Active depth-keeping and heave compensation method for marine seismic exploration buoy

Abstract

The application discloses a kind of active depth setting and heave compensation method of marine seismic exploration buoy, belong to marine seismic exploration technical field.The application constructs the four-in-one compound depth setting control system of "wave disturbance feedforward, depth feedback, multi-buoy cooperation, parameter self-adaption", including: S1 system initialization and parameter configuration, S2 wave disturbance perception and acceleration signal acquisition, S3 heave acceleration signal preprocessing and heave motion state estimation, S4 heave displacement prediction based on recursive least squares, S5 feedforward compensation amount calculation and advance driving, S6 depth feedback online correction, S7 multi-buoy cooperative depth setting, S8 control effect evaluation and parameter self-adaptive adjustment.The application realizes the rapid inhibition of high-frequency wave disturbance and the accurate elimination of static error through the parallel superposition and time scale cooperation of feedforward control and feedback control;Through multi-buoy array cooperative control and full-link parameter self-adaptive adjustment, the consistency of each seismic source depth and full-sea-state adaptive control are realized, which significantly improves the consistency of marine seismic exploration source excitation and data quality.

Description

Technical Field

[0001] This invention belongs to the field of marine seismic exploration technology, specifically relating to an active depth determination and heave compensation method for marine seismic buoys. Background Technology

[0002] In marine seismic exploration, air gun seismic sources are suspended underwater at a predetermined depth by a towed buoy to generate seismic waves. The depth of the seismic source directly affects the morphology of the seismic wavelet and the suppression effect of virtual reflections, thus determining the signal-to-noise ratio and imaging quality of the raw data. As exploration progresses into deeper and more complex geological structures, the requirements for the consistency of seismic source excitation are increasing, which places higher demands on the accuracy of buoy depth control and its resistance to wave disturbances.

[0003] The basic design of existing adjustable depth buoys involves a control system and a lifting device (including a gearbox, sensors, and a line winder) installed within the buoy body. The depth sensor monitors the current depth in real time, and the control system compares the measured depth with a preset target depth before driving the lifting device to raise or lower the cable or adjust buoyancy until the set depth is reached. Some existing solutions achieve constant depth control through a combination of multiple buoy systems; buoys are commonly equipped with IMU (Incremental Measurement Unit) accelerometers for attitude monitoring and wave statistics; and in the field of heave compensation, devices based on MRU (Medium-Range Unit) sensors for predicting the heave motion of the buoy are already available.

[0004] However, the existing technologies have the following shortcomings: First, they employ a pure feedback mode of "setting, measuring, comparing, and adjusting," where control actions only begin after depth deviation occurs, resulting in inherent lag for high-frequency wave disturbances with periods of 2-5 seconds, and failing to effectively suppress instantaneous depth fluctuations. Second, existing solutions are geared towards marine environmental monitoring scenarios, with accuracy only at the meter level, while seismic detection requires centimeter-level accuracy, representing a difference of orders of magnitude. Third, in multi-buoy arrays, the waves experienced by each buoy have phase differences, and independent depth determination cannot guarantee that all seismic sources are in the same depth plane, directly affecting the consistency of seismic data. Fourth, accelerometer data is only used for monitoring and recording and is not integrated into the control loop, thus failing to fully exploit the sensor's value. More fundamentally, these shortcomings are interconnected: feedback lag limits accuracy improvement, independent control sacrifices array consistency, and the lack of data closure makes feedforward compensation lack a foundation. To address these issues, this invention constructs a four-in-one composite depth determination control system integrating "wave disturbance feedforward, depth feedback, multi-buoy collaboration, and parameter adaptation," simultaneously solving multiple interrelated technical problems within the same framework. Summary of the Invention

[0005] The present invention aims to solve at least the above-mentioned technical problems existing in the depth control of existing marine seismic buoys, and to provide an active depth control and heave compensation method for marine seismic buoys.

