Intelligent cabin washing control system for dynamic pressure regulation of ship cargo cabin based on wall surface dynamic pressure feedback

Abstract

The invention discloses a ship cargo hold dynamic pressure regulation intelligent tank washing control system based on wall surface dynamic pressure feedback. A real-time jet flow distance is obtained through a sensing module; the valve element stroke of the needed nozzle is obtained through a control function constructed by the control module; constructing a prediction network model through an intelligent algorithm module; a real-time adjustment signal of the stroke of the valve element is obtained through the control module; a valve opening degree control signal of an electric control stop valve preset on a water inlet pipeline corresponding to the nozzle is obtained through an execution module according to the adjusting signal; linear driving of valve element displacement is achieved through a driving module based on a valve opening degree control signal, and the fluctuation range of wall surface dynamic pressure is obtained through a pressure sensor; and whether the fluctuation range of the wall surface dynamic pressure meets a preset range threshold value or not is judged through the wall surface dynamic pressure evaluation module, and meanwhile, the displacement of the valve element is indirectly adjusted according to the updated driving parameters, so that the ship cargo hold is cleaned. The problems that according to an existing traditional fixed pressure spraying mode, the value of cabin washing dynamic pressure borne by the cabin wall in the area close to a cabin washing machine is too high, erosive abrasion of cabin wall materials is accelerated, and then the structural integrity and the service life of a ship are affected are solved.

Description

Technical Field

[0001] This invention relates to the field of automated cleaning control technology for ship cargo holds, and in particular to an intelligent cargo hold cleaning control system based on wall dynamic pressure feedback. Background Technology

[0002] Current automated tank cleaning methods for chemical tankers primarily involve adjusting the rotation speed of the cleaning machine during its automatic operation to complete the cleaning work according to the actual cargo hold conditions. However, current automated tank cleaning methods still suffer from problems such as poor adaptability, low level of automation, long cleaning cycles, and high consumption of cleaning resources. During the cleaning process, the cleaning machine cleans the cargo hold walls with a constant jet pressure. Because the jet distance varies depending on the location of the tank wall, the dynamic pressure value reaching the wall varies significantly. This means that tank walls farther from the cleaning machine's outlet may not be cleaned effectively, thus prolonging operation time and reducing economic efficiency. Increasing the constant cleaning pressure of the cleaning machine would result in excessively high dynamic pressure values ​​on tank walls closer to the machine, accelerating erosion and wear of the tank wall materials, thereby affecting the structural integrity and service life of the ship. Summary of the Invention

[0003] This invention provides an intelligent tank cleaning control system for dynamic pressure regulation of ship cargo holds based on wall dynamic pressure feedback, in order to overcome the above-mentioned technical problems.

[0004] To achieve the above objectives, the technical solution of the present invention is as follows: A ship cargo hold dynamic pressure regulation intelligent tank cleaning control system based on wall dynamic pressure feedback includes a sensing module, a control module, an intelligent algorithm module, an execution module, a drive module, and a wall dynamic pressure assessment module; The sensing module is used to obtain the real-time jet distance between the current nozzle position and the bulkhead when the preset cleaning nozzle rotates at different angles; The control module is used to calculate the required nozzle valve core stroke based on the constructed control function and the real-time jet distance, with the set target wall dynamic pressure as the control target. The intelligent algorithm module is used to construct a predictive network model based on a pre-set feedforward BP neural network to predict the mapping relationship between real-time jet distance, target wall dynamic pressure and valve core stroke, and to embed the predictive network model into the pre-set controller for operation. The control module is also used to achieve real-time adjustment of the valve core stroke based on the predictive network model combined with the driving parameters of the preset stepper motor device, and to obtain the real-time valve core stroke adjustment signal. Furthermore, the driving parameters include at least the number of steps converted from stepper motor and lead screw transmission, lead screw lead, reduction ratio, basic number of steps, and subdivision coefficient; The execution module is used to obtain the valve opening control signal of the electrically controlled shut-off valve preset on the corresponding water inlet pipeline of the nozzle according to the adjustment signal, so as to realize the adjustment of the nozzle inlet pressure; The drive module is used to linearly drive the valve core displacement based on the valve opening control signal and according to the mechanical transmission components of the preset stepper motor and lead screw / linear guide rail. It also obtains the fluctuation data of the dynamic pressure on the wall in real time through the sampling data fed back by the pressure sensor preset on the inner wall of the ship's cargo hold. The wall dynamic pressure assessment module is used to determine whether the fluctuation data of the wall dynamic pressure meets the preset range threshold. If it does, no processing is required and the cleaning operation of the ship's cargo hold is continued through the cleaning nozzle. If it does not meet the threshold, the control module adjusts the drive parameters of the preset stepper motor device to obtain updated drive parameters. At the same time, the valve core displacement is indirectly adjusted based on the updated drive parameters, and the cleaning operation of the ship's cargo hold is continued through the cleaning nozzle.

