Intelligent control system of automatic sling special for ecological fish reef installation

Abstract

The application relates to the technical field of ocean engineering equipment control, and discloses an intelligent control system of an automatic sling special for ecological fish reef installation, which comprises a tension sensor, an inertial measurement unit, a ranging sonar, a hydraulic execution assembly and a control logic module; the control logic module performs variational mode decomposition on the tension data to extract an effective payload background tension component; Kalman filtering is performed on the acceleration and ranging data to obtain a relative seabed vertical displacement amount; a phase plane state space and a stable support geometric domain are constructed based on the background tension component and the vertical displacement amount, the stable support geometric domain boundary is adjusted by using wave disturbance energy variance; a feedforward instruction is issued by calculating the evolution slope of the phase trajectory; and a release instruction is sent when a state point falls into the stable support geometric domain and the residence time meets the standard. The application excludes the influence of wave disturbance, realizes adaptive adjustment of the release condition, prevents fish reef collision and damage, and guarantees the safety of hoisting operation under complex sea conditions.

Description

Technical Field

[0001] This invention relates to the field of marine engineering equipment control technology, specifically to an intelligent control system for a special automatic lifting device for installing ecological artificial reefs. Background Technology

[0002] Artificial reefs play a crucial role in nearshore ecological restoration, and their installation typically relies on crane vessels and automated spreader systems (AS / RS) to deploy them to designated seabed locations. Existing AS / RS often use a single tension threshold or depth arrival condition as the criterion for disengagement. In actual marine operations, the combined effects of wave motion and the elastic deformation of the lifting cables create disturbances, causing interference signals in the data collected by tension sensors. Conventional control logic struggles to separate the effective load support status from wave disturbances, easily leading to premature disengagement of the spreader before the reef fully contacts the seabed. Furthermore, the lack of multi-dimensional state observation and velocity feedforward control during the deployment process means that improper deployment speed control or soft seabed conditions can prevent timely and smooth intervention by the spreader, resulting in destructive collisions between the reef and the seabed. In addition, distortion of ranging signals in complex underwater acoustic environments and transient stress changes caused by uneven seabed can also lead to misjudgments in the control system, reducing the reliability of marine equipment operations. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides an intelligent control system for a special automatic lifting tool for ecological artificial reef installation. This system solves the problem that existing automatic lifting tools are unable to accurately determine the actual support status of the artificial reef in wave disturbance and complex seabed environments, leading to misjudgment of the timing of unhooking and damage from collisions during lowering.

[0004] To achieve the above objectives, the present invention provides the following technical solution: An intelligent control system for an automatic lifting device for installing ecological artificial reefs includes a tension sensor, an inertial measurement unit, a ranging sonar, a hydraulic proportional relief valve, a release electromagnetic hydraulic lock, and a control logic module. The control logic module includes: The signal decoupling module performs variational mode decomposition on the transient tension data collected by the tension sensor and outputs the effective load background tension component and wave and cable elastic disturbance components. The pose observation module performs Kalman filtering on the vertical acceleration data collected by the inertial measurement unit and the vertical distance data collected by the ranging sonar, and outputs the vertical displacement relative to the seabed. The topology mapping module obtains the wave excitation energy variance from the wave and cable elastic disturbance components, constructs a two-dimensional phase plane state space and a stable support geometric domain from the relative seabed vertical displacement and the effective load background tension components, and adjusts the boundary of the stable support geometric domain using the wave excitation energy variance. The feedforward control module performs tracking differentiation on the background tension component of the effective load and the vertical displacement relative to the seabed to obtain the phase trajectory evolution slope, and sends a feedforward voltage command to the hydraulic proportional relief valve. When the state point formed by the vertical displacement relative to the seabed and the base tension component of the effective load in the two-dimensional phase plane state space falls into the stable support geometric domain and the integral of the dwell time meets the system preset conditions, the unhooking execution module sends a high-level release command to the unhooking electromagnetic hydraulic latch.

[0005] The signal decoupling module executes a timestamp synchronization mechanism to establish clock synchronization between the tension sensor, the inertial measurement unit, and the ranging sonar; The signal decoupling module acquires the transient tension data according to a preset acquisition frequency. When the transient tension data is detected to be missing, transient interpolation processing is performed based on the valid data corresponding to adjacent timestamps. Finally, peak threshold points that exceed the preset physical limit range are removed; The preset physical limit range is determined based on the maximum breaking force of the lifting device when fully loaded and the upper limit of the sensor's range.

[0006] The variational mode decomposition includes: Establish a constrained variational model that includes the superposition of multiple eigenmode function components with discrete center frequencies; Introducing a quadratic penalty factor and Lagrange multipliers into the constrained variational model transforms it into an unconstrained variational problem; The intrinsic mode function components and the discrete center frequencies are iteratively updated in the frequency domain by the alternating direction multiplier method. In the denominator of the underlying operation of the iterative update, a regularization term with a small positive value is superimposed. After the iterative update converges, the continuous low-frequency dominant mode component with a center frequency less than the frequency division threshold and a proportion greater than the preset weight in the total signal energy is extracted. The continuous low-frequency dominant mode component is determined as the effective load background tension component. Multiple high-frequency modal components with a center frequency greater than the frequency division threshold are superimposed and reconstructed to generate the wave and cable elastic disturbance components. The frequency division threshold is adaptively assigned based on the peak frequency of the average wave spectrum of the target sea area, and the preset weight is set based on the empirical proportion of tension energy of the effective load in the static state of seawater to the total signal energy.

[0007] The pose observation module constructs a linear state prediction equation and generates a priori state estimation vector based on the vertical acceleration data. The vertical distance data is used as the observation input vector to construct the observation equation, and the state estimate is updated by combining the observation equation with the prior state estimation vector. Acquire the acoustic echo signal-to-noise ratio data collected by the ranging sonar. When the acoustic echo signal-to-noise ratio data is lower than the safety threshold, amplify the element values ​​of the measurement noise covariance matrix in the Kalman filter by a nonlinear scaling factor. Anti-divergence constraint terms are superimposed on the main diagonal of the gain matrix in the Kalman filter matrix inversion operation stage. The safety threshold is determined based on the static ranging error variance benchmark calibrated at the factory of the sonar equipment.

[0008] The pose observation module includes a mutation outlier isolator based on kinematic envelope; Calculate the theoretical maximum displacement change envelope interval at the current moment based on the prior state estimation vector from the previous operation cycle; When the step change between the vertical distance data and the vertical distance data of the previous calculation cycle exceeds the theoretical maximum displacement change envelope, the acoustic measurement data of the current calculation cycle is isolated, and the relative seabed vertical displacement is obtained by pure inertial deduction relying solely on the linear state prediction equation within the current calculation cycle.

[0009] The topology mapping module performs energy variance integration within a time window for the wave and cable elastic disturbance components to obtain the wave disturbance energy variance. The wave disturbance energy variance is input into the mapping function, and combined with the nominal displacement trigger threshold and the nominal tension trigger threshold, the dynamic displacement threshold and the dynamic tension threshold are calculated and obtained. Next, the dynamic displacement threshold and the dynamic tension threshold are clamped and limited above the lower bound of the bottom safety limit; The greater the variance of the wave disturbance energy, the greater the contraction of the dynamic displacement threshold and the dynamic tension threshold towards the origin of the coordinate system. The shape of the stable support geometry is an elliptical envelope surface whose center coincides with the origin of the state space coordinate of the two-dimensional phase plane and whose major and minor semi-axes are respectively a quarter-elliptical envelope surface of the dynamic displacement threshold and the dynamic tension threshold after being clamped and restricted. The nominal displacement trigger threshold is determined based on the sonar measurement blind zone and the geometric height of the lifting device chassis, and the nominal tension trigger threshold is determined based on the static buoyancy of the lifting device in still water.

