0045] In some specific embodiments, an ultrasonic signal is sent to the discharge port of each screw feeder 2, and the flow sensor and the concentration and particle size sensor receive the returned ultrasonic signal. For the measurement of sediment particle size and concentration distribution, the attenuation effect of different ore particle sizes on ultrasonic waves is considered, and the inversion formula of ore concentration and particle size is obtained in combination with the optimal regularization algorithm. The formula is as follows:

[0046]

[0047] Where A is the coefficient matrix, F is the discretized ore size frequency distribution, matrix H is the smoothing matrix, G is the vector of ultrasonic attenuation coefficients at different frequencies obtained from actual measurements, and the parameter γ can be obtained by minimizing the following γ function:

[0048]

[0049] Where I is the identity matrix, m is the number of ultrasonic frequencies used and is also the order of the unit matrix.

[0050] In some specific embodiments, since the flow rate of the water supply pump 5 in this solution is a fixed value, this solution needs to ensure that the ore concentration at the outlet of the main conveying pipeline 3 meets the preset standard. It is sufficient to ensure that the conveying capacity of each screw feeder 2 meets the standard. The conveying capacity formula of each screw feeder 2 is expressed as follows:

[0051]

[0052] Among them, Q0 is the conveying capacity, ω is the rotation speed, c is the ore concentration, θ is the opening of the intelligent valve, Q is the flow rate of the seawater pump, c 0 is the ore concentration at the outlet of the main conveying pipeline 3, where the conveying capacity is the weight of material conveyed per unit time.

[0053] In some embodiments, an Actor network is set up and its weight coefficients are initialized. The current state parameters are input into the Actor network. The formula for predicting action parameters based on the current state parameters is as follows:

[0054]

[0055] in, d t is the action parameter predicted at the current time t, f() is the Actor network, δ f is the Actor network weight coefficient. γ t for t Random noise at any moment.

[0056] In this solution, the current state parameters are first converted into a data format that can be processed by the algorithm and then input into the Actor network. The formula is as follows:

[0057]

[0058] Among them, ω is the rotation speed, c is the ore concentration, π is the opening of the intelligent valve, Q is the flow rate of the seawater pump, c 0 is the ore concentration at the outlet of the main feed pipeline 3.

[0059] Specifically, the Actor network predicts the probability distribution of various actions based on the current state parameters, and selects the action with the highest probability as the action parameter based on the probability distribution of various action parameters.

[0060] In some embodiments, the motion parameters are transmitted to the control system of the corresponding screw feeder 2 in the form of electrical signals, and the control system adjusts the speed of the screw feeder and the opening of the main and auxiliary feeding valve groups 4.

[0061] Furthermore, when the action parameters are executed to adjust the opening of the main and auxiliary material delivery valve groups 4, if the opening adjustment value of the main and auxiliary material delivery valve groups 4 is greater than or equal to the set threshold, the opening of the main material delivery valve 401 is adjusted, and the opening of the auxiliary material delivery valve 402 remains unchanged; if the opening adjustment value of the main and auxiliary material delivery valve groups 4 is less than the set threshold, the opening of the auxiliary material delivery valve 402 is adjusted, and the opening of the main material delivery valve 401 remains unchanged.

[0062] That is to say, since the flux of the main material delivery valve 401 in this solution is greater than that of the auxiliary material delivery valve 402, if the ore concentration at the outlet of the main material delivery pipeline 3 changes greatly, this solution will adjust the opening of the main material delivery valve 401; if the ore concentration at the outlet of the main material delivery pipeline 3 changes slightly, this solution will adjust the opening of the auxiliary material delivery valve 402, thereby ensuring that the main and auxiliary material delivery valves work in coordination and the main and auxiliary material delivery valve group 4 can adapt to more highly fluctuating working conditions.

