A marine intelligent disengaging hook control system and method

Abstract

The application discloses a kind of marine intelligent unhooking hook control system and method, it is related to ship control technical field, including data acquisition module, spectrum analysis module, strategy selection module, blocked determination module and stable control module;The application is effectively responded to complex and changeable sea state environment by intelligent data acquisition, analysis and strategy decision, realizes the efficient, stable and accurate control of unhooking hook action, specifically by wavelet analysis and particle swarm optimization are combined, can accurately analyze and predict the trend of sea state change, make unhooking action strategy and actual environment highly match, avoid blind or lagged operation decision, significantly improve the adaptability of system to respond to severe sea state;Real-time state feedback and dynamic control technology are fused, realize the fine control in the process of unhooking action, with the help of dynamic electromagnetic structure release intensity adjustment and accurate blocked state detection, effectively avoid the occurrence of unhooking hook action anomaly or instability phenomenon.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of ship control, in particular to a marine intelligent unhooking hook control system and method. BACKGROUND

[0002] During the sailing and berthing of a ship, the unhooking hook, as one of the key safety guarantee devices, its performance reliability, stability and response accuracy are directly related to the efficiency and safety of the ship operation, especially in severe sea conditions, it is impossible to ensure that the unhooking hook action is accurately matched with the actual environmental conditions to avoid abnormal or failure of action caused by environmental disturbance.

[0003] The current marine unhooking hook control technology mostly adopts manual experience decision or simple sensor signal feedback mechanism, and responds to general sea condition changes through pre-set action strategies. This way is relatively mature in technical implementation, has the advantages of simple structure, convenient implementation, etc., and can basically meet the needs of ship unhooking under general conditions.

[0004] However, with the increasing complexity and frequency of changes in the marine operating environment, the existing method often has problems such as strategy decision lag and large action response deviation when facing high-intensity and rapidly changing complex sea conditions, which further easily leads to unstable or even blocked unhooking action, increases equipment wear and tear and operation risk, and reduces the efficiency and safety of ship operation. SUMMARY

[0005] In order to solve the above technical problems, the present application provides a marine intelligent unhooking hook control system, comprising:

[0006] The data acquisition module S11 acquires the sea condition fluctuation data and the cable load data of the environment in which the unhooking hook is located; establishes an environmental disturbance state sequence according to the sea condition fluctuation data, and generates a dynamic load sequence according to the cable load data;

[0007] The spectrum analysis module S12 performs fast Fourier transform on the dynamic load sequence to extract the frequency spectrum features of the sequence, the frequency spectrum features including the dominant frequency spectrum amplitude and the frequency spectrum energy distribution; and constructs an unhooking hook response action preliminary strategy library according to the frequency spectrum features;

[0008] The strategy selection module S13 inputs the environmental disturbance state sequence into a pre-established deep recurrent neural network model to generate an environmental state prediction sequence; and selects an unhooking initial action strategy from the unhooking hook response action preliminary strategy library based on the environmental state prediction sequence;

[0009] The blocked judgment module S14 triggers the release action of the permanent magnet electromagnetic structure in the unhooking hook mechanism according to the unhooking initial action strategy, acquires the displacement and action torque of the release action at the release moment, and judges whether the unhooking action is blocked based on the displacement and action torque.

[0010] The stable control module S15: if the unhooking action is not blocked, real-time cable load data and unhooking mechanism displacement data are obtained, and the permanent magnet electromagnetic structure release current is dynamically adjusted to keep the unhooking mechanism action smooth.

[0011] Further, the sea state fluctuation data includes wave height, wave period and ship body attitude angle change; and the environmental disturbance state sequence is established according to the sea state fluctuation data, and specifically includes:

[0012] Collecting the sea state fluctuation data within a preset time period;

[0013] The wave height and wave period are input into a wavelet analysis model for time-frequency decomposition;

[0014] The construction steps of the wavelet analysis model are:

[0015] The historical wave height and wave period are selected as training samples to construct an initial wavelet decomposition function set, and the wavelet decomposition function is defined by two key parameters of center frequency and bandwidth;

[0016] An adaptive particle swarm optimization algorithm is used to optimize the parameters of the wavelet decomposition function set;

[0017] The optimized wavelet decomposition function set is used to generate a wavelet analysis model;

[0018] Based on the optimized wavelet analysis model, the time-frequency decomposition result of the wave signal is obtained, and the time-frequency energy distribution features are extracted, including the frequency spectrum energy distribution range, the main frequency area and the frequency change value of the wave signal energy at different frequencies and time periods;

[0019] The environmental disturbance state sequence is constructed according to the time-frequency energy distribution features and the ship body attitude angle change, and the real-time ship body attitude change data includes the roll change angle, the pitch change angle and the yaw change angle.

[0020] Further, the step of optimizing the wavelet decomposition function set parameters by using the adaptive particle swarm optimization algorithm includes:

[0021] Randomly initializing a particle swarm, each particle of the particle swarm represents a combination of wavelet decomposition function parameters, and the wavelet decomposition function parameter combination includes the center frequency and bandwidth of the wavelet, to form an initial particle swarm;

[0022] The time-frequency energy error of the training sample is used as a fitness function to evaluate the fitness value of each particle;

[0023] The particle speed and position are adjusted according to the fitness value, and the iteration is performed until the fitness reaches a predetermined convergence threshold, and the optimal parameters are output.

[0024] Further, the step of generating a dynamic load sequence according to the cable load data is:

[0025] Based on the cable tension sensor arranged on the disengaging hook bearing structure, the cable tension size borne by the disengaging hook is measured in real time and continuously;

[0026] The sampling frequency is set to 5-20 Hz, and the load raw data of 30 seconds to 2 minutes is measured and stored continuously to obtain an initial cable load data sequence

[0027] The obtained initial load data sequence is subjected to data denoising and abnormal data point correction to obtain high-quality dynamic load data;

[0028] The high-quality dynamic load data is subjected to normalization processing to generate high-quality normalized load values;

[0029] Based on multiple sets of high-quality normalized load values, a dynamic load sequence is formed.