[0006] The system upon which the method of the present invention is based includes: a sensor module consisting of an accelerometer IMU, a depth sensor, and a wind speed and direction sensor; a controller consisting of a signal processing module, a prediction module, a feedforward controller, a feedback controller, a cooperative controller, a parameter adaptive module, and a data storage unit; a lifting device actuator consisting of a buoyancy adjustment mechanism or a cable retraction mechanism; and an underwater acoustic communication module and a radio communication module.

[0007] The aforementioned sensor module, controller, lifting device actuator, and communication module are all existing technologies. Their specific hardware selection and implementation methods are common knowledge in the field. This invention does not involve any structural or functional improvements to the hardware of the aforementioned modules. The innovation of this invention lies in the active depth-keeping and heave compensation control method carried by the aforementioned modules.

[0008] This invention provides an active depth determination and heave compensation method for marine seismic exploration buoys, which constructs a four-in-one composite depth determination control system integrating "wave disturbance feedforward + depth feedback + multi-buoy collaboration + parameter adaptation". The method includes the following steps:

[0009] S1: System initialization and parameter configuration. Load target depth, actuator parameters, and control algorithm parameters, establish a data buffer, and provide initial operating conditions for subsequent real-time control.

[0010] S2: Wave disturbance sensing and acceleration signal acquisition. The buoy's heave motion is sensed in real time via an IMU (Integrated Velocity Measurement Unit). After coordinate transformation and subtraction of the gravity component, the heave acceleration is obtained. At the same time, marine environmental parameters are collected to provide real-time wave disturbance sensing data for feedforward control.

[0011] S3: Heave Acceleration Signal Preprocessing and Heave Motion State Estimation. The heave acceleration is sequentially subjected to bandpass filtering, second-order numerical integration, and high-pass filtering to eliminate drift, extracting a clean wave heave displacement signal to provide accurate heave motion state estimates for the prediction module.

[0012] S4: Heave displacement prediction based on recursive least squares. Using historical heave displacement data, the recursive least squares method is used to identify the coefficients of the prediction model online, predict the heave displacement in the future preset time domain, and evaluate the prediction confidence level, providing quantitative prediction input for advance compensation.

[0013] S5: Feedforward compensation calculation and advance drive. The feedforward compensation control quantity is calculated based on the predicted heave displacement and output to the actuator in advance for reverse compensation before the actual arrival of wave disturbance, realizing "compensation before disturbance".

[0014] S6: Online depth feedback correction. The depth of the seismic source in the water is measured by a depth sensor, and the feedback compensation is calculated using a PID control algorithm. This compensation is then added to the feedforward compensation to form the total control quantity, eliminating prediction model errors and cumulative drift, and ensuring long-term depth-keeping accuracy.

[0015] S7: Multi-buoy collaborative depth determination. In a multi-buoy array, the predicted displacement and prediction confidence of each buoy are shared through underwater acoustic communication; the main buoy calculates the average prediction deviation of the array, and the weighted average prediction deviation is obtained by weighted averaging based on the prediction confidence of each buoy; each buoy calculates the collaborative feedforward compensation amount according to a unified control law that integrates absolute depth control terms and synchronization control terms; the depth difference between each buoy is monitored in real time, and an alarm is triggered when the difference exceeds the limit.

[0016] S8: Control effect evaluation and adaptive parameter adjustment. Real-time evaluation of the depth-controlled effect, and coordinated adjustment of all parameters in the signal processing, prediction, feedforward, and feedback links according to sea state level to achieve adaptive optimal control under all sea state conditions.

[0017] Preferably, in step S1:

[0018] The S11 controller loads a preset target depth value and lifting device actuator parameters. The target depth value is determined by the operation plan, and the actuator parameters include response time constant, maximum stroke, maximum speed, and gain characteristics.

[0019] The S12 controller loads signal processing parameters, including the passband frequency range of the bandpass filter, the order of the prediction model, the prediction time domain, the feedforward gain coefficient, the PID feedback control parameters, and the synchronization gain coefficient. The feedforward gain coefficient is pre-calibrated through system identification experiments based on the transfer function characteristics of the lifting device, and the prediction time domain is matched with the response time constant of the actuator.