[0005] Furthermore, the formula for constructing the control function is as follows: , In the formula: Indicates the required valve core stroke for the nozzle; This indicates the set target wall dynamic pressure; This indicates the real-time jet distance.

[0006] Furthermore, based on a pre-set feedforward BP neural network, a predictive network model is constructed to predict the mapping relationship between real-time jet distance, target wall dynamic pressure, and valve core stroke. This includes the following steps: S100: The real-time jet distance, target wall dynamic pressure, and required nozzle valve core stroke are used as sample data; the real-time jet distance and target wall dynamic pressure are used as feature data, and the valve core stroke is used as data label. The sample data are divided into training set and test set according to a preset ratio. S101: Preset a feedforward BP neural network and set the initial weights and initial biases of the network, wherein the feedforward BP neural network includes an input layer, a hidden layer and an output layer connected in sequence. The input layer is used to process sample data vectors. x = [P, D]^T input to the hidden layer, where P represents the target wall dynamic pressure, D represents the real-time jet distance, and T represents the transpose; The hidden layer is used to perform nonlinear mapping on the sample data based on the activation function to obtain the mapping vector; and the activation function The expression is , The output layer is used to obtain the predicted valve core stroke based on the mapping vector; S102: Train the pre-set feedforward BP neural network using the training set to obtain the trained feedforward BP neural network; S103: Use the mean squared error function as the network loss function, and use the test set to confirm whether the network output of the trained feedforward BP neural network has converged. If so, the feedforward BP neural network trained at this time will be used as a prediction network model for the mapping relationship between real-time jet distance, target wall dynamic pressure and valve core stroke. If not, then based on the backpropagation method, the weights and biases of the trained feedforward BP neural network are adaptively updated, and S102 is executed repeatedly.

[0007] Furthermore, the method for obtaining the real-time valve core stroke adjustment signal specifically includes the following steps: S200: Based on a predictive network model combined with preset stepper motor drive parameters, it achieves real-time adjustment of the valve core stroke to obtain the valve core displacement error. ; S201: Based on valve core displacement error Real-time valve core stroke adjustment signal; That is, to determine the valve core displacement error. The range of values ​​determines the rotation direction of the stepper motor; like Then, the displacement of the valve core stroke is increased by driving the stepper motor to rotate clockwise; like Then, the displacement of the valve core stroke is reduced by driving the stepper motor to rotate counterclockwise; The valve stroke corresponding to the cleaning nozzle used for cleaning ship cargo holds is adjusted by obtaining the real-time valve stroke adjustment signal.

[0008] Furthermore, the expression for obtaining the valve core displacement error in S200 is as follows: , , In the formula: This represents the error in the real-time valve core stroke, i.e., the valve core displacement error. This indicates the number of steps converted between a stepper motor and a lead screw drive. Indicates the lead of the lead screw; Indicates the reduction ratio; This indicates the number of rotations of the stepper motor; Indicates the base number of steps; Indicates the subdivision coefficient; This indicates the current position of the valve core stroke.