[0030] Further, the step of constructing a disengaging hook response action preparation strategy library according to the spectral features specifically includes:

[0031] The spectral features of the dynamic load sequence are input as input variables into a nonlinear dynamics simulation model;

[0032] Based on the nonlinear dynamics simulation model, the mechanical response features of the disengaging hook action under the corresponding spectral features are simulated;

[0033] According to the mechanical response features, a disengaging hook response action strategy sequence is generated;

[0034] The spectral features and the action strategy sequence form a mapping relationship to form a preparation strategy library.

[0035] Further, the construction steps of the nonlinear dynamics simulation model are:

[0036] The friction coefficient, elastic modulus, and motion constraint conditions of the disengaging hook mechanism are collected, and a nonlinear dynamics control equation of the disengaging hook action is constructed;

[0037] Based on the actual motion characteristics of the disengaging hook mechanism, the generalized coordinates of the disengaging hook mechanism are defined;

[0038] Based on the defined generalized coordinates, the kinetic energy of the disengaging hook mechanism is calculated and obtained, and the potential energy of the elastic components of the disengaging hook mechanism is calculated and obtained; based on the obtained friction coefficient, the friction torque is calculated and obtained, and the dynamics equation of the disengaging hook mechanism is constructed according to the kinetic energy of the disengaging hook mechanism, the potential energy of the elastic components of the disengaging hook mechanism, and the friction torque;

[0039] The control equation is solved by a finite element numerical method to construct a nonlinear dynamics simulation model;

[0040] The steps of constructing the nonlinear dynamic simulation model are:

[0041] A finite element geometric model of the uncabling hook mechanism is established; the obtained friction coefficient, elastic modulus and motion constraint conditions are set as parameters and boundary conditions of the finite element simulation; numerical solution is performed on the established nonlinear control equation by using a finite element simulation software to obtain detailed dynamic response data of the uncabling hook mechanism under the frequency spectrum characteristic input condition, so as to construct a complete nonlinear dynamic simulation model.

[0042] Further, the generation steps of the deep recurrent neural network model include:

[0043] Obtaining historical environment prediction data, dividing the historical environment prediction data into a first test set and a first training set, the historical environment prediction data including an environment disturbance state sequence and a corresponding environment state prediction sequence;

[0044] Constructing a regression network, taking the environment disturbance state sequence in the first training set as the input of the regression network, taking the corresponding environment state prediction sequence as the output of the regression network, training the regression network, and obtaining an initial deep recurrent neural network;

[0045] Model verification is performed on the initial deep recurrent neural network by using the first test set, and an initial deep recurrent neural network with an output less than or equal to a preset test error threshold is output as the deep recurrent neural network model;

[0046] The logic for selecting an uncoupling initial action strategy from the uncoupling hook response action preparation strategy library based on the environment state prediction sequence is:

[0047] The predicted wave time-frequency energy distribution characteristics and the predicted ship body attitude change data in the environment state prediction sequence are jointly processed in the time domain-frequency domain to determine the corresponding environment state prediction feature vector;

[0048] The obtained environment state prediction feature vector is used as an input feature, and the most similar stored frequency spectrum feature vector to the current input feature vector is searched in the uncoupling hook response action preparation strategy library;

[0049] The action strategy sequence corresponding to the frequency spectrum feature vector with the largest similarity is determined, and the first uncoupling action strategy is selected from the action strategy sequence as the uncoupling initial action strategy.

[0050] Further, the permanent magnet electromagnetic structure release action is achieved by eliminating the adsorption force of the permanent magnet structure through a transient current pulse; the process of determining whether the uncoupling action is blocked according to the displacement amount and the action torque specifically includes:

[0051] A motion resistance determination feature vector is constructed according to the displacement and the action torque at the moment of releasing the uncabling hook mechanism;

[0052] The motion resistance determination feature vector is input into a preset resistance classification model for classification determination to obtain a determination label, the determination label including a motion resistance label and a motion non-resistance label;

[0053] Based on the motion resistance label, a strategy correction scheme is triggered, the strategy correction scheme being:

[0054] Based on the current environment state prediction sequence and the motion resistance label, the strategy selection module S13 is repeatedly executed to match a suitable uncabling initial action strategy;

[0055] The generation step of the resistance classification model includes:

[0056] Historical resistance determination data is obtained, and the historical resistance determination data is divided into a second training set and a second test set, the historical resistance determination data including the motion resistance determination feature vector and a corresponding determination label;

[0057] An initial classifier is configured, the motion resistance determination feature vector in the second training set is taken as input data of the initial classifier, the corresponding determination label in the second training set is taken as output data of the initial classifier, the initial classifier is trained, and an initial classification network is obtained;

[0058] The initial classification network is verified through the second test set, and an initial classification network with a test accuracy greater than or equal to a preset test accuracy is output and taken as a pre-constructed resistance classification model.

[0059] Further, the logic for dynamically adjusting the release force of the permanent-magnet electromagnetic structure to keep the uncabling hook mechanism moving smoothly is:

[0060] A load deviation value of real-time cable load data and preset target load data is calculated, and a displacement deviation value between uncabling hook mechanism displacement data and preset target displacement data is calculated;

[0061] It is determined whether the absolute values of the load deviation value and the displacement deviation value exceed a preset deviation threshold value, if there is a load deviation value greater than a preset load deviation positive threshold value or a displacement deviation value greater than a preset displacement deviation positive threshold value, the release current is reduced;

[0062] If there is a load deviation value less than a preset load deviation negative threshold value or a displacement deviation value less than a preset displacement deviation negative threshold value, the release current is increased.

[0063] A ship intelligent uncabling hook control method is applied to any one of the ship intelligent uncabling hook control systems, and the method is characterized in that the method includes:

[0064] S21: Collect sea state fluctuation data and cable load data of the environment where the uncoupling hook is located; establish an environmental disturbance state sequence according to the sea state fluctuation data, and generate a dynamic load sequence according to the cable load data;

[0065] S22: Perform fast Fourier transform on the dynamic load sequence to extract the frequency spectrum features of the sequence, the frequency spectrum features including dominant frequency spectrum amplitude and spectrum energy distribution; and construct an uncoupling hook response action preparation strategy library according to the frequency spectrum features;

[0066] S23: Input the environmental disturbance state sequence into a pre-established deep recurrent neural network model to generate an environmental state prediction sequence; and select an uncoupling initial action strategy from the uncoupling hook response action preparation strategy library based on the environmental state prediction sequence;

[0067] S24: Trigger the permanent magnet electromagnetic structure in the uncoupling hook mechanism to release action according to the uncoupling initial action strategy, collect the displacement amount and action torque of the release action at the release moment, and judge whether the uncoupling action is blocked based on the displacement amount and the action torque, if the uncoupling action is blocked, feedback and correct the uncoupling initial action strategy and return to step S23;

[0068] S25: If the uncoupling action is not blocked, real-time cable load data and uncoupling hook mechanism displacement data are obtained, and the release current of the permanent magnet electromagnetic structure is dynamically adjusted to keep the uncoupling hook mechanism action smooth.