[0020] The S13 controller performs self-tests and communication handshakes on each sensor and communication module to confirm that each module is working properly.

[0021] The S14 controller initialization data buffer is used to store historical data on heave acceleration, heave displacement, and multi-buoy coordination.

[0022] Further, in step S2:

[0023] S21 Coordinate Transformation: Since the buoy may experience roll and pitch during towing, and the accelerometer measurements are located in the carrier coordinate system, a coordinate transformation matrix is ​​required. Transform it to the navigation coordinate system. Let the buoy's roll angle be... The pitch angle is , heading angle is The attitude transformation matrix is ​​provided in real time by the IMU's attitude calculation module, and is then:

[0024]

[0025] The vertical acceleration in the navigation coordinate system is:

[0026]

[0027] in , , These are the triaxial acceleration measurements in the carrier coordinate system. The acceleration due to gravity; the vertical acceleration This refers to the heave acceleration mentioned in step S3. .

[0028] S22 Environmental Parameter Auxiliary Acquisition: Simultaneously acquire environmental parameters such as wind speed, wind direction, significant wave height, and spectral peak period for subsequent sea state classification.

[0029] Further, in step S3:

[0030] S31 Bandpass Filter: A bandpass filter is used for... After filtering, the lower passband frequency corresponds to the lower limit of the long-period wave frequency, and the upper passband frequency corresponds to the upper limit of the short-period wave frequency. Bandpass filtering removes high-frequency vibration noise and low-frequency zero-drift components, resulting in a clean wave heave acceleration signal. .

[0031] S32 numerical integration to calculate heave velocity: This involves applying the filtered acceleration signal... The real-time heave velocity of the buoy was obtained by numerical integration using the trapezoidal integral method. :

[0032]

[0033] in The sampling interval is... This represents the current sampling time.

[0034] S33 Numerical integration to calculate heave displacement: For velocity signals... The trapezoidal integral method was used again for numerical integration to obtain the real-time heave displacement of the buoy. :

[0035]

[0036] S34 High-Pass Filter Eliminates Integral Drift: A high-pass filter is used to eliminate integral drift. Processing is performed to eliminate the DC drift component generated by long-term integration:

[0037]

[0038] After high-pass filtering, a drift-free heave displacement signal is obtained. .

[0039] S35 State Estimation Update: Adjust the current heave / sag displacement. Store in historical data cache It executes in a preset cycle.

[0040] Further, in step S4:

[0041] S41 Prediction Model Construction: Heave displacement is described by a third-order linear model in the short time domain:

[0042]

[0043] in These are the model coefficients. To predict residuals.

[0044] S42 Model Coefficient Identification: Using data points from the historical data cache, the model coefficients are identified online using the recursive least squares method. The recursive formula is as follows:

[0045]

[0046] in For the model coefficient vector, For the regression vector, Let covariance matrix be the variance matrix. It is a forgetting factor.

[0047] S43 Multi-Step Prediction: Based on the model coefficients identified at the current time step The heave and displacement over a predetermined time domain are predicted. The number of prediction steps is determined by the ratio of the prediction time domain to the sampling interval. The recursive formula for multi-step prediction is:

[0048]

[0049] in The unknown historical values ​​are replaced by previous predicted values, ultimately yielding... .

[0050] S44 Prediction Confidence Assessment: Calculate the variance of the prediction residuals:

[0051]

[0052] As a quantitative indicator of prediction confidence, when the prediction residual variance exceeds a preset threshold, the feedforward gain coefficient is reduced, and feedback control is prioritized to ensure system stability.

[0053] Specifically, in step S5:

[0054] S51 Feedforward Compensation Calculation: The formula for calculating the feedforward compensation control quantity is as follows:

[0055]

[0056] in The value of the feedforward gain coefficient is determined in advance through system identification experiments based on the transfer function characteristics of the lifting device, and the predicted heave displacement is converted into the control command quantity of the lifting device actuator.