[0063] In some embodiments, the actual reward value obtained based on the execution result of the action parameters is determined by a reward function, and the formula of the reward function is expressed as follows:

[0064]

[0065] in, p 1 is the concentration difference reward function, p 2 is a sparse reward function, p 3 is the formal reward function.

[0066] The concentration difference reward function calculates rewards based on the difference between the actual concentration and the target concentration. The sparse reward function only rewards when the goal is achieved, otherwise the reward is 0. The formal reward function breaks down the task objectives into detailed components and designs multiple intermediate rewards.

[0067] In some embodiments, multiple sets of experience data are obtained and stored in an experience database, and the experience data are used to train the Actor network and the Critic network, wherein each set of experience data includes state parameters, action parameters, real reward values ​​at the current moment, and state parameters at the next moment.

[0068] In some embodiments, the parameters of the Actor network are updated by gradient descent, and the parameters of the Critic network are updated by minimizing the reward difference.

[0069] Specifically, in the training of the Actor network, the policy gradient theorem is used to calculate the policy gradient, and the weight parameters of the policy network are updated by the gradient descent method. The goal is to optimally approximate the expected policy function so that the optimal action parameters can be selected under given state parameters.

[0070] Specifically, in the training of the Critic network, the Critic network is trained through multiple sets of experience data in the experience database, and the parameters of the Critic network are continuously optimized during the training process by minimizing the reward difference.

[0071] In some specific embodiments, in order to stabilize the training process and improve the convergence speed, the target actor network and the target critic network are used for training, thereby avoiding the tight coupling between the actor network and the critic network, thereby reducing the impact of gradient noise. The formula is expressed as follows:

[0072]

[0073] in, is the target Actor network, is the target Critic network, is the weight coefficient of the target Actor network, is the weight coefficient of the target Critic network, and λ is the update coefficient.

[0074] In some embodiments, a torque sensor and a vibration sensor are arranged on the screw shaft of each screw feeder 2. When the torque sensor detects that the torque value is greater than the torque threshold or the vibration sensor detects that the amplitude value is greater than the amplitude threshold, the rotational speed of the corresponding screw feeder 2 is reduced to a safe speed. If the torque value is still greater than the torque threshold or the amplitude value is still greater than the amplitude threshold at the safe speed, the screw shaft is reversed for a first period of time.

[0075] Specifically, since the working environment of the screw feeder 2 is complex, internal blockage is prone to occur. This solution uses a torque sensor and a vibration sensor to detect the blockage of the screw feeder 2. When the torque value is detected to be greater than the torque threshold or the amplitude value is greater than the amplitude threshold, it means that the screw feeder 2 is blocked. At this time, the blockage is first relieved by reducing the speed. If the blockage is not relieved after reducing the speed, the screw shaft is controlled to reverse to loosen the blocked ore.

[0076] Furthermore, when the torque value of the screw feeder 2 is greater than the torque threshold or the amplitude value is greater than the amplitude threshold, the screw feeder 2 is considered to be a blocked screw feeder, and the execution of the action parameter corresponding to the screw feeder 2 is stopped.

[0077] Example 2

[0078] A maintenance and detection method for screw feeder based on intelligent monitoring and control, reference Figure 3 ,include:

[0079] Obtaining the corrosion state parameters of the corrosion maintenance components of each screw feeder 2 at the current moment;

[0080] Based on the optimization objective, a current network and a target network with the same initial parameters are constructed. The current corrosion state is input into the current network to obtain the Q value of the current network. Based on the current network Q value, a maintenance action is selected and executed using the ε-greedy strategy, where the Q value is the weight of each maintenance action under the current corrosion state. The maintenance actions include full maintenance, partial maintenance, and no maintenance.

[0081] At each moment, the corrosion state of the current moment is input into the target network to obtain the Q value of the target network. The target value is calculated based on the Q value of the current network and the Q value of the target network. The loss function is defined according to the target value, and the parameters of the current network are updated based on the result of the loss function. The parameters of the current network are synchronized to the target network at fixed intervals.