[0069] Compared with the prior art, the beneficial effects of the present application are as follows:

[0070] Through intelligent data collection, analysis and strategy decision, the present application effectively responds to complex and variable sea state environment, realizes efficient, stable and accurate control of the uncoupling hook action, accurately analyzes and predicts the sea state change trend through the combination of wavelet analysis and particle swarm optimization, makes the uncoupling action strategy highly match the actual environment, avoids blind or lag operation decision, and significantly improves the adaptability of the system to severe sea state.

[0071] In addition, through the fusion of real-time state feedback and dynamic regulation technology, fine control in the uncoupling action process is realized, the dynamic electromagnetic structure release intensity adjustment and accurate blocked state detection are used, the occurrence of uncoupling hook action abnormalities or instability is effectively avoided, the reliability and service life of the uncoupling mechanism are greatly improved, and thus the safety and operation efficiency of the ship operation are enhanced. BRIEF DESCRIPTION OF DRAWINGS

[0072] In order to more clearly illustrate the technical solutions in the embodiments or the prior art, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments described in the present application, and other drawings can also be obtained by those skilled in the art according to these drawings.

[0073] Figure 1 A module diagram of a ship intelligent unhooking hook control system provided for the embodiment 1 of the present application;

[0074] Figure 2 A flow chart of a ship intelligent unhooking hook control method provided for the embodiment 2 of the present application. DETAILED DESCRIPTION

[0075] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0076] Embodiment 1

[0077] Please refer to Figure 1 The present embodiment discloses a ship intelligent unhooking hook control system, which comprises:

[0078] The data acquisition module S11 acquires the sea state fluctuation data and the cable load data of the environment where the unhooking hook is located; the environmental disturbance state sequence is established according to the sea state fluctuation data, and the dynamic load sequence is generated according to the cable load data;

[0079] Specifically, the sea state fluctuation data comprises wave height, wave period and ship body attitude angle change; the establishment of the environmental disturbance state sequence according to the sea state fluctuation data specifically comprises:

[0080] S111: acquiring the sea state fluctuation data within a preset time length;

[0081] It should be noted that the wave height is measured in real time by using a shipborne wave height meter; the wave period is measured in real time by using a period sensor; and the ship body attitude angle change is measured in real time by using an inertial navigation system (INS), including real-time angle data of three axial directions of roll, pitch and yaw;

[0082] The wave height, the wave period and the ship body attitude angle change are all collected and stored in real time at a specified sampling frequency (for example, 5 Hz to 10 Hz);

[0083] S112: inputting the wave height and the wave period into a wavelet analysis model for time-frequency decomposition;

[0084] Specifically, the construction steps of the wavelet analysis model are:

[0085] S112.1: Select the historical wave height and wave period as the training sample, and construct an initial set of wavelet decomposition functions, which are defined by two key parameters of center frequency and bandwidth;

[0086] It should be noted that the wavelet decomposition function is a special mathematical waveform, which is similar to an oscillating wave packet, and can effectively detect the wave signal changes in different time and frequency ranges. In short, this function can disassemble complex wave signals according to different frequencies and time periods, so as to clearly know the intensity and frequency distribution of waves in different time periods.

[0087] S112.2: Use an adaptive particle swarm optimization algorithm to optimize the parameters of the set of wavelet decomposition functions;

[0088] Specifically, the adaptive particle swarm optimization algorithm is used to optimize the parameters of the set of wavelet decomposition functions, which includes:

[0089] Randomly initialize the particle swarm, and each particle of the particle swarm represents a set of wavelet decomposition function parameter combinations, including the center frequency and bandwidth of the wavelet, to form an initial particle swarm;

[0090] Use the time-frequency energy error of the training sample as the fitness function to evaluate the fitness value of each particle, and the fitness function is expressed as:

[0091]

[0092] Where β is the fitness function, E real,n is the actual time-frequency energy value of the nth historical sample point, E pred,n is the wavelet predicted time-frequency energy value of the nth sample point;

[0093] Adjust the particle speed and position according to the fitness value, and the speed update formula of the particle is:

[0094] v i (t+1)=w·v i (t)+c1·r1·[pbest i (t)-x i (t)]+c2·r2·[gbest(t)-x i (t)];

[0095] In the formula, v i (t) is the speed of the ith particle at the tth iteration, w is a preset inertia weight factor, c1 and c2 are acceleration factors of the particle swarm algorithm, and r1 and r2 are random numbers in the interval [0, 1], which are used to improve the random search ability of the algorithm, pbest i(t) is the historical optimal position found by the ith particle by the tth iteration, gbest(t) is the global optimal position found by the entire particle swarm by the tth iteration;

[0096] It should be noted that w is set by the experimenter based on historical data to balance the global search ability and local search ability of the algorithm, and the acceleration factor is a learning factor, which respectively represents the weight of the individual experience of the particle and the weight of the overall experience of the group;

[0097] The position update formula of the particle is represented as:

[0098] x i (t+1)=x i (t)+v i (t+1);

[0099] x i (t) is the position of the ith particle at the tth iteration;

[0100] Iterate until the fitness reaches a predetermined convergence threshold, and output the optimal parameters;

[0101] S112.3: Use the optimized wavelet decomposition function set to generate a wavelet analysis model;

[0102] S113: Obtain the time-frequency decomposition result of the wave signal based on the optimized wavelet analysis model, extract the time-frequency energy distribution features, including the frequency spectrum energy distribution range of the energy of the wave signal at different frequencies and time periods, the main frequency area, and the frequency change value;

[0103] S114: Construct an environmental disturbance state sequence according to the time-frequency energy distribution features and the ship body attitude angle change, wherein the ship body real-time attitude change data includes the roll change angle, the pitch change angle, and the yaw change angle, and the environmental disturbance state sequence is represented as:

[0104] S env (t)={E tf (t),θ roll (t),θ pitch (t),θ yaw (t)};

[0105] E tf (t) is the wave time-frequency energy distribution feature, θ roll (t), θ pitch (t), θ yaw (t) are the roll change angle, the pitch change angle, and the yaw change angle of the ship body, respectively;

[0106] Specifically, the steps of generating a dynamic load sequence according to the cable load data are:

[0107] Based on the cable tension sensor arranged on the disengaging hook bearing structure, the cable tension received by the disengaging hook is measured in real time and continuously;

[0108] It should be noted that the cable tension sensor includes but is not limited to strain and pressure sensors.