[0010] The feedforward control module performs a timing alignment operation based on a hardware timestamp cache queue for the effective load background tension component and the relative seabed vertical displacement. The aligned data is input into a discrete nonlinear tracking differentiator to obtain the smoothed derivative signal of the effective load's base tension component and the smoothed derivative signal of the relative seabed vertical displacement. The smoothed derivative signal of the effective load's base tension component and the smoothed derivative signal of the relative seabed vertical displacement are used to calculate the quotient, thereby obtaining the phase trajectory evolution slope. When the absolute value of the smoothed derivative signal of the vertical displacement relative to the seabed is less than the small motion threshold, the quotient calculation is stopped and the slope of the phase trajectory evolution is cleared to zero. The micro-motion threshold characterizes the velocity boundary of the lifting device when it is in a transient static hovering state in the vertical direction. The absolute deviation between the phase trajectory evolution slope and the hard landing threshold is calculated. The effective over-limit slope deviation is obtained by subtracting the adaptive dead zone bandwidth. The adaptive dead zone bandwidth and the wave disturbance energy variance form a positive proportional function mapping relationship. When the effective over-limit slope deviation is greater than zero, a smooth saturation function containing a hyperbolic tangent function with a scaling factor is used to nonlinearly map the effective over-limit slope deviation to generate the feedforward voltage command. The upper limit of the numerical output of the feedforward voltage command is clamped and limited to the maximum control voltage limit value allowed by the hydraulic proportional relief valve. The hard landing threshold is calculated based on the elastic modulus of the cable of the mother ship's heave compensation winch and the ultimate compressive yield strength of the ecological reef frame.

[0011] The decoupling execution module defines a discrete state characteristic function based on whether the state point falls within the stable support geometry. The discrete state characteristic function values ​​within the preset observation time window are summed over time, and the dwell reliability index is obtained by dividing the summation result by the actual number of effective sampling points within the window. When the dwell confidence index is not less than the trigger confidence threshold, it is determined that the dwell time integral meets the system preset condition and the high-level release command is generated. After the high-level release command is issued, the physical separation secondary verification logic is activated, the separation monitoring window is started to read the transient tension data, and when the transient tension data does not drop to the minimum value range of no load and the vertical displacement relative to the seabed shows a reverse pulling trend, an audible and visual alarm is triggered and a constant tension follow-up command is issued. The trigger confidence threshold is set according to the probability constraint limit of the unplanned accident rate of unhooking in the deep-sea construction operation procedure; the minimum value range of no-load is determined according to the static wet weight of the lifting gear when it is unloaded in seawater.

[0012] This invention provides an intelligent control system for a special automatic lifting tool for installing ecological artificial reefs. It has the following beneficial effects: 1. This invention obtains the effective load background tension component by performing variational mode decomposition on tension sensor data through a signal decoupling module, and uses Kalman filtering to fuse inertial measurement unit and ranging sonar data, combined with abrupt outlier isolation mechanism to obtain the vertical displacement relative to the seabed; it eliminates the influence of high-frequency wave interference and cable sway under complex sea conditions, maintains the continuity and numerical reliability of the attitude calculation process, and provides an accurate data basis for subsequent unhooking determination.

[0013] 2. This invention constructs a phase plane state space using the vertical displacement relative to the seabed and the background tension component, utilizes the dynamic contraction of the wave excitation energy variance to stabilize the geometric domain boundary, and combines the tracking differential results of two sets of parameters to generate a feedforward voltage command to intervene in the hydraulic system; it achieves adaptive adjustment of the unhooking trigger condition, and while smoothly controlling the lowering speed of the lifting gear to prevent collision damage between the artificial reef and the seabed, it avoids premature unhooking caused by changes in sea state.

[0014] 3. During the unhooking execution phase, this invention calculates the dwell reliability index through discrete state indicator functions to determine the effective dwell time of the state point within the stable support geometric domain. After the release command is issued, it activates a secondary physical verification logic that includes transient tension and reverse displacement trend judgment. This can effectively eliminate erroneous unhooking signals caused by local seabed unevenness or bottom sediment subsidence, ensuring the equipment safety of the entire process of ecological reef installation. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the system architecture of the present invention; Figure 2 This is a flowchart of the method of the present invention; Figure 3 This is a comparison diagram of the transient tension evolution of the sling during the landing process of this invention; Figure 4 This is a comparative chart of the statistical evaluation of the multi-dimensional operation indicators of the system of the present invention. Detailed Implementation

[0016] The technical solutions in 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, and 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.

[0017] Please see the appendix Figure 1 This invention provides an intelligent control system for a special automatic lifting tool for installing ecological artificial reefs, including physical hardware devices and a control logic module. The physical hardware devices mainly consist of a sensing layer, a transmission layer, a main control layer, and an execution layer.

[0018] The sensing layer, mounted on the automated spreader body, acquires multi-source physical quantities during the deployment of the artificial reef. This layer includes tension sensors, an inertial measurement unit (IMU), and a ranging sonar. Tension sensors are installed at the stress connection points between the slings and the spreader to acquire transient tension data. The IMU is installed within the main structural frame of the spreader to collect vertical acceleration data. The ranging sonar is installed at the bottom of the spreader, with its beam perpendicular to the seabed, to acquire the vertical distance between the spreader and the seabed.

[0019] The transmission layer includes umbilical cables or fiber optic composite cables. It transmits data acquired by the sensing layer to the main control layer via an industrial fieldbus protocol. The main control layer is located in the mother ship's control room and employs a programmable logic controller (PLC) with multi-core processing capabilities and a real-time operating system. The execution layer includes a hydraulic proportional relief valve and a release electromagnetic hydraulic latch. The hydraulic proportional relief valve is installed in the hydraulic control circuit of the mother ship's heave compensation winch and is used to adjust the cable release damping. The release electromagnetic hydraulic latch is installed at the bearing lifting point at the end of the automatic spreader and is used to lock or release the ecological reef. An electrical signal control link is established between the main control layer and the execution layer.

[0020] Please see the appendix Figure 2 This invention provides an intelligent control method for a special automatic lifting tool for ecological artificial reef installation, comprising the following steps: S10, the signal decoupling module continuously receives transient tension data from the tension sensor, performs variational mode decomposition calculation, and decomposes the tension signal containing environmental interference into the effective load background tension component and wave and cable elastic disturbance component in the frequency domain. S20, the pose observation module receives the vertical acceleration data from the inertial measurement unit and the vertical distance data from the ranging sonar, performs covariance update and gain fusion calculation, and outputs the vertical displacement relative to the seabed. S30, the topology mapping module receives the wave and cable elastic disturbance components, calculates the wave excitation energy variance within a specific time window, constructs a two-dimensional phase plane state space based on the relative seabed vertical displacement and the effective load background tension component, and uses the wave excitation energy variance as a control variable to generate and adjust the boundary of the stable support geometry within the two-dimensional phase plane state space. S40, the feedforward control module performs phase-lag-free differential processing on the effective load's base tension component and the vertical displacement relative to the seabed to obtain the phase trajectory evolution slope. Combined with the preset hard landing threshold and adaptive dead zone bandwidth, it calculates and generates a feedforward voltage command, which is then sent to the hydraulic proportional relief valve to actively intervene in the evolution path of the system state point in the two-dimensional phase plane state space. S50, the uncoupling execution module continuously monitors the position of the state point in the two-dimensional phase plane state space. When the state point evolves into the stable support geometric domain and the integral of the state point's dwell time meets the system's preset conditions, it outputs a high-level release command to the uncoupling electromagnetic hydraulic latch, triggering the execution of the physical uncoupling action.