[0057] S52 control quantity limiting: To prevent the actuator from exceeding its physical limits, Amplitude limiting is applied:

[0058]

[0059] in and These are the maximum and minimum control command limits for the actuator, determined by the physical characteristics of the lifting device.

[0060] S53 advance drive output: The feedforward compensation amount after amplitude limiting is output to the actuator of the lifting device in advance before the actual arrival of wave disturbance. The actuator drives the buoyancy adjustment mechanism or cable release mechanism to perform reverse motion compensation according to the control command: when it is predicted that the buoy will move upward, the lifting device is controlled to increase the buoy load or release the cable, so that the seismic source moves downward relative to the buoy; when it is predicted that the buoy will move downward, the lifting device is controlled to decrease the buoy load or retract the cable, so that the seismic source moves upward relative to the buoy; the compensation amplitude is proportional to the predicted heave displacement.

[0061] S54 Actuator Status Monitoring: During actuator operation, the system monitors feedback signals in real time via the actuator's built-in encoder or pressure sensor to confirm that the actuator is acting according to the expected instructions. If an abnormal actuator response is detected, a safety protection mechanism is immediately triggered.

[0062] Specifically, in step S6:

[0063] S61 Depth Deviation Calculation: Calculate the depth deviation at the current moment:

[0064]

[0065] S62 PID feedback control quantity calculation: The feedback compensation quantity is calculated using the positional PID control algorithm.

[0066]

[0067] in , , These are the proportional, integral, and differential gain coefficients, respectively. For feedback control cycle, This is the sequence number of the current feedback control cycle.

[0068] S63 Integral Separation and Anti-Saturation Processing: An integral separation strategy is adopted. When the absolute value of the depth deviation is greater than a preset threshold, the integral gain coefficient is forced to zero, and only proportional and derivative control is used; when the absolute value of the depth deviation is not greater than the preset threshold, integral control is enabled to eliminate static error. At the same time, the integral term is limited to prevent overshoot and oscillation caused by integral saturation.

[0069] S64 Total Control Quantity Synthesis: The feedforward control quantity and the feedback control quantity are superimposed to obtain the total control quantity.

[0070]

[0071] The total control quantity is output to the actuator after being limited.

[0072] S65 Dual Time Scale Control: Feedforward control is executed in the first preset cycle, and feedback control is executed in the second preset cycle. The feedforward control cycle is shorter than the feedback control cycle. The two run in parallel and the control quantities are superimposed for output.

[0073] Three-time-scale coordination mechanism: The feedforward control cycle, the feedback control cycle, and the multi-buoy cooperative communication cycle in step S71 constitute a three-time-scale coordination architecture. The three satisfy a progressive relationship where the feedforward control cycle is shorter than the feedback control cycle, and the feedback control cycle is shorter than the cooperative communication cycle.

[0074] Further, in step S7:

[0075] S71 data sharing among buoys: Each buoy sends its own predicted heave displacement, actual depth, and prediction confidence level to the main buoy via underwater acoustic communication in a coordinated communication cycle.

[0076] S72 array average prediction deviation calculation: The main buoy collects prediction data from all buoys and calculates the array's average predicted depth deviation.

[0077]

[0078] in This represents the number of buoys.

[0079] Step S73: Calculation of the coordinated compensation amount for each buoy:

[0080] First, a weighted average prediction bias is obtained by averaging the prediction confidence levels of each buoy:

[0081] Then, the coordinated feedforward compensation for each buoy is:

[0082] The first term is the absolute depth feedforward compensation for each buoy, and the second term is the synchronization control term. This is the synchronization gain coefficient; For the first The weight of each buoy, For the first The variance of the prediction residuals for each buoy.

[0083] S74 Synchronization Error Monitoring and Alarm: Real-time monitoring of the depth difference between each buoy. An alarm is triggered when the depth difference between any two buoys exceeds a preset threshold and continues for a preset duration.

[0084] Further, in step S8:

[0085] S81 control effect evaluation: In each feedback control cycle, calculate the root mean square error of depth, peak depth error, and standard deviation of depth error.