[0109] The sampling frequency is set to 5-20 Hz, and the load raw data of 30 seconds to 2 minutes is continuously measured and stored to obtain the initial cable load data sequence

[0110] The obtained initial load data sequence is subjected to data denoising and abnormal data point correction to obtain high-quality dynamic load data.

[0111] It should be noted that the data denoising uses existing signal filtering methods (such as Butterworth low-pass filter) to remove high-frequency noise interference to obtain a relatively smooth load signal; the abnormal data point correction uses median filtering or sliding window average filtering method to correct the abnormal data points in the load data sequence, ensuring the accuracy and stability of the load data.

[0112] The high-quality dynamic load data is subjected to normalization processing and is expressed as:

[0113]

[0114] Wherein, L(t) is the high-quality dynamic load data at the t time, L max is the maximum high-quality dynamic load data, L min is the minimum high-quality dynamic load data, L norm (t) is the high-quality normalized load value at the t time.

[0115] Based on a plurality of high-quality normalized load values, a dynamic load sequence is formed.

[0116] The above steps automatically optimize the parameters of the wavelet decomposition function through the adaptive particle swarm optimization algorithm to adapt to the analysis of complex sea state fluctuation characteristics. Compared with the traditional fixed parameter wavelet analysis or the wavelet analysis with artificially adjusted parameters, the adaptive particle swarm optimization algorithm has obvious advantages and can more accurately reflect the true characteristics of ocean fluctuations, providing a high-quality data basis for subsequent disengaging action strategies.

[0117] The spectrum analysis module S12: performs fast Fourier transform on the dynamic load sequence to extract the frequency spectrum characteristics of the sequence, including the dominant frequency spectrum amplitude and the frequency spectrum energy distribution; and constructs a disengaging hook response action preparation strategy library according to the frequency spectrum characteristics.

[0118] It should be noted that: the main frequency is the main frequency component in the load change; the spectral amplitude is the amplitude of each frequency component; the spectral energy distribution is the energy proportion of the load change in each frequency range;

[0119] Specifically, the constructing the uncoupling hook response action preparation strategy library according to the spectrum characteristics specifically includes:

[0120] S121: input the spectrum characteristics of the dynamic load sequence as an input variable into the nonlinear dynamics simulation model;

[0121] Specifically, the construction steps of the nonlinear dynamics simulation model are:

[0122] S121.1: collect the friction coefficient, elastic modulus and motion constraint conditions of the uncoupling hook mechanism, and construct the nonlinear dynamics control equation of the uncoupling hook action;

[0123] It should be noted that: the friction coefficient is obtained by standard friction test; the elastic modulus is obtained by standard material mechanics test (such as tensile test or compression test); the motion constraint condition is determined according to the uncoupling hook structure design and installation condition, including the constraint type and motion range of rotating pair and sliding pair;

[0124] S121.2: based on the actual motion characteristics of the uncoupling hook mechanism, define the generalized coordinates q i of the uncoupling hook mechanism;

[0125] It should be noted that: the uncoupling hook mechanism represents the uncoupling hook mechanical components, and the rotation angle of the uncoupling hook mechanism rotating part is taken as the generalized coordinate, or the displacement of the sliding part is taken as the generalized coordinate;

[0126] Based on the defined generalized coordinates, the kinetic energy of the uncoupling hook mechanism is calculated and obtained, which is represented as:

[0127]

[0128] In the formula, T is the kinetic energy of the uncoupling hook mechanism, J is the established moment of inertia, is the angular velocity of the rotation angle;

[0129] It should be noted that: the moment of inertia J is determined by the material density, geometric size (such as cross-sectional area, length, mass distribution) and other mechanical design parameters of the uncoupling hook mechanism;

[0130] The potential energy of the elastic part of the uncoupling hook mechanism is calculated and obtained, which is represented as:

[0131]

[0132] In the formula, V is the potential energy of the uncoupling hook mechanism, k is the established elastic stiffness coefficient, and x is the elastic variable of the elastic part.

[0133] It should be noted that the elastic component includes a spring or a beam, and k is determined based on the elastic modulus;

[0134] Based on the obtained friction coefficient, the friction torque is calculated, which is represented as:

[0135] Q i = -μNr

[0136] Wherein, Q i is the friction torque corresponding to the i-th generalized coordinate, μ is the obtained friction coefficient, N is the normal load value, and r is the effective friction force action radius;

[0137] It should be noted that when measuring the friction coefficient or simulating the actual working condition, the loading device is used to apply a normal force to the contact surface of the cable hook; a force sensor (such as a pressure sensor or a strain gauge type force meter) is installed to record the size of the applied normal force; the data output by the sensor is recorded and converted into a clear normal load value through a data acquisition system;

[0138] r refers to the actual effective action distance of the friction force on the contact surface, which is usually determined according to the geometric size of the cable hook mechanism and the actual distribution of the contact surface;

[0139] The dynamics equation of the cable hook mechanism is constructed:

[0140]

[0141] Wherein, L is the Lagrange function, L = T-V; is the partial derivative of the Lagrange function L with respect to the generalized coordinate q i is the partial derivative, is the partial derivative of the Lagrange function L with respect to the generalized velocity of the change rate with respect to time, is the derivative of the i-th generalized coordinate with respect to time;

[0142] It should be noted that the generalized coordinate q i is used to describe the position, angle and other degrees of freedom of the cable hook mechanism, represents the motion speed of the generalized coordinate;

[0143] S121.3: Solve the control equation by finite element numerical method to construct a nonlinear dynamics simulation model;