[0021] In this embodiment, in step S10 of the system execution control method, the signal decoupling module specifically executes the following logic to extract and decouple the original environmental interference data in the frequency domain.

[0022] Before data processing, rigorous fault tolerance verification and preprocessing are required for the multi-source heterogeneous data. The sensor components include a tension sensor, an inertial measurement unit, and a ranging sonar. To establish a common time reference for the multi-source heterogeneous data, a timestamp synchronization mechanism is implemented. This mechanism uses the IEEE 1588 precise time protocol or sends hardware pulse-second signals via the main control layer to achieve microsecond-level clock synchronization between sensor nodes. Considering potential data loss due to electromagnetic and underwater acoustic interference in the marine environment, a sliding compensation logic based on cubic spline interpolation is introduced in the timestamp alignment stage. When a missing sensor data frame is detected, transient interpolation is performed based on valid data from adjacent timestamps to ensure the completeness of each physical quantity on the same time slice. The data acquisition frequency is set according to the system's preset control response bandwidth, and the set value is greater than twice the Nyquist sampling standard of the highest frequency of sea surface wave disturbance and the highest frequency of cable elastic vibration. The main control layer acquires transient tension data according to the set data acquisition frequency and records it as... Simultaneously, outlier points that exceed the physical limits are removed.

[0023] Acquire transient tension data after cleaning Subsequently, the system does not directly use a conventional digital low-pass filter for processing because conventional filters introduce unavoidable phase delays, thereby degrading the real-time tracking performance of the control system. Based on these considerations, this embodiment constructs a variational mode decomposition mathematical model to perform phase-hysteresis-free non-recursive decoupling of the signal. The variational mode decomposition operation converts the transient tension data... Assuming the system consists of a superposition of multiple intrinsic mode function (EMF) components with discrete center frequencies, a constrained variational model is constructed to achieve compact separation in the frequency domain. The objective is to minimize the sum of the one-sided spectral estimation bandwidths of each EMF component, while being constrained by the fact that the sum of the EMF components is strictly equal to the transient tension data. Physical equality constraints.

[0024] To solve the constrained variational model described above, the system introduces a quadratic penalty factor and Lagrange multipliers into the original objective function. In this embodiment, the quadratic penalty factor is set to a value between 1000 and 3000, chosen to balance data fidelity and filtering smoothness; a value that is too high will cause the reconstructed signal to lose high-frequency details, while a value that is too low will weaken the ability to resist white noise interference. Lagrange multipliers are used to strengthen the strictness of the equality constraints. After introducing these two parameters, the constrained variational model is transformed into an unconstrained variational problem. The system iteratively updates each eigenmode function component and its center frequency in the frequency domain using the alternating direction multiplier method. During the frequency domain update iteration calculation process, to prevent matrix calculation singularities and iteration divergence risks caused by the denominator approaching zero due to the overlap of noise frequency components and the center frequency, the system constantly superimposes a multiplier with a value of 10 into the denominator term of the underlying operation. −6 The small positive regularization term. For the specific iterative derivation process of the alternating direction multiplier method for finding the optimal solution in the frequency domain, those skilled in the art can refer to existing nonlinear programming and numerical optimization algorithms, which are well-known techniques in this field and will not be elaborated upon here.

[0025] After the iterative calculation converges, multiple decoupled intrinsic mode function components and their respective center frequencies are obtained. At this point, physical determination cannot rely solely on a single extreme value or frequency; the system introduces a multi-dimensional evaluation logic based on center frequency and modal energy proportion. Based on the physical characteristics of wave periods in marine engineering, the system pre-sets a frequency division threshold. This threshold is not a fixed constant but is adaptively assigned based on the average peak frequency of the wave spectrum from meteorological and hydrological stations in the target sea area for that year. The system extracts continuous low-frequency dominant modal components whose center frequency is lower than the threshold and whose proportion in the total signal energy is greater than a preset weight, and determines them as effective payload background tension components. The effective load's background tension component The system characterizes the basic physical loads of the ecological artificial reef, consisting of its static gravity, buoyancy, and steady hydrodynamic drag. Subsequently, the system superimposes and reconstructs multiple high-frequency modal components with center frequencies greater than the frequency threshold and corresponding to the frequency band of sea surface swell disturbance, generating wave and cable elastic disturbance components. A general description of signal dynamic decoupling corresponds, in its physical implementation, to the process in this step of mapping the mixed original signal to independent physical variables based on adaptive frequency thresholds and energy proportion rules.

[0026] In this embodiment, in step S20 of the system execution control method, the pose observation module is responsible for optimizing the estimation of multi-source sensor data, thereby outputting a high-precision vertical displacement relative to the seabed. .

[0027] In complex marine engineering applications, the direct application of data from multiple sensors carries significant risks due to their inherent physical limitations. Relying solely on vertical acceleration data from the inertial measurement unit (IMU) for secondary integration to obtain displacement is highly susceptible to divergent integral drift caused by low-frequency sensor bias errors and vibration noise. Based on these physical limitations, this embodiment establishes a Kalman filter data fusion model, incorporating vertical distance data from ranging sonar, which possesses absolute reference properties, into the kinematic derivation system.

[0028] The system combines the discrete-time sampling period of the inertial measurement unit (IMU) with real-time acquired vertical acceleration data to construct a linear state prediction equation. This equation aims to independently predict the prior state estimation vector, including prior displacement and prior velocity, based on the internal system state transition matrix and control input gain matrix, independent of external observations. The discrete-time sampling period ranges from 0.005 seconds to 0.02 seconds, and its value is directly controlled by the hardware polling cycle of the fieldbus between the main control layer and the IMU.

[0029] After completing the internal state deduction, the system needs to incorporate external measurements for correction. The vertical distance data provided by the ranging sonar does not exhibit integral divergence over long timescales and possesses reliable absolute position boundary reference constraints. However, this type of sensor is susceptible to high-frequency underwater acoustic multipath reflection noise interference caused by uneven impedance of the seabed sediment. Based on this consideration, the system extracts the vertical distance data output by the ranging sonar and uses it as the observation input vector to construct an observation equation that correlates the internal state with the external measurements. The relative pose fusion observation feature based on Kalman filtering in the claims, at the implementation level, relies on this observation equation to incorporate the absolute distance reference of the external physical space into the kinematic inference algorithm framework.

[0030] After synchronously acquiring the prior state estimation vector and the observation input vector at the current moment, the system performs covariance update and Kalman gain fusion operations. The system derives and calculates the current Kalman gain matrix based on the prior error covariance matrix of the previous cycle. This gain matrix is ​​used to dynamically allocate the system decision weights between the internal kinematic prediction confidence and the external sonar measurement confidence. The calculation involves inverting the innovation variance matrix, which contains the observation matrix and the measurement noise covariance matrix. During the matrix inversion operation, to prevent singularity divergence caused by rank deficiency in the measurement noise covariance matrix due to extreme underwater acoustic blind zones, the system algorithm forcibly superimposes a small constant anti-divergence constraint term on the main diagonal of the matrix to be inverted. This constant value is preferably set to 10. −6 .