[0082] In some specific embodiments, the corrosion maintenance components of the screw feeder 2 include a barrel, a spiral blade, a valve core, and an inner wall of a pipe. The corrosion state parameter at the current moment is the remaining wall thickness of the corrosion maintenance component at the current moment divided by the initial wall thickness, and the formula is expressed as:

[0083]

[0084] in, is the corrosion parameter at the current time t, is the remaining wall thickness at the current time t, and D is the initial wall thickness.

[0085] In some embodiments, the Savitzky-Golay wavelet smoothing method is used to perform noise reduction processing on the corrosion pipeline degradation data obtained by online monitoring to obtain the wall thickness of each corrosion repair component, and a preventive maintenance threshold and a failure threshold are defined for each corrosion repair component. When the corrosion state parameter of the corrosion repair component exceeds the preventive maintenance threshold, it indicates that the component should be repaired. When the corrosion state parameter of the corrosion repair component exceeds the failure threshold, it indicates that the corrosion repair component should be replaced.

[0086] In some specific embodiments, the optimization goal is one of minimizing equipment loss, minimizing maintenance intervals, minimizing unplanned downtime, minimizing maintenance costs, and minimizing energy consumption. The optimization goal is set according to the needs of the implementation application.

[0087] Specifically, minimizing equipment loss is used as the optimization goal to keep the corrosion rate of the device material within the set range, minimizing the maintenance interval is used as the optimization goal to avoid additional costs caused by excessive maintenance and insufficient maintenance, minimizing unplanned downtime is used as the optimization goal to maintain the efficient operation of the screw feeder 2, minimizing maintenance cost is used as the optimization goal to reduce the cost of anti-corrosion coating, replacement parts and maintenance operations, and minimizing energy consumption is used as the optimization goal to reduce the increase in flow resistance and energy consumption caused by corrosion.

[0088] Furthermore, the reward function of the current network is defined based on the optimization objective. The formula of the reward function is expressed as:

[0089]

[0090] in, For corrosion control rewards, Penalty for unplanned downtime, Penalty for repair costs, Penalty for energy consumption, Remaining life bonus, 、 、 、 、 is the corresponding weight parameter.

[0091] In some specific embodiments, in the ε-greedy strategy, by setting the initial exploration probability And the initial exploration probability decays with the Q value training process, and the formula is expressed as:

[0092]

[0093] Among them, the initial exploration probability The value is 0.5, λ is the decay rate of exploration probability, which is 1- The probability of selecting the highest value of the current network Q value is The probability of randomly selecting a maintenance action is .

[0094] In some embodiments, the actions that can be selected at each moment for the current network and the target network in this solution include full repair, partial repair, and no repair, as expressed by the following formula:

[0095]

[0096] in, Indicates partial repair, Indicates overall maintenance, Indicates no maintenance.

[0097] In some embodiments, a state transition function is defined to represent the state transition relationship of a corrosion repair component from time t to time t+1 after the repair action, and the formula is as follows:

[0098]

[0099] in, It is the remaining wall thickness after maintenance at the last maintenance moment. For maintenance actions, is the corrosion parameter at the current time t.

[0100] In some specific embodiments, the Q value of the current network output is Q( S t , a, θ), the Q value of the target network output is ( S t , a, θ'), where is the corrosion state at the current moment, a is the corresponding maintenance action, θ is the weight of each maintenance action in the current network, and θ' is the weight of each maintenance action in the target network. The calculation formula for the target value at the current moment is as follows:

[0101]

[0102] Among them, is the target value at the current time t, is the reward value at time t+1.

[0103] The formula of the loss function defined based on the target value is expressed as:

[0104]

[0105] in, is the loss function, and N is the number of samples.

[0106] That is to say, after obtaining the loss function, the parameters of the current network are updated according to the value of the loss function, so that the maintenance actions selected by the current network are getting better and better. The formula is as follows:

[0107]

[0108] Among them, α is the learning rate.