[0144] Specifically, the steps of constructing the nonlinear dynamics simulation model are:

[0145] establishing a finite element geometric model of the uncabling hook mechanism; setting the obtained friction coefficient, elastic modulus and motion constraint conditions as parameters and boundary conditions of the finite element simulation; using a finite element simulation software to numerically solve the established nonlinear control equation to obtain detailed dynamic response data of the uncabling hook mechanism under the frequency spectrum characteristic input condition, so as to construct a complete nonlinear dynamic simulation model;

[0146] It should be noted that the finite element simulation software includes but is not limited to ABAQUS, ANSYS;

[0147] S122: Simulate the mechanical response characteristics of the uncabling hook action under the corresponding frequency spectrum characteristics based on the nonlinear dynamic simulation model;

[0148] The dominant frequency, the frequency spectrum amplitude and the frequency spectrum energy distribution in the frequency spectrum characteristics are taken as the excitation input of the nonlinear dynamic simulation model. The dominant frequency is used to construct a simulation load input signal, which is represented as:

[0149] F(t)=A f sin(2πf dom t)

[0150] Wherein, F(t) is the simulation load input signal, t is the time, f dom is the dominant frequency, A f is the frequency spectrum amplitude corresponding to the dominant frequency f dom ;

[0151] Based on the frequency spectrum energy distribution, the secondary frequency adjacent to the dominant frequency is constructed as an auxiliary excitation input signal as an external excitation;

[0152] The simulation duration is set to 30s-120s, the time step is set to 0.01s-0.1s, and the nonlinear dynamic finite element simulation is run to obtain the motion trajectory data of the uncabling hook mechanism response;

[0153] The mechanical response characteristics of the uncabling hook mechanism response in the simulation results are extracted, including the initial time of the uncoupling threshold, the uncoupling action duration and the peak action torque of the uncabling hook mechanism;

[0154] It should be noted that the initial time when the uncabling hook mechanism reaches the preset uncoupling threshold is determined according to the displacement or angle data; the duration of the uncoupling action is determined according to the time length required for the uncabling hook mechanism action from the starting action time to complete uncoupling in the simulation results; the peak action torque of the uncabling hook mechanism is determined according to the peak value of the action torque experienced by the uncabling hook mechanism during the uncoupling process;

[0155] The beneficial effects of steps S121 and S122 are:

[0156] By collecting the friction coefficient, elastic modulus and motion constraint conditions of the uncoupling hook mechanism, a nonlinear dynamics simulation model is constructed, and the specific motion response process of the uncoupling hook mechanism under the action of complex dynamic load is accurately simulated by taking the spectral characteristics of the dynamic load sequence as input, so that the initial triggering time of the uncoupling threshold, the duration of the uncoupling action and the peak torque borne by the mechanism during the uncoupling process are obtained, thereby avoiding the problem that the traditional empirical method cannot accurately describe the nonlinear friction, elasticity and constraint conditions, and providing data support for the subsequent development of uncoupling action strategy, and improving the safety and general applicability of the uncoupling hook control.

[0157] S123: generating an uncoupling response action strategy sequence according to the mechanical response characteristics;

[0158] Based on the initial time of the uncoupling threshold, the duration of the uncoupling action and the peak torque of the uncoupling hook mechanism, an uncoupling response action strategy is constructed, which is represented as:

[0159] P k ={t start,k ,t dur,k ,M act,k}

[0160] P k is the kth uncoupling response action strategy, t start,k is the starting time point t dur,k is the action duration, and M act,k is the peak torque of the uncoupling hook mechanism;

[0161] For each group of simulation analysis data corresponding to the spectral characteristics, repeat the above steps to generate a complete uncoupling response action strategy sequence;

[0162] S124: forming a preliminary strategy library by forming a mapping relationship between the spectral characteristics and the action strategy sequence;

[0163] The strategy library data structure is constructed, the spectral characteristics are taken as the mapping input, and the corresponding uncoupling response action strategy is taken as the mapping output, a one-to-one mapping relationship between the spectral characteristics and the uncoupling response action strategy is established, and the mapping input spectral characteristics are represented as:

[0164] f={f dom ,A f ,E dist}

[0165] Wherein, f is the mapping input spectral characteristics, f dom is the dominant frequency, A f is the spectral amplitude corresponding to the dominant frequency, and E dist is the spectral energy distribution characteristic;

[0166] The uncoupling response action strategy of the mapping output is represented as:

[0167] P k = {t start,k , t dur,k , M act,k}

[0168] Based on the multi-dimensional interpolation algorithm, the spectrum features formed by all simulation data and the corresponding action strategy are taken as training data to fit and establish a nonlinear mapping relationship from the spectrum features to the unhooking response action strategy;

[0169] It should be noted that the multi-dimensional interpolation method includes but is not limited to radial basis function interpolation (RBF interpolation) and cubic spline interpolation;

[0170] Based on the obtained nonlinear mapping relationship stored in the computer database, a complete unhooking response action preparation strategy library is constructed;

[0171] The beneficial effects of steps S123 and S124 are:

[0172] By establishing the mapping relationship between the load spectrum features and the unhooking action strategy, a pre-optimized strategy library is formed, so that in actual work, the dynamic load working condition can be quickly matched and the corresponding unhooking strategy can be provided in time, avoiding the response delay and error caused by real-time calculation; at the same time, the multi-dimensional interpolation algorithm is used to fit the nonlinear relationship between the spectrum features and the action strategy, which improves the adaptability of the strategy library to unknown load working conditions, effectively expanding the application range and control accuracy of the system under complex dynamic conditions;

[0173] Strategy selection module S13: inputting the environmental disturbance state sequence into the pre-established deep recurrent neural network model to generate an environmental state prediction sequence; selecting an unhooking initial action strategy from the unhooking response action preparation strategy library based on the environmental state prediction sequence;

[0174] Specifically, the generation steps of the deep recurrent neural network model include:

[0175] Obtain historical environmental prediction data, and divide the historical environmental prediction data into a first test set and a first training set, wherein the historical environmental prediction data includes an environmental disturbance state sequence and a corresponding environmental state prediction sequence;

[0176] It should be noted that the environmental state prediction sequence is the wave time-frequency energy distribution feature and the ship attitude change data in the future time period;

[0177] Construct a regression network, take the environmental disturbance state sequence in the first training set as the input of the regression network, take the corresponding environmental state prediction sequence as the output of the regression network, train the regression network, and obtain an initial deep recurrent neural network;

[0178] The initial deep recurrent neural network is verified by using the first test set, and an initial deep recurrent neural network with an output less than or equal to a preset test error threshold is taken as the deep recurrent neural network model.