[0031] To enhance the system's fault tolerance in harsh sea conditions, the system avoids relying on a single fixed parameter and instead introduces multi-dimensional weighted logic based on signal-to-noise ratio (SNR). The initial diagonal elements of the measurement noise covariance matrix are assigned values ​​based on the static ranging error variance benchmark calibrated at the sonar equipment's factory. During operation, the system extracts the acoustic echo SNR data returned from the sonar hardware in real time. When the current SNR value is determined to be lower than the system's preset safety threshold, the system synchronously amplifies the specific element values ​​of the measurement noise covariance matrix using a nonlinear scaling factor, thereby automatically reducing the Kalman gain's decision-making confidence in the current unreliable acoustic measurement. Considering that fish schools may cross the sea or bottom sediment may blow sand in the ecological reef deployment area, causing accidental blockage of the sonar beam, the system adds a sudden outlier isolator based on kinematic envelope before introducing external measurements into the observation equation. The system calculates the theoretical maximum displacement change envelope interval at the current moment based on the prior state estimation vector (including velocity and acceleration information) of the previous moment; when the step change of the vertical distance data output by the ranging sonar exceeds the envelope interval, the system determines that it has encountered hard obstruction by underwater organisms or suspended objects.

[0032] After acquiring the gain weights, the system applies the Kalman gain to the latest measurement residual to perform posterior correction of the state vector, ultimately outputting the corrected optimal posterior state estimation vector. Simultaneously, a closed-loop update of the posterior error covariance matrix is ​​completed, providing a reliable initial data basis for the iterative derivation in the next computation cycle. After the aforementioned closed-loop iterative fusion operation, the system extracts the first vector element representing the position from the optimal posterior state estimation vector. This extracted value serves as the relative vertical displacement of the seabed after successfully eliminating low-frequency cumulative drift error and high-frequency random acoustic noise. The vertical displacement relative to the seabed Subsequently, it is directly used as the core coordinate parameter output for phase plane topology mapping operations.

[0033] In this embodiment, in step S30 of the system execution control method, the topology mapping module performs the following logical calculations to establish an adaptive determination mechanism that correlates the system operation status with the severity of marine environmental disturbances.

[0034] In marine construction operations, sea state disturbances exhibit high nonlinearity and randomness. Directly using a single transient excitation extreme value as the criterion can easily introduce quantization noise and cause malfunctions in the control system. Based on these considerations, the system performs time-sliding window integration and quantization calculations of wave excitation energy. The system extracts the wave and cable elastic disturbance components output from the signal decoupling step and performs energy variance integration within the time sliding window. The technical purpose of this calculation is to smoothly transform the high-frequency alternating time-domain disturbance signal into a slowly varying state variable reflecting the macroscopic intensity of the marine excitation kinetic energy. The system sets the time sliding window parameters as follows: Calculate the variance of wave excitation energy The mathematical model is as follows: ; in, This is the current system timestamp. For the variable in the integration time The components of wave and cable elasticity disturbance at any given moment. Time sliding window parameters Mathematical expectation and mean of the internal wave and cable elastic disturbance components. Time sliding window parameters. The value is set based on the typical long-cycle swell frequency of the meteorological statistics of the target sea area in that year. It is usually set to include the duration of 1 to 2 complete wave cycles, and its value range is preferably between 5 and 15 seconds, so as to ensure that the integration result can filter out transient spikes while keeping a keen follow of the trend of sea state deterioration.

[0035] Since the main control layer is a discrete digital system, the above integral formula is converted into a summation form based on discrete sampling sequences during actual hardware deployment. When performing discrete mean square error calculation, the system internally incorporates division-by-zero error prevention logic for the denominator. The system continuously counts the total number of valid and non-missing discrete sampling points within the current time window. When the total number of valid sampling points is detected to be close to zero (e.g., during system initial power-on or in the event of a severe bus interruption), the algorithm forcibly assigns the denominator to a preset default safety constant, thereby avoiding hardware division-by-zero overflow faults and ensuring the continuity of the underlying control timing.

[0036] Traditional methods rely solely on instantaneous threshold values ​​from a single sensor to determine the seabed status. When the spreader is affected by lateral ocean current shear or the mother ship by vertical wave heave, false disengagement judgments often occur due to sudden drops in tension. To establish a multi-dimensional, comprehensive safety assessment basis, the system constructs a two-dimensional phase plane state space based on multi-dimensional state parameters. Before extracting the state parameters, the system uses a unified global timestamp reference to measure the vertical displacement relative to the seabed output in each step. With the background tension component of the effective load Strict timing alignment is performed. The system uses the aligned vertical displacement relative to the seabed. As the horizontal axis, the effective load background tension component Using the vertical axis as the coordinate axis, an orthogonal two-dimensional phase plane state space is constructed. This state space characterizes the relative kinematic relationship and force coupling state between the ecological reef and the seabed. Within this two-dimensional phase plane state space, the origin of the coordinate system is defined as the ideal absolutely static landing state where the physical displacement is reduced to the limit zero point and the effective load is fully supported and borne by the seabed.

[0037] The system generates and adjusts the boundary of the stable support geometric domain within a two-dimensional phase plane state space. Within this state space, the system delineates a specific decision region, defined as the stable support geometric domain. This domain defines the physical state range in which the ecological reef completely penetrates the surface mud and obtains stable support from the seabed. Considering the increased wave disturbance kinetic energy under severe sea conditions, the risk of impact rebound upon bottoming increases significantly. The system then calculates the variance of the wave disturbance energy... As a penalty variable for dynamic topology mapping, adaptive shrinkage or expansion adjustment is performed on the boundary of the stable supporting geometry.

[0038] The system internally stores nominal displacement and nominal tension trigger thresholds calibrated under calm, wave-free conditions. The nominal displacement trigger threshold is determined based on the sonar measurement blind zone and the geometric height of the spreader chassis; the nominal tension trigger threshold is determined based on the spreader's static buoyancy in the water. The system constructs an internal mapping function based on the variance of wave excitation energy. The two nominal thresholds mentioned above are dynamically corrected as independent variables to generate dynamic displacement threshold and dynamic tension threshold.

[0039] In this embodiment, the stable support geometric domain is specifically constructed in the two-dimensional phase plane state space as a quarter-elliptical envelope centered at the origin and with the dynamic displacement threshold and dynamic tension threshold as its major and minor axes. Its geometric boundary satisfies the following inequality equation: ; in, For dynamic displacement threshold, The dynamic tension threshold is used. The advantage of using an elliptical nonlinear boundary instead of a traditional rectangular orthogonal boundary is that, when displacement and tension simultaneously approach the critical value, the elliptical boundary provides a smoother coupling chamfer transition, avoiding high-frequency oscillating transitions at the edge of the decision domain caused by minor boundary overflows in a single dimension. The core penalty control logic of this mapping function is: when the wave excitation energy variance... An increase in the numerical value indicates an intensification of sea surface swell. At this point, the mapping function, by introducing a non-positive mapping coefficient, forces the dynamic displacement threshold and dynamic tension threshold to decrease synchronously from their nominal values, causing the boundary of the stable support geometry to converge towards the origin. This contraction mapping mechanism forces the system to meet stringent conditions of smaller residual displacement and lower residual tension under severe sea conditions in order to generate a bottoming trigger signal. Conversely, when the wave disturbance energy variance... When the temperature is reduced and the sea conditions are stable, the dynamic boundary expands outward toward the nominal threshold to improve operational efficiency.