[0109] In some specific embodiments, the formula for synchronizing the parameters of the current network to the target network at fixed time intervals is expressed as follows:

[0110]

[0111] That is to say, in this scheme, the ε-greedy strategy is used to utilize the target network Q value and the current network Q value to improve the prediction accuracy of the current network for maintenance actions.

[0112] Specifically, the iteration of the current network is terminated when the life of the corrosion maintenance component is exhausted or the number of unplanned downtimes exceeds a set threshold.

[0113] In some specific embodiments, the current network Q value at each moment and the actual corrosion state at the next moment are combined into a state-action pair. Input into the experience pool, so as to train the current network and the target network through the data in the experience pool.

[0114] Example 3

[0115] Based on the same concept, Figure 4 This application also proposes a deep-sea mining precision feeding device based on intelligent monitoring and control, including:

[0116] A setting module is provided for setting a plurality of feed ports at the bottom of the storage tank 1, each feed port is connected to a screw feeder 2, and each screw feeder 2 has a different pitch, and the discharge ports of each screw feeder are respectively connected to the same main feed pipeline 3, and a main and auxiliary feed valve group 4 is provided at the discharge port of each screw feeder 2, and a water supply pump 5 is connected to the main feed pipeline, and the main and auxiliary feed valve group 4 is composed of a main feed valve 401 and an auxiliary feed valve 402 connected in parallel, and the flux of the main feed valve 401 is greater than that of the auxiliary feed valve 402;

[0117] The prediction module is used to set up the Actor network and obtain the current state parameters. The Actor network predicts the action parameters based on the current state parameters. The state parameters include the speed of each screw feeder, the ore concentration at the discharge port, the opening of the main and auxiliary feed valve groups, the flow rate of the water supply pump, and the ore concentration at the outlet of the main feed pipeline. The action parameters include the speed of each screw feeder and the opening of the main and auxiliary feed valve groups.

[0118] An iterative module is used to execute action parameters and obtain a real reward value and a state parameter at the next moment based on the execution result of the action parameters, set a Critic network, and the Critic network performs a value evaluation on the action parameters based on the state parameters at the current moment to obtain a predicted reward value. The parameters of the Actor network are updated based on the predicted reward value, and the difference between the predicted reward value and the real reward value is calculated to obtain a reward difference. The parameters of the Critic network are updated with the reward difference. The Actor network and the Critic network after the parameter update are used to predict the action parameters at the next moment. When the ore concentration at the outlet of the main feed pipeline 3 in the state parameters at the next moment meets the preset standard, the prediction of the action parameters is stopped.

[0119] Example 4

[0120] This embodiment also provides an electronic device, referring to Figure 5 , includes a memory 404 and a processor 402, wherein the memory 404 stores a computer program, and the processor 402 is configured to run the computer program to perform the steps in any of the above method embodiments.

[0121] Specifically, the processor 402 may include a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0122] Memory 404 may include a large-capacity memory 404 for data or instructions. By way of example, and not limitation, memory 404 may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 404 may include removable or non-removable (or fixed) media. Where appropriate, memory 404 may be internal or external to the data processing device. In certain embodiments, memory 404 is non-volatile memory. In certain embodiments, memory 404 includes read-only memory (ROM) and random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically alterable ROM (EAROM) or a flash memory (FLASH), or a combination of two or more of these. In appropriate circumstances, the RAM may be a static random access memory (SRAM) or a dynamic random access memory (DRAM), wherein the DRAM may be a fast page mode dynamic random access memory 404 (FPMDRAM), an extended data output dynamic random access memory (EDODRAM), a synchronous dynamic random access memory (SDRAM), etc.

[0123] The memory 404 may be used to store or cache various data files required for processing and / or communication, as well as possible computer program instructions executed by the processor 402 .