[0009] Beneficial Effects: This invention provides an intelligent tank cleaning control system for ship cargo holds based on wall dynamic pressure feedback. By constructing an integrated closed-loop control system with dynamic sensing, adaptive adjustment, and real-time feedback capabilities, compared with the traditional fixed spray pressure method, this system can automatically adjust the outlet pressure during the rotation of the cleaning nozzles, ensuring that the sprayed water jets maintain constant wall dynamic pressure at different distances. This improves cleaning efficiency, protects the bulkhead structure, and saves energy and water. It is also suitable for various ship types and complex working conditions. By embedding the predictive network model obtained through a feedforward BP neural network into the pre-set controller, intelligent and unmanned control logic is further realized, reducing manual intervention, ensuring operational safety, and significantly enhancing the system's engineering applicability and promotional value in the field of intelligent ship tank cleaning. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic diagram of the intelligent tank cleaning control system for dynamic pressure regulation of ship cargo holds based on wall dynamic pressure feedback, as described in this invention. Figure 2 This is a flowchart of the intelligent tank cleaning control method for dynamic pressure regulation of ship cargo holds based on wall dynamic pressure feedback in this embodiment; Figure 3 This is a schematic diagram of the cleaning nozzle in this embodiment; Figure 4 This is a block diagram of the intelligent tank cleaning control technology roadmap for dynamic pressure regulation of ship cargo holds based on wall dynamic pressure feedback in this embodiment. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0013] This embodiment provides an intelligent tank cleaning control system for dynamic pressure regulation in ship cargo holds based on wall dynamic pressure feedback, such as... Figure 1 and Figure 2As shown, it includes a sensing module, a control module, an intelligent algorithm module, an execution module, a drive module, and a wall dynamic pressure assessment module; The sensing module is used to obtain the real-time jet distance between the current nozzle position and the bulkhead when the preset cleaning nozzle rotates at different angles. The sensing module in this embodiment can be used to acquire the jet distance between the nozzle and the bulkhead and optional process quantities such as pressure / flow rate in real time, so as to provide input for closed-loop control. Specifically, the sensing module can use any one or more combinations of millimeter wave ranging, laser ranging or visual ranging. It is preferable to deploy a ranging unit with strong resistance to steam and water mist interference near the cleaning nozzle, and communicate with the control module through a preset analog / serial interface to balance measurement stability and adaptability to the cabin environment. The configuration of this sensing module makes it easy to maintain reliable sensing under different ship types, different media and different operating conditions. In this embodiment, the cleaning nozzle is specifically a movable inner needle nozzle, such as... Figure 3 As shown, it includes a nozzle body 1, an inner needle assembly 2, a sealing mechanism 3, and a connection interface 4. The nozzle body 1 is a corrosion-resistant metal tube connected to a high-pressure water inlet pipeline. The inner needle assembly 2 is located inside the nozzle and is axially movable. By changing the extension of the inner needle, the equivalent area of ​​the nozzle outlet is adjusted, thereby affecting the flow rate and jet shape. The sealing mechanism 3 is used to ensure reliable sealing under high-pressure conditions. The connection interface 4 is used to connect to external pipelines and support structures for easy assembly and maintenance. In addition, the inner needle assembly 2 can be displacement controlled by a linear drive mechanism. Preferably, a stepper motor-screw or electric push rod structure is used to achieve micro-step adjustment to match the requirements of different working distances and cleaning intensities. In this embodiment, to improve environmental adaptability, the end of the inner needle can be made of wear-resistant and corrosion-resistant materials according to the properties of the medium. The nozzle and support structure can be fixed by flanges or quick-release clamps for easy replacement and maintenance. The control module is used to calculate the required nozzle valve core stroke based on the constructed control function and the real-time jet distance, with the set target wall dynamic pressure as the control target. Specifically, the formula for constructing the control function is as follows: (1), In the formula: Indicates the required valve core stroke for the nozzle; This indicates the set target wall dynamic pressure; Indicates the real-time jet distance; The intelligent algorithm module is used to construct a predictive network model based on a pre-set feedforward BP neural network to predict the mapping relationship between real-time jet distance, target wall dynamic pressure and valve core stroke, and to embed the predictive network model into a pre-set controller for operation. In this embodiment, a predictive network model is constructed to map the real-time jet distance, target wall dynamic pressure, and valve core stroke. The specific steps include: S100: The real-time jet distance, target wall dynamic pressure, and required nozzle valve core stroke are used as sample data; the real-time jet distance and target wall dynamic pressure are used as feature data, and the valve core stroke is used as data label. The sample data are divided into training set and test set according to a preset ratio. S101: Preset a feedforward BP neural network and set the initial weights and initial biases of the network, wherein the feedforward BP neural network includes an input layer, a hidden layer and an output layer connected in sequence. The input layer is used to process sample data vectors. x = [P, D]^T input to the hidden layer, where P represents the target wall dynamic pressure, D represents the real-time jet distance, and T represents the transpose; The hidden layer is used to perform nonlinear mapping on the sample data based on the activation function to obtain the mapping vector; and the activation function The expression is (2), The output layer is used to obtain the predicted valve core stroke based on the mapping vector; S102: Train the pre-set feedforward BP neural network using the training set to obtain the trained feedforward BP neural network; S103: Use the mean squared error function as the network loss function, and use the test set to confirm whether the network output of the trained feedforward BP neural network has converged. If so, the feedforward BP neural network trained at this time will be used as a prediction network model for the mapping relationship between real-time jet distance, target wall dynamic pressure and valve core stroke. If not, then based on the backpropagation method, the weights and biases of the trained feedforward BP neural network are adaptively updated, and S102 is executed repeatedly.