[0040] Furthermore, to provide deadlock prevention and fault tolerance logic, the system sets a lower bound for the underlying safety limit of the mapping function. This lower bound is obtained through a conservative estimate of the maximum expected sea state lower limit displacement and residual tension. This applies regardless of the variance of the wave excitation energy. No matter how extreme the magnitude, the dynamic threshold output by the mapping function is clamped and limited above the lower bound of the bottom safety limit to prevent the geometric domain from shrinking to zero or negative values ​​due to extreme waves, thereby causing the control system to fall into a permanent closed-loop stagnation state that cannot be triggered.

[0041] In this embodiment, in step S40 of the system execution control method, the feedforward control module performs the following logical operation: actively intervenes in the evolution path of the system state point in the two-dimensional phase plane state space, thereby suppressing the physical oscillation at the bottom instant.

[0042] In conventional automatic control systems, the rate of change of physical quantities is typically obtained using direct differential calculations or by adding a digital low-pass filter. However, direct differential calculations severely amplify sensor quantization noise, while low-pass filters introduce unavoidable phase hysteresis, which can easily lead to control divergence in hydraulic damping control scenarios requiring millisecond-level response. Based on these physical limitations, this embodiment introduces a nonlinear tracking differentiator to perform hysteresis-free smoothing of the signal.

[0043] Before performing differential processing, the heterogeneous delay problem of multi-source data must be addressed. This takes into account the background tension component of the payload. Vertical displacement relative to the seabed The variational mode decomposition module and the Kalman filter observation module output data independently in the parallel computing branch. Due to the difference in algorithm complexity, there is a significant discrepancy in the underlying microprocessor computation time between the two. To eliminate the timing misalignment illusion introduced by parallel computing, the system sets up a hardware timestamp cache queue based on a global high-precision clock. The main control layer compares the timestamps of the two data streams in real time, performs a cache delay on the faster-computed data stream, and forcibly extracts the effective load background tension component with strict contractual timing. Vertical displacement relative to the seabed .

[0044] After timing alignment, the system will effectively control the background tension component of the payload. Vertical displacement relative to the seabed The input is synchronously fed into a discrete nonlinear tracking differentiator. The nonlinear tracking differentiator, through its internal speed-range optimal control synthesis function, filters out high-frequency noise while outputting a smooth differential derivative signal in real time. To prevent high-frequency chattering in the digital hardware, the system sets the speed factor parameter of the differentiator based on the highest physical response frequency of the hydraulic proportional relief valve. The value of this speed factor parameter is preferably limited to between 50 and 150 to ensure that the frequency of the generated derivative signal does not exceed the physical tracking limit of the underlying hydraulic actuator. After tracking calculations, the system obtains the smoothed derivative signal of the effective load's base tension component. and the smoothed derivative signal of the vertical displacement relative to the seabed .

[0045] After acquiring the smoothed derivative signal, the system calculates the phase trajectory evolution slope based on the geometric relationships within the two-dimensional phase plane. (Phase trajectory evolution slope) The formula used to characterize the intensity of the landing kinetic energy conversion during the process of the current ecological artificial reef approaching the seabed is as follows: ; in, and These are the smoothed derivative signals of the effective load's background tension component and the smoothed derivative signal of the vertical displacement relative to the seabed, respectively, calculated above. Their units are relative to the rate of stress change (e.g., Newtons per second) and velocity (e.g., meters per second), respectively. During the division operation described above, to prevent issues arising from the smoothed derivative signal of the vertical displacement relative to the seabed... The slope of the phase trajectory evolution tends to infinity when the value approaches zero (i.e., the lifting device is in a transient static hovering state in the vertical direction). Therefore, the system forcibly implants zero-division clamping logic in the denominator during low-level calculations. When this is detected... The absolute value is less than the system's preset micro-motion threshold (preferably set to 10). −3 When the speed reaches (m / s), the algorithm determines that there is no substantial vertical relative approximation, immediately stops performing the division operation, and directly changes the phase trajectory evolution slope. Forced zeroing prevents overflow in hardware logic operations.

[0046] To avoid excessive intervention by the system in response to minor hydrodynamic disturbances, an adaptive dead-time bandwidth is defined within the module. This adaptive dead-time bandwidth is related to the variance of the wave excitation energy output from the step calculation. This directly constitutes a proportional function mapping logic. The scaling factor of this proportional function is comprehensively calibrated based on the inherent mechanical clearance of the mother ship's heave compensation winch and the dead zone leakage of the hydraulic motor. When the wave excitation energy variance... When the background sea conditions are rough and environmental clutter is increased, the system proportionally widens the adaptive dead zone bandwidth. The technical purpose of this parameter selection logic is to use the increased dead zone to absorb trajectory jitter caused by high-frequency waves and avoid high-frequency false triggering of the hydraulic proportional relief valve; while in calm sea conditions, the adaptive dead zone bandwidth is adaptively reduced, improving the system's ability to sensitively detect minute force changes at the moment of bottoming.

[0047] During the output control command phase, the system avoids using traditional hard step switching functions for control logic decisions. A hard landing threshold is pre-set within the system. This threshold is determined based on the elastic modulus of the mother ship's heave compensation winch cable and the ultimate compressive yield strength of the ecological reef concrete frame, and is used to quantify the maximum bottoming impact gradient the system can tolerate. The system calculates the phase trajectory evolution slope in real time. The effective over-limit slope deviation is obtained by subtracting the current adaptive dead zone bandwidth from the absolute deviation between the current and the hard landing threshold. When the effective over-limit slope deviation is greater than zero, it is determined that the current system has a strong tendency for hard landing impact. At this time, the system uses a preset smoothing saturation function to perform a nonlinear mapping on the effective over-limit slope deviation, thereby generating a feedforward voltage command. This smoothing saturation function has continuous first-order mathematical derivative characteristics, and its output upper limit is strictly clamped to the maximum control voltage limit allowed by the hydraulic proportional relief valve.

[0048] The smooth saturation function is constructed using a hyperbolic tangent function (Tanh function) model with a scaling factor introduced. Feedforward voltage command. The calculation formula is set as follows: ; in, This is the maximum control voltage limit for the hydraulic proportional relief valve. For effective over-limit slope deviation, The feedforward gain coefficient is used to calibrate the hydraulic valve spool opening curve. Because the hyperbolic tangent function exhibits a smooth, approximately linear following characteristic near zero, and smoothly approaches the asymptote at the limit deviations... This mapping mechanism effectively avoids severe pressure pulsations in the hydraulic circuit caused by discrete control steps, thereby fundamentally suppressing the destructive water hammer effect in the hydraulic pipeline. The purpose of selecting this type of nonlinear function with continuous and smooth transition characteristics is to completely eliminate the destructive water hammer effect induced by sudden changes in control voltage in the hydraulic circuit. After the calculation is completed, the main control layer sends the generated feedforward voltage command to the hydraulic proportional relief valve in the execution layer, actively increasing the cable release damping of the mother ship's heave compensation winch, successfully achieving a flexible soft landing of the ecological artificial reef.