[0124] The processor 402 reads and executes computer program instructions stored in the memory 404 to implement any one of the deep-sea mining precision feeding methods based on intelligent monitoring and control in the above embodiments.

[0125] Optionally, the electronic device may further include a transmission device 406 and an input / output device 408 , wherein the transmission device 406 is connected to the processor 402 , and the input / output device 408 is connected to the processor 402 .

[0126] Transmission device 406 can be used to receive or transmit data via a network. Specific examples of such networks may include wired or wireless networks provided by the electronic device's communications provider. In one embodiment, the transmission device includes a network interface controller (NIC), which can be connected to other network devices via a base station to enable communication with the Internet. In another embodiment, transmission device 406 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0127] The input / output device 408 is used to input or output information. In this embodiment, the input information may be state parameters at the current moment, and the output information may be reward values, action parameters, etc.

[0128] Optionally, in this embodiment, the processor 402 may be configured to execute the following steps through a computer program:

[0129] A plurality of feed ports are provided at the bottom of the storage tank 1, each of which is connected to a screw feeder 2, and each screw feeder 2 has a different pitch. The discharge ports of the respective screw feeders are connected to the same main feed pipeline 3, and a main and auxiliary feed valve group 4 is provided at the discharge port of each screw feeder 2. A water supply pump 5 is connected to the main feed pipeline. The main and auxiliary feed valve group 4 consists of a main feed valve 401 and an auxiliary feed valve 402 connected in parallel, and the flux of the main feed valve 401 is greater than that of the auxiliary feed valve 402;

[0130] Set up an Actor network and obtain the current state parameters. The Actor network predicts the action parameters based on the current state parameters. The state parameters include the speed of each screw feeder, the ore concentration at the discharge port, the opening of the main and auxiliary feed valve groups, the flow rate of the water supply pump, and the ore concentration at the outlet of the main feed pipeline. The action parameters include the speed of each screw feeder and the opening of the main and auxiliary feed valve groups.

[0131] Execute the action parameters, and obtain the real reward value and the state parameters at the next moment based on the execution results of the action parameters, set a Critic network, the Critic network performs a value evaluation on the action parameters based on the state parameters at the current moment to obtain a predicted reward value, update the parameters of the Actor network based on the predicted reward value, calculate the difference between the predicted reward value and the real reward value to obtain a reward difference, update the parameters of the Critic network with the reward difference, use the Actor network and the Critic network after the updated parameters to predict the action parameters at the next moment, and stop predicting the action parameters when the ore concentration at the outlet of the main feed pipeline 3 in the state parameters at the next moment meets the preset standard.

[0132] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementation modes, and this embodiment will not be repeated here.

[0133] In general, various embodiments may be implemented in hardware or dedicated circuitry, software, logic, or any combination thereof. Some aspects of the invention may be implemented in hardware, while other aspects may be implemented in firmware or software executed by a controller, microprocessor, or other computing device, but the invention is not limited thereto. Although various aspects of the invention may be shown and described as block diagrams, flow charts, or using some other graphical representation, it should be understood that, as non-limiting examples, the blocks, devices, systems, techniques, or methods described herein may be implemented in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or a controller or other computing device, or some combination thereof.

[0134] The embodiments of the present invention may be implemented by computer software that is executable by a data processor of a mobile device, such as in a processor entity, or by hardware, or by a combination of software and hardware. Computer software or programs (also referred to as program products) including software routines, applets and / or macros may be stored in any device-readable data storage medium and include program instructions for performing specific tasks. A computer program product may include one or more computer executable components configured to perform the embodiments when the program is run. One or more computer executable components may be at least one software code or a portion thereof. In addition, it should be noted at this point that, for example, Figure 5Any block of the logic flow in the program may represent program steps, or interconnected logic circuits, blocks and functions, or a combination of program steps and logic circuits, blocks and functions. The software may be stored on physical media such as memory chips or memory blocks implemented within the processor, magnetic media such as hard disks or floppy disks, and optical media such as, for example, DVDs and their data variants, CDs, etc. Physical media are non-transitory media.