[0179] Specifically, the logic for selecting the unhooking initial action strategy from the unhooking response action preparation strategy library based on the environment state prediction sequence is as follows:

[0180] The predicted wave time-frequency energy distribution feature and the predicted ship body posture change data in the environment state prediction sequence are jointly processed in the time domain and the frequency domain to determine the corresponding environment state prediction feature vector;

[0181] The obtained environment state prediction feature vector is used to search for a stored frequency spectrum feature vector most similar to the current input feature vector in the unhooking response action preparation strategy library with the environment state prediction feature vector as the input feature;

[0182] It should be noted that the similarity between the feature vectors is realized by a distance measurement algorithm, including but not limited to the Euclidean distance, the cosine similarity, or the Mahalanobis distance method.

[0183] The action strategy sequence corresponding to the frequency spectrum feature vector with the maximum similarity is determined, and the first unhooking action strategy in the action strategy sequence is selected as the unhooking initial action strategy.

[0184] The blocked determination module S14 triggers the permanent magnet electromagnetic structure in the unhooking hook mechanism to release the action according to the unhooking initial action strategy, collects the displacement amount and the action torque at the release moment, and determines whether the unhooking action is blocked based on the displacement amount and the action torque;

[0185] It should be noted that the displacement amount data is collected by a displacement sensor installed on the moving part of the unhooking hook mechanism, and the specific sensor type includes but is not limited to a linear variable differential transformer (LVDT) or a grating displacement sensor;

[0186] The action torque data is obtained by real-time measurement of a torque sensor arranged on the rotating shaft or bearing part of the unhooking hook mechanism, and the torque sensor type includes but is not limited to a strain torque sensor or a piezoelectric torque sensor;

[0187] Specifically, the permanent magnet electromagnetic structure release action is realized by eliminating the adsorption force of the permanent magnet structure through a transient current pulse; and the process of determining whether the unhooking action is blocked based on the displacement amount and the action torque specifically includes:

[0188] S141: constructing an action blocking determination feature vector according to the displacement amount and the action torque at the release moment of the unhooking hook mechanism;

[0189] which is expressed as:

[0190] X b = [Δd, M act ] T

[0191] In the formula, X b is an action blocking judgment feature vector, Δd is a displacement change value at the release moment of the cable hook release mechanism, M act is an action moment value at the release moment of the cable hook release mechanism, and the superscript T represents the transpose of the vector;

[0192] It should be noted that Δd represents the displacement increment from the action start moment of the cable hook release mechanism to the current measurement moment, M act represents the real-time measured cable hook release mechanism axial moment value at the release moment of the unhooking action, and the superscript T represents the transpose of the vector, that is, the conversion of the row vector into a column vector;

[0193] S142: input the action blocking judgment feature vector into a preset blocked classification model for classification judgment to obtain a judgment label, the judgment label including an action blocked mark and an action unblocked mark;

[0194] Specifically, the generation step of the blocked classification model includes:

[0195] Obtain historical blocked judgment data, divide the historical blocked judgment data into a second training set and a second test set, and the historical blocked judgment data includes an action blocking judgment feature vector and a corresponding judgment label;

[0196] Configure an initial classifier, use the action blocking judgment feature vectors in the second training set as the input data of the initial classifier, use the corresponding judgment labels in the second training set as the output data of the initial classifier, train the initial classifier, and obtain an initial classification network;

[0197] Verify the initial classification network through the second test set, output the initial classification network with a test accuracy greater than or equal to a preset test accuracy, and use it as a pre-constructed blocked classification model;

[0198] It should be noted that the blocked classification model can use a support vector machine;

[0199] S143: based on the action blocked mark, trigger a strategy correction scheme, and the strategy correction scheme is:

[0200] Based on the current environment state prediction sequence and the action blocked mark, repeatedly execute the strategy selection module S13 to match a suitable unhooking initial action strategy.

[0201] The stable control module S15: if the unhooking action is not blocked, obtain real-time cable load data and cable hook release mechanism displacement data, and dynamically adjust the release current of the permanent magnet electromagnetic structure to keep the cable hook release mechanism action smooth;

[0202] Specifically, the logic for dynamically adjusting the release force of the permanent-magnetic electromagnetic structure to keep the decoupling hook mechanism moving smoothly is as follows:

[0203] The load deviation value of the real-time cable load data and the preset target load data is calculated, and the displacement deviation value between the displacement data of the decoupling hook mechanism and the preset target displacement data is calculated.

[0204] is expressed as:

[0205] ΔL = L real (τ) - L ref

[0206] ΔS = S real (τ) - S ref

[0207] In the formula, ΔL is the load deviation value, L real (τ) is the real-time cable load data at time τ, L ref is the preset target load data, ΔS is the displacement deviation value, S real (τ) is the real-time measured displacement data of the decoupling hook mechanism, S ref is the preset target displacement data.

[0208] It is judged whether the absolute values of the load deviation value and the displacement deviation value exceed the preset deviation threshold value. If there is a load deviation value greater than the preset load deviation positive threshold value, or a displacement deviation value greater than the preset displacement deviation positive threshold value, the release current is reduced.

[0209] If there is a load deviation value less than the preset load deviation negative threshold value, or a displacement deviation value less than the preset displacement deviation negative threshold value, the release current is increased.

[0210] In the execution steps of the strategy selection module S13 and the blocked judgment module S14, the real-time collected environmental disturbance state sequence is predicted by using a deep recurrent neural network, so as to realize accurate estimation of the future environmental change trend; the prediction result is quickly matched with the pre-constructed action strategy library, so as to effectively improve the foresight and accuracy of the decoupling hook action decision. In addition, based on the blocked judgment model, it is monitored in real time whether the release action of the decoupling hook mechanism is blocked, so that the system can quickly identify the abnormality and timely trigger the correction process of the action strategy, thereby significantly reducing the response delay under abnormal state; at the same time, under the unblocked condition, the release current is dynamically adjusted, the load and displacement deviations are corrected in real time, the fine control of the decoupling hook mechanism is realized, and the stability, reliability and smoothness of the mechanism operation are ensured.