[0049] In this embodiment, in step S50 of the system execution control method, the decoupling execution module performs the following logical operation to realize the state monitoring closed loop and the final physical separation action.

[0050] In deep-water operations, due to the coupling effect of bottom current shear and cable elastic restoring force, ecological artificial reefs often experience high-frequency, slight bouncing in the initial stage of bottom contact. Relying solely on a single transient extreme value signal penetrating the geometric boundary to trigger action could easily lead to dangerous deployment of the load in a semi-suspended state. Based on these physical causal considerations, the system avoids using a one-sided, single-moment threshold comparison. Instead, it performs real-time comparison of the topological relationship between the state coordinate points and the stable support geometric domain, and establishes a multi-dimensional state dwell verification logic based on a time sliding window.

[0051] The system extracts the vertical displacement relative to the seabed in real time based on the underlying hardware polling cycle. With the background tension component of the effective load The two simultaneously constitute transient coordinate points. The system defines a discrete state indicator function, which takes a value of 1 when the current transient coordinate point falls completely within the constructed stable support geometry, and a value of 0 otherwise. Combined with a preset observation time sliding window length, the system calculates a dwell reliability index. To broaden the system's compatibility with various time-based filtering algorithms, the resident reliability index... The calculation logic is not limited to the arithmetic mean; its core rule is to sum the discrete state characteristic function values ​​within the observation time window over time, and then divide the sum by the actual number of effective sampling points within the window. This quantitatively assesses the probability that the stress and displacement state of the ecological artificial reef will remain stable within the safe landing limits over a continuous period of time. The length of the observation time window is determined based on the discrete sampling frequency of the underlying controller and the dead-zone response time of the hydraulic solenoid valve. The corresponding time span is usually limited to between 1.5 seconds and 3.0 seconds to balance the safety of system action triggering with the overall efficiency of offshore operations.

[0052] In calculating the credibility index of residency During the process, the system synchronously executes verification logic for time alignment and data loss tolerance of multi-source data. If some historical state sequences within the sliding window are lost due to transient interference on the communication bus, the actual number of valid sampling points will change dynamically. To prevent engineering computational risks, the system embeds anti-divergence and anti-zero-checking mechanisms at the underlying level. The system counts the actual number of valid sampling points within the current sliding window in real time. When it detects that the actual number of valid sampling points is close to zero, or the data loss ratio exceeds the preset safety tolerance, the algorithm determines that the current communication signal-to-noise ratio is insufficient to support high-precision judgment. At this time, the system forcibly terminates the decoupling judgment branch of the current cycle, stops executing the division operation to avoid hardware zero-overflow failure, and retains the resident reliability index. Reset to zero until the communication link is restored and the continuous valid state queue is refilled.

[0053] In obtaining reliable residency credibility indicators Subsequently, the system triggers a zero-stress uncoupling action based on this indicator and issues a low-level execution command. The system internally presets a trigger confidence threshold. The trigger confidence threshold The value is strictly limited to between 0.85 and 0.95, specifically set based on the statistical probability constraint limit for unplanned accidental unhooking in deep-sea construction operation procedures. The system will calculate the dwell reliability index in real time. With trigger confidence threshold Perform continuous comparison and judgment. When Continuously greater than or equal to the trigger confidence threshold At this point, the system physically determines that the gravity of the ecological reef has been completely and stably transferred to the seabed, and the current lifting system is in a completely relaxed, zero-stress safe state. The main control layer generates a disengagement trigger signal and sends the command to the hydraulic release circuit at the end of the lifting device via an opto-isolated output channel. The system collects the feedback current of the hydraulic solenoid valve coil in real time. When the time distribution of the characteristic inflection point of the current feedback curve conforms to the physical electromagnetic law of the valve core's full-stroke opening, the system finally confirms that the physical separation action has been completed.

[0054] In addition, to prevent physical jamming of the mechanical locking mechanism due to seabed sediment, the system immediately activates a secondary physical separation verification logic based on tension feedback after confirming valve core opening. The system initiates a 3-5 second disengagement monitoring window, continuously reading transient tension data. If, within this window, the transient tension data does not decrease to the minimum unloaded value range representing only the weight of the sling, and the vertical displacement of the sling relative to the seabed shows a reverse pulling trend with the rise and fall of the mother ship, the system determines that a mechanical jamming-type disengagement failure has occurred. The main control layer immediately triggers the highest-level audible and visual alarm and immediately issues a constant tension follow-up command to the heave compensation winch, preventing the winch from continuing to retrieve the cable and preventing the reef from being forcibly dragged and rolled in a semi-disengaged state, ensuring the absolute bottom-line safety of the equipment and construction operations. This execution mechanism, through decision support based on state-space topology evolution and dual verification of electrical characteristics, successfully ensures highly reliable control and smooth release of the reef during landing.

[0055] Specific application examples: Implementation background and hardware configuration: This embodiment provides an intelligent control system for a specialized automatic lifting device for installing ecological artificial reefs, set up for an ecological artificial reef deployment operation in an offshore area of ​​the South China Sea. The target sea area has a water depth of approximately 55 meters, and the sea state during the operation is level 3 (average wave height approximately 1.5 meters, typical long-period swell frequency approximately 0.12 Hz). The ecological artificial reef to be deployed is a reinforced concrete structure, with an aerial weight of 20 tons and a static buoyancy weight of approximately 12 tons in the water (corresponding to a nominal tension of approximately 117.6 kN).

[0056] The system hardware physical devices mainly consist of the perception layer, transmission layer, main control layer, and execution layer: Sensing layer: Arranged on the body of the automated spreader, including a tension sensor with a range of 50 tons installed at the force connection node of the spreader; an inertial measurement unit with a sampling rate of 100Hz installed in the main structural frame of the spreader; and a ranging sonar with a vertical orientation to the seabed and a measurement accuracy of 0.1 meters.

[0057] Transmission layer and main control layer: Data is transmitted to the programmable logic controller in the mother ship's control room via umbilical cable or optical fiber composite cable and industrial fieldbus.

[0058] The execution layer includes a hydraulic proportional relief valve (adjusting release damping) installed in the hydraulic circuit of the mother ship's heave compensation winch, and a solenoid hydraulic release latch at the end of the automatic spreader.

[0059] Algorithm dynamic execution process and timing diagram analysis: Please see Figure 3 This figure illustrates the trajectory of the sling tension over time (X-axis: operation time - seconds) during a complete ecological artificial reef deployment, from bottoming out to final release (Y-axis: sling tension - kN). High tension indicates suspension in the water, while zero tension indicates complete support on the seabed and detachment. The system executes the following intelligent control steps: S10, Multi-source heterogeneous data acquisition and dynamic signal decoupling: Implementation Details and Theory: The system achieves microsecond-level timestamp synchronization via the IEEE 1588 protocol or second pulses, and employs sliding compensation logic based on cubic spline interpolation to handle data packet loss caused by underwater acoustics. When the spreader is lowered to 10 meters above the seabed, the tension sensor is affected by the heave of sea state 3 waves and the elastic vibration of the cable, causing its original transient tension to fluctuate wildly between 90kN and 150kN. Figure 3 In this context, the light gray thin line appears as wavy fluctuations and spikes during the 0-10 second period, indicating that traditional systems directly read the raw data and are highly susceptible to environmental interference.