[0014] The intelligent algorithm module in this embodiment is used to establish the mapping relationship between "nozzle jet distance - cargo hold target wall dynamic pressure - valve position stroke or displacement". It can adopt any of the following strategies, including but not limited to neural network control, fuzzy control or PID-NN hybrid control. In this embodiment, a three-layer feedforward BP neural network is preferably used. After offline training, it is deployed in the control system with embedded code for online invocation. In this way, when the jet distance between the cleaning nozzle and the cargo hold wall changes due to the movement of the cleaning nozzle, it can quickly give the valve position target value and suppress jet pressure overshoot. This prevents the problem of excessively high cleaning dynamic pressure value on the tank wall in the area close to the cleaning machine, which accelerates the erosion and wear of the tank wall material. This greatly improves the structural integrity and service life of the ship's cargo hold. The control module is also used to achieve real-time adjustment of the valve core stroke based on the prediction network model combined with the preset driving parameters of the stepper motor device, and to obtain the real-time valve core stroke adjustment signal; and the driving parameters include at least the number of steps converted from the stepper motor and the lead screw drive, the lead screw lead, the reduction ratio, the basic number of steps, and the subdivision coefficient. The method for obtaining the adjustment signal of the real-time valve core stroke in this embodiment specifically includes the following steps: S200: Based on a predictive network model combined with preset stepper motor drive parameters, it achieves real-time adjustment of the valve core stroke to obtain the valve core displacement error. ; Specifically, the expression for obtaining the valve core displacement error is as follows: (3), , In the formula: This represents the error in the real-time valve core stroke, i.e., the valve core displacement error. This indicates the number of steps converted between a stepper motor and a lead screw drive. Indicates the lead of the lead screw; Indicates the reduction ratio; This indicates the number of rotations of the stepper motor; Indicates the base number of steps; Indicates the subdivision coefficient; Indicates the current valve core stroke position; S201: Based on valve core displacement error Real-time valve core stroke adjustment signal; That is, to determine the valve core displacement error. The range of values ​​determines the rotation direction of the stepper motor; like Then, the displacement of the valve core stroke is increased by driving the stepper motor to rotate clockwise; like Then, the displacement of the valve core stroke is reduced by driving the stepper motor to rotate counterclockwise; And by using the real-time valve core stroke adjustment signal obtained, the valve core stroke corresponding to the cleaning nozzle used for cleaning the ship's cargo hold is adjusted; In this embodiment, the control module is used to complete data acquisition, control calculation (by combining the constructed control function with the real-time jet distance to obtain the required nozzle valve core stroke) and command output of the sensing module. It can be implemented using a pre-set MCU, PLC or industrial computer platform, and can be configured with analog signal acquisition, pulse / step output and communication interface (such as RS-485 / Mod bus, CAN or Ethernet) to interface with the sensing system and the drive / execution system. Through the configuration of this multi-platform, standardized interface, it is easy to achieve flexible deployment and expansion under different scale and cost constraints. The execution module is used to obtain the valve opening control signal of the electrically controlled shut-off valve preset on the water inlet pipeline corresponding to the nozzle according to the adjustment signal, so as to realize the adjustment of the nozzle inlet pressure. In this embodiment, the nozzle inlet pressure is dynamically adjusted by an execution module, preferably an electrically controlled shut-off valve installed on the water inlet pipeline. This electrically controlled shut-off valve includes, but is not limited to, needle valves, ball valves, or butterfly valves, and its valve core can achieve smooth displacement within a controllable stroke. This allows for indirect adjustment of the cargo hold wall dynamic pressure by changing the nozzle inlet pressure. Furthermore, the execution module may include preset necessary limit and position detection structures to ensure the repeatability and safety of valve position control. The electrically controlled shut-off valve of the execution module preferably adopts a valve type suitable for high-pressure water media, with the valve core achieving smooth displacement within a controllable stroke (e.g., 0 to several millimeters) to change the nozzle inlet pressure and indirectly regulate the wall dynamic pressure. The drive module converts the pulse / direction signal output by the control module into valve core displacement and, in conjunction with mechanical transmission components (lead screw, linear guide, or push rod), achieves high repeatability and low hysteresis. The drive module is used to linearly drive the valve core displacement based on the valve opening control signal and according to the mechanical transmission components of the preset stepper motor and the lead screw / linear guide rail, and to obtain the fluctuation data of the dynamic pressure on the wall in real time through the sampling data fed back by the pressure sensor preset on the inner wall of the ship's cargo hold. In this embodiment, the method of linearly driving the valve core displacement based on the valve opening control signal and the mechanical transmission assembly of the preset stepper motor and the lead screw / linear guide is a known existing technical means; that is, the drive module is used to convert the control system command into the valve core displacement, and the conversion formula is shown in equation (3); specifically, it is preferable to use the mechanical transmission assembly of the stepper motor and the lead screw / linear guide to achieve linear drive; it can also be achieved by using a servo motor or electric actuator or other equivalent schemes; and in this embodiment, the drive module should have the characteristics of fine-tuning control, fast start and stop response and sufficient output torque to meet the adjustment requirements of small valve position step, low hysteresis and high repeatability accuracy.