[0135] Those skilled in the art should understand that the technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0136] The above embodiments merely illustrate several embodiments of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for precise feeding of deep-sea mining based on intelligent monitoring and control, characterized in that: The following steps are involved: A plurality of feed ports are provided at the bottom of the storage tank (1), each feed port is connected to a screw feeder (2), and each screw feeder (2) has a different pitch. The discharge ports of the respective screw feeders are respectively connected to the same main feed pipeline (3), and a main and auxiliary feed valve group (4) is provided at the discharge port of each screw feeder (2). A water supply pump (5) is connected to the main feed pipeline. The main and auxiliary feed valve group (4) is composed of a main feed valve (401) and an auxiliary feed valve (402) connected in parallel, and the flux of the main feed valve (401) is greater than that of the auxiliary feed valve (402); Set up an Actor network and obtain the current state parameters. The Actor network predicts the action parameters based on the current state parameters. The state parameters include the speed of each screw feeder, the ore concentration at the discharge port, the opening of the main and auxiliary feed valve groups, the flow rate of the water supply pump, and the ore concentration at the outlet of the main feed pipeline. The action parameters include the speed of each screw feeder and the opening of the main and auxiliary feed valve groups. Execute the action parameters, and obtain the real reward value and the state parameters of the next moment based on the execution results of the action parameters, set the Critic network, the Critic network evaluates the value of the action parameters based on the state parameters of the current moment to obtain the predicted reward value, update the parameters of the Actor network based on the predicted reward value, calculate the difference between the predicted reward value and the real reward value to obtain the reward difference, update the parameters of the Critic network with the reward difference, use the Actor network and the Critic network after the parameter update to predict the action parameters of the next moment, and stop the prediction of the action parameters when the ore concentration at the outlet of the main feed pipeline (3) in the state parameters of the next moment meets the preset standard.

2. The method for accurate feeding of deep-sea mining based on intelligent monitoring and control according to claim 1 is characterized in that: The staggered amount of the spiral blades of two adjacent spiral feeders (2) is half a pitch, and the spiral phase difference is 2π / n, where n is the total number of spiral feeders.

3. The method for accurate feeding of deep-sea mining based on intelligent monitoring and control according to claim 1, characterized in that: The rotation speed of each screw feeder (2), the opening of the main and auxiliary feeding valve groups (4) and the flow rate of the water supply pump (5) are obtained by capturing the parameters of the background control system. The ore concentration at the discharge port of each screw feeder (2) is obtained by installing a flow sensor and a concentration and particle size counter at the discharge port of each screw feeder (2). The ore concentration at the outlet of the main feeding pipeline (3) is then calculated based on the ore concentration at the discharge port of each screw feeder (2) and the flow rate of the water supply pump (5).

4. The method for accurate feeding of deep-sea mining based on intelligent monitoring and control according to claim 1, characterized in that: Set up the Actor network and initialize the weight coefficient of the Actor network. The formula for predicting action parameters based on the current state parameters of the Actor network is as follows: in, d t is the action parameter predicted at the current time t, γ t for t Random noise at the moment, f() is the Actor network, S t is the state parameter at the current moment, is the weight coefficient of the Actor network.

5. The method for accurate feeding of deep-sea mining based on intelligent monitoring and control according to claim 1 is characterized in that: When the action parameters are executed to adjust the openings of the main and auxiliary material delivery valve groups (4), if the opening adjustment value of the main and auxiliary material delivery valve groups (4) is greater than or equal to a set threshold, the opening of the main material delivery valve (401) is adjusted, and the opening of the auxiliary material delivery valve (402) remains unchanged; if the opening adjustment value of the main and auxiliary material delivery valve groups (4) is less than the set threshold, the opening of the auxiliary material delivery valve (402) is adjusted, and the opening of the main material delivery valve (401) remains unchanged.