[0060] To address the phase delay, a constrained variational mode decomposition mathematical model is constructed, introducing a quadratic penalty factor (1000-3000) and Lagrange multipliers. The alternating direction multiplier method is employed (with 10 multipliers superimposed in the denominator).−6 The singularity-preventing regularization term is non-recursively decoupled in the frequency domain. The system extracts the low-frequency dominant modes with a center frequency less than 0.12Hz (peak frequency of the target sea area) as the effective payload background tension components. In the figure, this corresponds to the smooth fluctuation of the black solid line in the 0-8.5 second range, and the pure 117.6kN effective load tension was successfully extracted.

[0061] S20, relative pose fusion observation based on Kalman filtering: Implementation Details and Theory: In seabed sandstorm areas, single ranging sonars are highly susceptible to range jumps (e.g., from 8 meters to 2 meters) due to multipath reflection. The system constructs a Kalman filter model, incorporating sonar data as the observation input vector into the IMU kinematics extrapolation system. When performing covariance updates and inverting the innovation variance matrix, a 10-fold matrix is ​​superimposed on the main diagonal. −6 The anti-divergence constraint.

[0062] The system extracts the acoustic echo signal-to-noise ratio in real time and automatically scales the observation noise covariance nonlinearly. It also incorporates a sudden outlier isolator based on the kinematic envelope, which forcibly isolates the data frame when the sonar step exceeds the envelope range. The system relies solely on internal state prediction equations for derivation, ultimately outputting a smooth and drift-free vertical displacement relative to the seabed. .

[0063] S30, Wave Energy Assessment and Two-Dimensional Phase Plane Dynamic Topological Mapping: Implementation details and theory: The system extracts wave components and calculates time sliding window parameters. (Set to 5-15 seconds) Wave excitation energy variance The formula is: ; The system uses displacement x-axis Construct an orthogonal two-dimensional phase plane state space for the ordinate.

[0064] Given the sea state of level 3 in this area, the wave kinetic energy is relatively large, and the system will... The decision boundary is adaptively shrunk as a penalty variable. The stable support geometry is constructed as an elliptical envelope surface: To prevent false triggering, the system automatically reduces the nominal tension trigger threshold from 12kN to 5kN (i.e., the dynamic tension threshold). The system requires extremely weak contact tension to determine landing and also has a bottom safety limit lower bound to prevent deadlock.

[0065] S40, nonlinear tracking differential and winch damping feedforward control linkage: Implementation details and theory: After time-stamped alignment, the data is input into a discrete nonlinear tracking differentiator (with a velocity factor set between 50-150 to avoid high-frequency chatter) to obtain a smooth derivative signal without phase hysteresis. The system calculates the phase trajectory evolution slope. (Includes zero-clamping logic to prevent small motion thresholds).

[0066] When the artificial reef is approximately 1.5 meters from the seabed (corresponding to the vertical black dashed line at 8.5 seconds in the diagram), the system calculates that the slope, after deducting the dead zone bandwidth, exceeds the hard landing threshold, indicating a strong risk of impact. The main control layer immediately generates a feedforward voltage command using the hyperbolic tangent function. .

[0067] A command such as 6.5V is sent to the hydraulic relief valve to actively increase the release damping. In the timing diagram, this is represented by a gentle downward slope (black solid line) between 8.5 and 10.5 seconds, with tension smoothly and without impact as it descends to 0, achieving a perfect soft landing. The vertical dashed line at 10.0 seconds represents the moment the reef physically contacts the seabed. In stark contrast, the light gray thin line shows violent oscillations after 10 seconds, indicating that the conventional system made a hard landing without deceleration, generating a water hammer recoil force of nearly 250kN and mechanical bouncing.

[0068] S50, topology trajectory triggering and zero-stress decoupling execution: Implementation details and theory: After the artificial reef smoothly touches the bottom, the system does not use a single extreme value trigger, but instead calculates the reliability index of the state point's residence in the stable support geometric domain in real time. (Sliding window length 1.5-3.0 seconds), and built-in anti-zero communication packet loss verification.

[0069] In the chart, this is represented by the solid black line in the low-level, flat segment between 10.5 and 12.5 seconds. Although the artificial reef bottomed out at 10 seconds, the system patiently observed for another 2.0 seconds, confirming that the force was relatively stable and there was no rebound. At the vertical dashed line at 12.5 seconds, the indicator... If the value is greater than the confidence threshold of 0.9, the dwell verification is deemed successful, a level signal is generated to trigger physical uncoupling, and the tension is instantly and completely reduced to zero.

[0070] After confirming that the valve core is open, activate the physical separation secondary verification: start a 3-5 second window. If the tension does not drop to the no-load range and displacement occurs, it is determined that the mechanical jamming has failed, and an alarm is forcibly triggered and a constant tension follow command is issued to prevent dragging and rolling.

[0071] Experimental verification and comparison of statistical effects of multi-dimensional work indicators: To verify the technical effectiveness, we compared the data from 20 consecutive deployments of the traditional control system and the system of this invention under the same sea area and operating conditions. Please refer to [link / reference]. Figure 4(Light gray bars represent traditional systems, and dark gray bars represent the system of this invention): Indicator 1: Maximum impact force upon bottoming out (kN) Meaning: The maximum peak tensile force that the cables and lifting gear withstand at the moment the artificial reef impacts the seabed.

[0072] Comparative performance: Traditional systems use simple limit switches and constant damping hydraulic control, resulting in an impact force as high as 245.8kN, generating tremendous destructive force; due to nonlinear feedforward intervention, the maximum impact force of this invention is only 125.4kN (extremely close to the static buoyancy weight of 117.6kN), equivalent to gently placing it on the seabed without additional impact force.

[0073] Indicator 2: Number of times the hook accidentally became unhooked during 20 operations Meaning: The number of serious accidents in which the tension of a ship drops suddenly due to the heave caused by wind and waves, and the system mistakenly believes that it has hit the bottom and releases prematurely in mid-air.

[0074] Comparative performance: The traditional system experienced three mid-air dangerous releases; the present invention, relying on two-dimensional phase plane dynamic topology mapping and time sliding window verification, reduced the number of accidental unhookings to zero, and there was no mechanical jamming, achieving a 100% success rate.

[0075] Indicator 3: Decoupling delay time (seconds) Meaning: The time difference between when the artificial reef physically touches the bottom and when the system finally unlocks and releases the reef.

[0076] Performance comparison: Traditional systems decouple immediately upon reaching the bottom (only 0.2 seconds), seemingly fast but prone to malfunctions; the system of this invention shows 2.0 seconds. This is not an algorithmic delay, but a deliberately designed safety observation window, sacrificing a minimal efficiency of 2 seconds for absolute safety with zero false decouplings.

Claims

1. An intelligent control system for a special automatic lifting tool for installing ecological artificial reefs, characterized in that, include: Tension sensor, inertial measurement unit, ranging sonar, hydraulic proportional relief valve, unhooking electromagnetic hydraulic latch and control logic module; The control logic module includes: The signal decoupling module performs variational mode decomposition on the transient tension data collected by the tension sensor and outputs the effective load background tension component and the wave and cable elastic disturbance component. The pose observation module performs Kalman filtering on the vertical acceleration data collected by the inertial measurement unit and the vertical distance data collected by the ranging sonar, and outputs the vertical displacement relative to the seabed. The topology mapping module obtains the wave excitation energy variance from the wave and cable elastic disturbance components, constructs a two-dimensional phase plane state space and a stable support geometric domain from the relative seabed vertical displacement and the effective load background tension component, and adjusts the boundary of the stable support geometric domain using the wave excitation energy variance. The feedforward control module performs tracking differentiation on the effective load's base tension component and the relative vertical displacement of the seabed to obtain the phase trajectory evolution slope, and sends a feedforward voltage command to the hydraulic proportional relief valve. When the state point formed by the vertical displacement relative to the seabed and the base tension component of the effective load in the two-dimensional phase plane state space falls into the stable support geometric domain and the integral of the dwell time meets the system preset conditions, the unhooking execution module sends a high-level release command to the unhooking electromagnetic hydraulic latch.