[0015] The wall dynamic pressure assessment module is used to determine whether the fluctuation data of the wall dynamic pressure meets the preset range threshold. If it does, no processing is required and the cleaning operation of the ship's cargo hold is continued through the cleaning nozzle. If it does not meet the threshold, the control module adjusts the drive parameters of the preset stepper motor device to obtain updated drive parameters. At the same time, the valve core displacement is indirectly adjusted based on the updated drive parameters, and the cleaning operation of the ship's cargo hold is continued through the cleaning nozzle.

[0016] The system workflow described in this embodiment is as follows: After the system is powered on, the control module performs initialization and self-test procedures, including: sensor linearity and communication status detection, drive module zero / limit position verification, execution module opening and closing sensitivity test, communication link and storage resource detection, etc.; if any critical unit malfunctions, it enters safety mode, maintains valve position safety, stops motion output, and prompts for maintenance.

[0017] During the operation, the control module follows Figure 4 The process operates continuously with closed-loop pressure regulation: when the nozzle rotates or translates, causing an increase in distance, the algorithm outputs a corresponding increase in valve opening to compensate for pressure attenuation; when the distance decreases, the valve converges to avoid near-field overpressure, thereby maintaining a relatively constant dynamic pressure on the wall surface at different spray distances, ensuring cleaning uniformity and reducing the risk of erosion to the bulkhead and coating. At the end of the operation, the system executes the finishing logic: the valve position returns to zero or to the maintenance position, the drive module is unloaded, key operating parameters (valve position—distance—time sequence, etc.) are recorded, and status reporting is completed; nozzle flushing and inner needle reset can be performed if necessary. To adapt to different ship types and media, this embodiment allows equivalent replacements without departing from the core concept: the sensing module can be replaced between millimeter-wave, laser, or vision; the execution module can be replaced between different valve types; the drive module can use stepper, servo, or electric actuators; the intelligent algorithm module can be replaced with fuzzy or adaptive control. Compared with existing constant pressure cleaning methods, the system described in this embodiment can dynamically adjust the pressure according to the actual distance of the jet, avoiding the problem of "overpressure in the near zone and underpressure in the far zone". At the same time, the system has a compact structure, is easy to integrate, and is suitable for upgrading existing tank cleaning pump systems or developing new intelligent tank cleaning equipment, showing broad engineering application prospects. The dynamic pressure adjustment workflow in this embodiment includes: the sensing system collects the real-time distance between the nozzle and the tank wall → the control system generates the target wall dynamic pressure → the intelligent algorithm system calculates the corresponding valve position target → the control system outputs a control signal to the drive system → the drive system drives the execution system to complete the valve position adjustment → the sensing / estimation link feeds back the latest process quantity → the control system compares the target with the feedback and iteratively updates until the cleaning task is completed. The above process corresponds to a three-layer structure of "module logic—action link—closed-loop feedback", which facilitates engineering implementation and maintenance.