[0211] Example 2

[0212] Please refer to Figure 2As shown, based on the unified inventive concept, the embodiment discloses a ship intelligent unhooking hook control method, comprising:

[0213] S21: collect the sea state fluctuation data and the cable load data of the unhooking hook environment; establish an environmental disturbance state sequence according to the sea state fluctuation data, and generate a dynamic load sequence according to the cable load data;

[0214] S22: perform fast Fourier transform on the dynamic load sequence to extract the frequency spectrum characteristics of the sequence, the frequency spectrum characteristics including the dominant frequency spectrum amplitude and the spectrum energy distribution; and construct an unhooking hook response action preparation strategy library according to the frequency spectrum characteristics;

[0215] S23: input the environmental disturbance state sequence into the pre-established deep recurrent neural network model to generate an environmental state prediction sequence; and select an unhooking initial action strategy from the unhooking hook response action preparation strategy library based on the environmental state prediction sequence;

[0216] S24: trigger the permanent magnet electromagnetic structure in the unhooking hook mechanism to release the action according to the unhooking initial action strategy, collect the displacement and the action torque at the release moment of the release action, and judge whether the unhooking action is blocked based on the displacement and the action torque, if the unhooking action is blocked, feedback corrects the unhooking initial action strategy and returns to step S23;

[0217] S25: if the unhooking action is not blocked, obtain real-time cable load data and unhooking hook mechanism displacement data, and dynamically adjust the release current of the permanent magnet electromagnetic structure to keep the unhooking hook mechanism action smooth.

[0218] The above-described embodiments only describe the preferred embodiments of the present application, and do not limit the scope of the present application. Without departing from the design spirit of the present application, various modifications and improvements to the technical solutions of the present application made by those skilled in the art shall fall within the protection scope of the present application determined by the claims.

Claims

1. A marine intelligent disengaging hook control system, characterized in that, The system comprises: A data acquisition module S11: collect the sea state fluctuation data and cable load data of the environment where the uncoupling hook is located; establish an environmental disturbance state sequence according to the sea state fluctuation data, and generate a dynamic load sequence according to the cable load data; A spectrum analysis module S12: perform fast Fourier transform on the dynamic load sequence to extract the spectral features of the sequence, wherein the spectral features include dominant frequency spectral amplitude and spectral energy distribution; and construct an uncoupling hook response action preparation strategy library according to the spectral features; A strategy selection module S13: input the environmental disturbance state sequence into a pre-established deep recurrent neural network model to generate an environmental state prediction sequence; and select an uncoupling initial action strategy from the uncoupling hook response action preparation strategy library based on the environmental state prediction sequence; A blocked judgment module S14: trigger the permanent magnet electromagnetic structure release action in the uncoupling hook mechanism according to the uncoupling initial action strategy, collect the displacement and action torque of the release action at the release moment, and judge whether the uncoupling action is blocked based on the displacement and action torque; A stable control module S15: if the uncoupling action is not blocked, obtain real-time cable load data and uncoupling hook mechanism displacement data, and dynamically adjust the release current of the permanent magnet electromagnetic structure to keep the uncoupling hook mechanism action smooth; Wherein the construction of the uncoupling hook response action preparation strategy library according to the spectral features specifically comprises: Input the spectral features of the dynamic load sequence as input variables into a nonlinear dynamics simulation model; Obtain the mechanical response features of the uncoupling hook action under the corresponding spectral features based on the nonlinear dynamics simulation model; Generate an uncoupling response action strategy sequence according to the mechanical response features; Form a mapping relationship between the spectral features and the action strategy sequence to form the preparation strategy library; The construction steps of the nonlinear dynamics simulation model are: Collect the friction coefficient, elastic modulus and motion constraint conditions of the uncoupling hook mechanism, and construct the nonlinear dynamics control equation of the uncoupling hook action; Define the generalized coordinates of the uncoupling hook mechanism based on the actual motion characteristics of the uncoupling hook mechanism; Based on the defined generalized coordinates, calculate the kinetic energy of the uncoupling hook mechanism, and calculate the potential energy of the elastic components of the uncoupling hook mechanism; based on the obtained friction coefficient, calculate the friction torque, and construct the dynamics equation of the uncoupling hook mechanism according to the kinetic energy of the uncoupling hook mechanism, the potential energy of the elastic components of the uncoupling hook mechanism and the friction torque; Solve the control equation by the finite element numerical method to construct the nonlinear dynamics simulation model; The steps for constructing the nonlinear dynamics simulation model are: Establish a finite element geometric model of the uncoupling hook mechanism; set the obtained friction coefficient, elastic modulus and motion constraint conditions as the parameters and boundary conditions of the finite element simulation; use finite element simulation software to numerically solve the established nonlinear control equation to obtain detailed dynamics response data of the uncoupling hook mechanism under the input condition of the spectral features, so as to construct a complete nonlinear dynamics simulation model.

2. A smart disengaging hook control system for a marine vessel as claimed in claim 1, wherein, The sea state fluctuation data includes wave height, wave period and ship attitude angle change; and the establishment of the environmental disturbance state sequence according to the sea state fluctuation data specifically comprises: Collect the sea state fluctuation data within a preset time period; The wave height and wave period are input into a wavelet analysis model for time-frequency decomposition; The wavelet analysis model is constructed by the following steps: Historical wave height and wave period are selected as training samples to construct an initial set of wavelet decomposition functions, which are defined by two key parameters, center frequency and bandwidth; An adaptive particle swarm optimization algorithm is used to optimize the parameters of the set of wavelet decomposition functions; The optimized set of wavelet decomposition functions is used to generate a wavelet analysis model; Based on the optimized wavelet analysis model, the time-frequency decomposition results of the wave signal are obtained, and the time-frequency energy distribution features are extracted, including the spectral energy distribution range of the wave signal energy at different frequencies and time periods, the main frequency region, and the frequency change value; The environmental disturbance state sequence is constructed based on the time-frequency energy distribution features and the ship body attitude angle changes, including the roll change angle, the pitch change angle, and the yaw change angle.