2. The intelligent control system according to claim 1, characterized in that, The signal decoupling module executes a timestamp synchronization mechanism to establish clock synchronization between the tension sensor, the inertial measurement unit, and the ranging sonar; The signal decoupling module acquires the transient tension data according to a preset acquisition frequency. When the transient tension data is detected, transient interpolation processing is performed based on the valid data corresponding to adjacent timestamps. Finally, outlier points that exceed the preset physical limits are removed; The preset physical limit range is determined based on the maximum breaking force of the lifting device when fully loaded and the upper limit of the sensor's range.

3. The intelligent control system according to claim 2, characterized in that, The variational mode decomposition includes: Establish a constrained variational model that includes the superposition of multiple eigenmode function components with discrete center frequencies; Introducing a quadratic penalty factor and Lagrange multipliers into the constrained variational model transforms it into an unconstrained variational problem; The intrinsic mode function components and the discrete center frequencies are iteratively updated in the frequency domain by the alternating direction multiplier method. A regularization term with a small positive value is superimposed on the denominator of the underlying operation in the iterative update.

4. The intelligent control system according to claim 3, characterized in that, After the iterative update converges, the continuous low-frequency dominant mode components with a center frequency lower than the frequency division threshold and a proportion greater than the preset weight in the total signal energy are extracted, and the continuous low-frequency dominant mode components are determined as the effective load background tension components. Multiple high-frequency modal components with a center frequency greater than the frequency division threshold are superimposed and reconstructed to generate the wave and cable elastic disturbance components. The frequency division threshold is adaptively assigned based on the peak frequency of the average wave spectrum of the target sea area, and the preset weight is set based on the empirical proportion of tension energy of the effective load in the static state of seawater to the total signal energy.

5. The intelligent control system according to claim 1, characterized in that, The pose observation module constructs a linear state prediction equation and generates a priori state estimation vector based on the vertical acceleration data. The vertical distance data is used as the observation input vector to construct the observation equation, and the state estimate is updated by combining the observation equation with the prior state estimation vector. Acquire the acoustic echo signal-to-noise ratio data collected by the ranging sonar. When the acoustic echo signal-to-noise ratio data is lower than the safety threshold, amplify the element values ​​of the measurement noise covariance matrix in the Kalman filter by a nonlinear scaling factor. Anti-divergence constraint terms are superimposed on the main diagonal of the gain matrix in the Kalman filter matrix inversion operation stage. The safety threshold is determined based on the static ranging error variance benchmark calibrated at the factory of the sonar equipment.

6. The intelligent control system according to claim 5, characterized in that, The pose observation module includes a mutation outlier isolator based on kinematic envelope; Calculate the theoretical maximum displacement change envelope interval at the current moment based on the prior state estimation vector from the previous operation cycle; When the step change between the vertical distance data and the vertical distance data of the previous calculation cycle exceeds the theoretical maximum displacement change envelope, the acoustic measurement data of the current calculation cycle is isolated, and the relative seabed vertical displacement is obtained by pure inertial deduction relying solely on the linear state prediction equation within the current calculation cycle.

7. The intelligent control system according to claim 1, characterized in that, The topology mapping module performs energy variance integration within a time window for the wave and cable elastic disturbance components to obtain the wave disturbance energy variance. The wave disturbance energy variance is input into the mapping function, and combined with the nominal displacement trigger threshold and the nominal tension trigger threshold, the dynamic displacement threshold and the dynamic tension threshold are calculated and obtained. Next, the dynamic displacement threshold and the dynamic tension threshold are clamped and limited above the lower bound of the bottom safety limit; The greater the variance of the wave disturbance energy, the greater the contraction of the dynamic displacement threshold and the dynamic tension threshold towards the origin of the coordinate system. The shape of the stable support geometry is an elliptical envelope surface whose center coincides with the origin of the state space coordinate of the two-dimensional phase plane and whose major and minor semi-axes are respectively a quarter-elliptical envelope surface of the dynamic displacement threshold and the dynamic tension threshold after being clamped and restricted. The nominal displacement trigger threshold is determined based on the sonar measurement blind zone and the geometric height of the lifting device chassis, and the nominal tension trigger threshold is determined based on the static buoyancy of the lifting device in still water.

8. The intelligent control system according to claim 1, characterized in that, The feedforward control module performs a timing alignment operation based on a hardware timestamp cache queue for the effective load background tension component and the relative seabed vertical displacement. The aligned data is input into a discrete nonlinear tracking differentiator to obtain the smoothed derivative signal of the effective load's base tension component and the smoothed derivative signal of the relative seabed vertical displacement. The smoothed derivative signal of the effective load's base tension component and the smoothed derivative signal of the relative seabed vertical displacement are used to calculate the quotient, thereby obtaining the phase trajectory evolution slope. When the absolute value of the smoothed derivative signal of the vertical displacement relative to the seabed is less than the small motion threshold, the quotient calculation is stopped and the slope of the phase trajectory evolution is cleared to zero. The micro-motion threshold characterizes the velocity boundary of the lifting device when it is in a transient static hovering state in the vertical direction.

9. The intelligent control system according to claim 8, characterized in that, Calculate the absolute deviation between the phase trajectory evolution slope and the hard landing threshold, subtract the adaptive dead zone bandwidth to obtain the effective over-limit slope deviation, and the adaptive dead zone bandwidth and the wave disturbance energy variance form a positive proportional function mapping relationship. When the effective over-limit slope deviation is greater than zero, a smooth saturation function containing a hyperbolic tangent function with a scaling factor is used to nonlinearly map the effective over-limit slope deviation to generate the feedforward voltage command. The upper limit of the numerical output of the feedforward voltage command is clamped and limited to the maximum control voltage limit value allowed by the hydraulic proportional relief valve. The hard landing threshold is calculated based on the elastic modulus of the cable of the mother ship's heave compensation winch and the ultimate compressive yield strength of the ecological reef frame.

10. The intelligent control system according to claim 1, characterized in that, The decoupling execution module defines a discrete state characteristic function based on whether the state point falls within the stable support geometry. The discrete state characteristic function values ​​within the preset observation time window are summed over time, and the dwell reliability index is obtained by dividing the summation result by the actual number of effective sampling points within the window. When the dwell confidence index is not less than the trigger confidence threshold, it is determined that the dwell time integral meets the system preset condition and the high-level release command is generated. After the high-level release command is issued, the physical separation secondary verification logic is activated, the separation monitoring window is started to read the transient tension data, and when the transient tension data does not drop to the minimum value range of no load and the vertical displacement relative to the seabed shows a reverse pulling trend, an audible and visual alarm is triggered and a constant tension follow-up command is issued. The trigger confidence threshold is set based on the probability constraint limit of the unplanned accident rate of accidental unhooking in the deep-sea construction operation procedures. The minimum unload range is determined based on the static wet weight of the lifting gear when it is unloaded in seawater.

Images