[0018] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A wall surface dynamic pressure feedback-based intelligent cleaning control system for dynamically adjusting the pressure of a cargo hold of a ship, characterized in that, The control module is configured to take the set target wall surface dynamic pressure as a control target, and calculate the required valve core stroke by combining the real-time jet flow distance based on the constructed control function. The control module is configured to take the set target wall surface dynamic pressure as a control target, and calculate the required valve core stroke by combining the real-time jet flow distance based on the constructed control function. The intelligent algorithm module is configured to construct a prediction network model of the mapping relationship among the real-time jet flow distance, the target wall surface dynamic pressure and the valve core stroke based on the preset feedforward BP neural network, and embed the prediction network model into a preset controller for internal operation. The control module is further configured to realize real-time adjustment of the valve core stroke based on the prediction network model and the driving parameters of the preset stepper motor device, and obtain an adjustment signal of the real-time valve core stroke. The driving parameters at least include the step number, lead screw pitch, reduction ratio, basic step number and subdivision coefficient of the conversion of the stepper motor and the lead screw transmission. The execution module is configured to obtain a valve opening degree control signal of an electric control stop valve preset on a water inlet pipeline corresponding to the nozzle based on the adjustment signal, so as to realize adjustment of the nozzle inlet pressure. The driving module is configured to realize linear driving of the valve core displacement based on the valve opening degree control signal and the preset mechanical transmission assembly of the stepper motor cooperating with the lead screw / linear guide rail, and obtain fluctuation data of the wall surface dynamic pressure in real time through the sampling data fed back by the pressure sensor preset on the inner wall of the ship cargo hold. The wall surface dynamic pressure evaluation module is configured to judge whether the fluctuation data of the wall surface dynamic pressure satisfies a preset range threshold value. The control function is constructed by the following formula 2. The intelligent cleaning control system based on wall dynamic pressure feedback for dynamic pressure regulation of ship cargo hold according to claim 1, characterized in that, The method for constructing the prediction network model of the mapping relationship among the real-time jet flow distance, the target wall surface dynamic pressure and the valve core stroke based on the preset feedforward BP neural network specifically includes the following steps: In the formula: represents the spool stroke of the required nozzle; represents the set target wall surface dynamic pressure; represents the real-time jet flow distance.

3. The intelligent cleaning control system based on wall dynamic pressure feedback for dynamic pressure regulation of ship cargo hold according to claim 1, characterized in that, S100: Take the real-time jet flow distance, the target wall surface dynamic pressure and the valve core stroke of the required nozzle as sample data; take the real-time jet flow distance and the target wall surface dynamic pressure as feature data, and take the valve core stroke as a data label; and divide the sample data into a training set and a test set according to a preset proportion. The output layer is configured to obtain a predicted valve core stroke according to a mapping vector. S102: Train the preset feedforward BP neural network through the training set to obtain a trained feedforward BP neural network. The input layer is used to input a sample data vector x = [P, D]^T into the hidden layer, wherein P represents a target wall surface dynamic pressure, D represents a real-time jet distance, and T represents transposition. The hidden layer is used for nonlinear mapping of sample data based on an excitation function to obtain a mapping vector; and the excitation function is expressed as S103: Take a mean square error function as a network loss function, and confirm whether the network output of the trained feedforward BP neural network converges through the test set. ​ ​ If yes, the trained feedforward BP neural network at this time is taken as a prediction network model for mapping relationship of real-time jet flow distance, target wall surface dynamic pressure and valve core stroke; If not, the weight and bias of the trained feedforward BP neural network are adaptively updated based on the back propagation method, and S102 is repeatedly executed.

4. The intelligent cleaning control system based on wall dynamic pressure feedback for dynamic pressure regulation of ship cargo hold according to claim 3, characterized in that, The method for obtaining the adjustment signal of the real-time valve core stroke specifically comprises the following steps: S200: Real-time adjustment of the valve core stroke is realized based on the prediction network model combined with the driving parameters of the preset stepping motor device to obtain a valve core displacement error amount ; S201: determining a valve core displacement error amount a real-time valve core stroke adjustment signal; That is, the valve core displacement error amount is judged to determine the rotation direction of the stepping motor. If then rotate the stepper motor clockwise to increase the displacement of the spool travel; If then rotate the stepper motor counterclockwise to reduce the displacement of the spool travel; The valve core stroke corresponding to the cleaning nozzle for cleaning the cargo hold of the ship is adjusted through the obtained adjustment signal of the real-time valve core stroke.

5. The intelligent cleaning control system based on wall dynamic pressure feedback for dynamic pressure regulation of ship cargo hold according to claim 4, characterized in that, The expression for obtaining the valve core displacement error amount in S200 is In the formula: represents the error amount of the real-time valve core stroke, i.e., the valve core displacement error amount; represents the step number converted by the stepper motor and the screw rod transmission; represents the lead of the screw rod; represents the reduction ratio; represents the rotation step number of the stepper motor; represents the basic step number; represents the subdivision coefficient; represents the current valve core stroke position.

Images