3. A smart disengaging hook control system for a marine vessel as claimed in claim 2, wherein, The steps of optimizing the parameters of the set of wavelet decomposition functions using the adaptive particle swarm optimization algorithm include: Randomly initialize the particle swarm, and each particle in the particle swarm represents a combination of wavelet decomposition function parameters, including the center frequency and bandwidth of the wavelet, to form an initial particle swarm; Use the time-frequency energy error of the training samples as the fitness function to evaluate the fitness value of each particle; Adjust the particle speed and position based on the fitness value, and iterate until the fitness value reaches a predetermined convergence threshold, and output the optimal parameters.

4. A smart disengaging hook control system for a marine vessel as claimed in claim 3, wherein, The steps of generating a dynamic load sequence based on cable load data include: Based on the cable tension sensor installed on the disengaging hook bearing structure, the cable tension on the disengaging hook is measured in real time and continuously; Set the sampling frequency to 5-20 Hz, continuously measure and store 30 seconds to 2 minutes of raw load data to obtain an initial cable load data sequence; Perform data denoising and abnormal data point correction on the obtained initial load data sequence to obtain high-quality dynamic load data; Normalize the high-quality dynamic load data to generate high-quality normalized load values; Based on multiple sets of high-quality normalized load values, a dynamic load sequence is formed.

5. The intelligent off- cleat hook control system for marine vessels as claimed in claim 1 wherein, The steps of generating a deep recurrent neural network model include: Obtain historical environmental prediction data, and divide the historical environmental prediction data into a first test set and a first training set, including environmental disturbance state sequences and corresponding environmental state prediction sequences; Construct a regression network, use the environmental disturbance state sequences in the first training set as the input of the regression network, use the corresponding environmental state prediction sequences as the output of the regression network, train the regression network, and obtain an initial deep recurrent neural network; Use the first test set to verify the initial deep recurrent neural network model, and output the initial deep recurrent neural network with a test error threshold less than or equal to the preset test error threshold as the deep recurrent neural network model; The logic for selecting a disengaging initial action strategy from the disengaging hook response action preparation strategy library based on the environmental state prediction sequence is: The predicted wave time-frequency energy distribution features in the environment state prediction sequence and the predicted ship body posture change data are jointly processed in time domain and frequency domain to determine corresponding environment state prediction feature vectors; Using the obtained environment state prediction feature vectors, an input feature is taken as an environment state prediction feature vector, and a stored frequency spectrum feature vector most similar to the current input feature vector is searched in a disengaging hook response action preparation strategy library; The action strategy sequence corresponding to the frequency spectrum feature vector with the largest similarity is determined, and a first disengaging action strategy is selected from the action strategy sequence as a disengaging initial action strategy.

6. The intelligent off- cleat hook control system for marine vessels as claimed in claim 1 wherein, The permanent magnet electromagnetic structure release action is achieved by eliminating the adsorption force of the permanent magnet structure through a transient current pulse; and the process of determining whether the disengaging action is blocked according to the displacement and the action torque specifically includes: An action blockage judgment feature vector is constructed according to the displacement and the action torque at the moment of the disengaging hook mechanism release; The action blockage judgment feature vector is input into a preset blocked classification model for classification judgment to obtain a judgment label, the judgment label including an action blocked label and an action unblocked label; Based on the action blocked label, a strategy correction scheme is triggered, and the strategy correction scheme is: Based on the current environment state prediction sequence and the action blocked label, the strategy selection module S13 is repeatedly executed to match a suitable disengaging initial action strategy; The generation steps of the blocked classification model include: Historical blocked judgment data is obtained, and the historical blocked judgment data is divided into a second training set and a second test set, the historical blocked judgment data including an action blockage judgment feature vector and a corresponding judgment label; An initial classifier is configured, the action blockage judgment feature vector in the second training set is taken as input data of the initial classifier, the corresponding judgment label in the second training set is taken as output data of the initial classifier, the initial classifier is trained, and an initial classification network is obtained; The initial classification network is verified through the second test set, and an initial classification network with a test accuracy greater than or equal to a preset test accuracy is output and taken as a pre-constructed blocked classification model.

7. The intelligent off- cleat hook control system for marine vessels as claimed in claim 1 wherein, The logic for dynamically adjusting the release intensity of the permanent magnet electromagnetic structure to keep the disengaging hook mechanism action smooth is: A load deviation value of real-time cable load data and a preset target load data is calculated, and a displacement deviation value between disengaging hook mechanism displacement data and a preset target displacement data is calculated; It is judged whether the absolute values of the load deviation value and the displacement deviation value exceed a preset deviation threshold value, if there is a load deviation value greater than a preset load deviation positive threshold value, or a displacement deviation value greater than a preset displacement deviation positive threshold value, the release current is reduced; If there is a load deviation value less than a preset load deviation negative threshold value, or a displacement deviation value less than a preset displacement deviation negative threshold value, the release current is increased.

8. A method for controlling an intelligent disengaging hook for a ship, applied to the intelligent disengaging hook control system of any one of claims 1-7, characterized in that, The method includes: S21: collecting sea state fluctuation data and cable load data of an environment where a disengaging hook is located; establishing an environment disturbance state sequence according to the sea state fluctuation data, and generating a dynamic load sequence according to the cable load data; S22: Fast Fourier transform is performed on the dynamic load sequence to extract the spectral features of the sequence, including dominant frequency spectral amplitude and spectral energy distribution; a decoupling hook response action preparation strategy library is constructed according to the spectral features; S23: The environmental disturbance state sequence is input into a pre-established deep recurrent neural network model to generate an environmental state prediction sequence; and a decoupling initial action strategy is selected from the decoupling hook response action preparation strategy library based on the environmental state prediction sequence; S24: The decoupling initial action strategy is used to trigger the release action of the permanent magnet electromagnetic structure in the decoupling hook mechanism; the displacement and action torque at the release moment of the release action are collected, and whether the decoupling action is blocked is judged based on the displacement and action torque; if the decoupling action is blocked, the decoupling initial action strategy is corrected and step S23 is returned; S25: If the decoupling action is not blocked, real-time cable load data and decoupling hook mechanism displacement data are obtained, and the release current of the permanent magnet electromagnetic structure is dynamically adjusted to keep the decoupling hook mechanism action smooth.

Images