[0086] S82 Sea State Level Determination: Based on the wave spectrum analysis results obtained in step S22 and the standard deviation of heave displacement obtained in step S35, the current sea state level is automatically determined.

[0087] S83 Control Parameter Adaptive Adjustment: Based on sea state level and control effect evaluation results, automatically adjust four control parameters: prediction time domain, feedforward gain coefficient, PID parameters, and bandpass filter passband range. These four adjustments are coordinated and linked across steps S3 to S6 based on the same sea state determination result and the same set of control effect evaluation indicators. The adjusted parameters take effect simultaneously within the same control cycle.

[0088] S84 control log recording and data storage: Records key data throughout the entire operation process to the local storage medium at a frequency no less than a preset frequency.

[0089] The aforementioned S2 to S8 constitute an inseparable and complete closed-loop control system. Specifically: the acceleration acquisition in S2 provides the data source for S3; the displacement estimation result of S3 provides the sole predictive input for S4; the prediction result of S4 directly determines the effectiveness of the advance compensation in S5; S5 and S6 work synergistically in a parallel superposition manner to jointly constitute the complete depth-fixed control of a single buoy. They complement each other in terms of control objectives: S5 is responsible for "fast" disturbance suppression, while S6 is responsible for "accurate" error elimination, coordinating on a time scale, and superimposing control quantities; S7 introduces an array coordination dimension based on the aforementioned single buoy prediction and compensation, with the synchronous control term superimposed with incremental corrections on the feedforward compensation of S5, expanding the control objective from "absolute depth of a single buoy" to "relative depth consistency of the array"; the adaptive adjustment of S8 is simultaneously fed back to S3, S4, S5, and S6, enabling the parameters of each stage—signal processing, prediction, feedforward, and feedback—to be collaboratively optimized under all sea state conditions. The above steps are interconnected by multiple coupling relationships, including data dependence, complementary control, timing coordination, and parameter feedback. The absence or performance degradation of any link will lead to a significant decrease in the overall control effect.

[0090] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0091] (1) Balancing control accuracy and response speed. Feedforward control executes with a short cycle to suppress high-frequency wave disturbances in advance; feedback control executes with a longer cycle to eliminate prediction model errors and sensor drift. After the two operate in parallel with a specific time scale ratio and are superimposed, for the first time in buoy depth control, both rapid suppression of high-frequency disturbances and precise elimination of static errors are achieved simultaneously, avoiding the technical contradiction of "fast response and high accuracy cannot be achieved at the same time" in traditional schemes. The typical accuracy of existing pure feedback control schemes is at the meter level, while this invention can improve the control accuracy to the centimeter level, and the response mode changes from "lag compensation" to "advance compensation".

[0092] (2) Unity of absolute depth and relative consistency. The fusion control law includes both absolute depth control and synchronization control terms in the same formula. The former makes each buoy approach the target depth, while the latter eliminates the relative deviation of each buoy caused by wave phase difference. A single absolute depth control cannot eliminate the relative deviation within the array, and a single synchronization control cannot guarantee that the overall average depth does not deviate from the target. After the two are algebraically superimposed in the same control law, the simultaneous achievement of "absolute accuracy" and "relative consistency" is realized. The mutual deviation of the depths of each source in a multi-buoy array can be reduced from the meter level to the centimeter level.

[0093] (3) All-sea-state adaptive capability. Based on the same sea state determination result and the same set of control effect evaluation indicators, the parameters of the entire link of signal processing, prediction, feedforward, and feedback are adjusted in a coordinated manner, and the adjusted parameters take effect simultaneously within the same control cycle. Independent parameter adjustment of a single link cannot guarantee the coordinated matching between links. The coordinated linkage of the entire link enables each link to achieve optimal matching simultaneously under all sea state conditions, which is suitable for the operational needs under different sea areas and different towing speed conditions.