6. The method for accurate feeding of deep-sea mining based on intelligent monitoring and control according to claim 1, characterized in that: A torque sensor and a vibration sensor are arranged on the screw shaft of each screw feeder (2). When the torque sensor detects that the torque value is greater than the torque threshold or the vibration sensor detects that the amplitude value is greater than the amplitude threshold, the rotation speed of the corresponding screw feeder (2) is reduced to a safe rotation speed. If the torque value is still greater than the torque threshold or the amplitude value is still greater than the amplitude threshold at the safe rotation speed, the screw shaft is reversed for a first time.

7. A deep-sea mining precision feeding system based on intelligent monitoring and control, characterized in that: include: A module is provided for arranging a plurality of feed ports at the bottom of a storage tank (1), each feed port being connected to a screw feeder (2), and each screw feeder (2) having a different pitch, and the discharge ports of each screw feeder being respectively connected to the same main feed pipeline (3), and a main and auxiliary feed valve group (4) being provided at the discharge port of each screw feeder (2), a water supply pump (5) being connected to the main feed pipeline, and the main and auxiliary feed valve group (4) being composed of a main feed valve (401) and an auxiliary feed valve (402) connected in parallel, and the flux of the main feed valve (401) being greater than that of the auxiliary feed valve (402); The prediction module is used to set up the Actor network and obtain the current state parameters. The Actor network predicts the action parameters based on the current state parameters. The state parameters include the speed of each screw feeder, the ore concentration at the discharge port, the opening of the main and auxiliary feed valve groups, the flow rate of the water supply pump, and the ore concentration at the outlet of the main feed pipeline. The action parameters include the speed of each screw feeder and the opening of the main and auxiliary feed valve groups. An iterative module is used to execute action parameters and obtain a real reward value and a state parameter at the next moment based on the execution result of the action parameters, set a critic network, and the critic network evaluates the value of the action parameters based on the state parameters at the current moment to obtain a predicted reward value. The parameters of the actor network are updated based on the predicted reward value, and the difference between the predicted reward value and the real reward value is calculated to obtain a reward difference. The parameters of the critic network are updated with the reward difference. The actor network and the critic network after the parameter update are used to predict the action parameters at the next moment. When the ore concentration at the outlet of the main feed pipeline (3) in the state parameters at the next moment meets the preset standard, the prediction of the action parameters is stopped.

8. A maintenance and inspection method for the deep-sea mining precision feeding system according to claim 7, characterized in that: include: Obtaining corrosion state parameters of the corrosion maintenance components of each screw feeder (2) at the current moment; Based on the optimization objective, a current network and a target network with the same initial parameters are constructed. The current corrosion state is input into the current network to obtain the Q value of the current network. Based on the current network Q value, a maintenance action is selected and executed using the ε-greedy strategy, where the Q value is the weight of each maintenance action under the current corrosion state. The maintenance actions include full maintenance, partial maintenance, and no maintenance. At each moment, the corrosion state of the current moment is input into the target network to obtain the Q value of the target network. The target value is calculated based on the Q value of the current network and the Q value of the target network. The loss function is defined according to the target value, and the parameters of the current network are updated based on the result of the loss function. The parameters of the current network are synchronized to the target network at fixed intervals.

9. An electronic device comprising a memory and a processor, characterized in that: The memory stores a computer program, and the processor is configured to run the computer program to execute the deep-sea mining precision feeding method based on intelligent monitoring and control as described in any one of claims 1 to 6 or the maintenance and detection method as described in claim 8.

10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, which includes a program code for controlling a process to execute a process, wherein the process includes a deep-sea mining precision feeding method based on intelligent monitoring and control according to any one of claims 1 to 6 or a maintenance and detection method according to claim 8.

Images