[0094] (4) Systemic Synergistic Effect. S2 to S8 of this invention constitute a complete closed-loop system. Data link coupling makes S3 and S4 mutually dependent, with the former providing the only input for the latter. Only when the two are combined can they simultaneously possess prediction accuracy and advance compensation capability. Control quantity coupling makes S5 and S6 indispensable. Only when the two are matched in parallel at a specific time scale and superimposed on the output can they simultaneously achieve rapid suppression of high-frequency disturbances and precise elimination of static errors. Cooperative dimension coupling makes absolute depth control and synchronous control algebraically superimposed in the same control law, so that they can simultaneously achieve the dual goals of each buoy approaching the target depth and maintaining depth consistency within the array. Adaptive coupling makes all link parameters take effect simultaneously within the same control cycle, so that each link can simultaneously achieve optimal matching. The four couplings work together to make the overall control method exhibit performance that no single technology or simple combination can achieve, realizing a systemic technical effect that balances "fast" and "accurate", unifies absolute depth and relative consistency, and adapts to all sea states. Attached Figure Description

[0095] Figure 1 This is a control flowchart of the active depth determination and heave compensation method for marine seismic buoys of the present invention;

[0096] Figure 2 This is a timing diagram of the dual-timescale and three-timescale collaborative mechanism of the present invention;

[0097] Figure 3 This is a schematic diagram of the heave displacement prediction algorithm of the present invention;

[0098] Figure 4 This is a block diagram illustrating the feedforward and feedback composite control principle of the present invention.

[0099] Figure 5 This is a flowchart of the multi-buoy collaborative depth control process of the present invention;

[0100] Figure 6 This is the buoy status overview interface for the active depth determination and heave compensation control system of marine seismic buoys.

[0101] Figure 7 This is the heave displacement prediction interface for the active depth determination and heave compensation control system of marine seismic buoys.

[0102] The accompanying figure shows the reference numerals, with S1 to S8 corresponding to the eight steps in the method flow. , , , These represent the acceleration, velocity, displacement, and predicted displacement of the heave motion, respectively. , , These represent the target depth, actual depth, and depth deviation, respectively. , , These represent the feedforward compensation, feedback compensation, and total control quantity, respectively. , , These symbols represent the feedforward control cycle, the feedback control cycle, and the cooperative communication cycle, respectively. Each symbol is defined in detail in the specific embodiments, and those skilled in the art, in conjunction with the accompanying drawings and specific embodiments, can clearly understand the meaning of each symbol. Detailed Implementation

[0103] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0104] Method Implementation Examples

[0105] This embodiment provides an active depth determination and heave compensation method for marine seismic exploration buoys, which constructs a four-in-one composite depth determination control system of "wave disturbance feedforward + depth feedback + multi-buoy collaboration + parameter self-adaptation".

[0106] The system upon which the method described in this embodiment relies includes a sensor module, a controller, a lifting device actuator, and a communication module.

[0107] The sensor module consists of an accelerometer (IMU), a depth sensor, and a wind speed and direction sensor, which are used to collect the buoy's triaxial acceleration and attitude data in real time, monitor the actual water depth of the seismic source, and collect wind speed and direction data, respectively.

[0108] The controller consists of a signal processing module, a prediction module, a feedforward controller, a feedback controller, a cooperative controller, a parameter adaptive module, and a data storage unit, which are used for acceleration signal preprocessing, heave displacement prediction, feedforward compensation calculation, feedback compensation calculation, multi-buoy cooperative control, adaptive adjustment of control parameters, and key data recording, respectively.

[0109] The lifting device's actuator is either a buoyancy adjustment mechanism or a cable retraction mechanism, used to execute the controller's depth adjustment commands.

[0110] The communication module includes an underwater acoustic communication module and a radio communication module, which are used for data sharing between buoys in the multi-buoy array and data exchange with the shipboard central controller or shore base station.

[0111] The aforementioned sensor module, controller, lifting device actuator, and communication module are all existing technologies. Their specific hardware selection and implementation methods are common knowledge in the field. This invention does not involve any structural or functional improvements to the hardware of the aforementioned modules. The innovation of this invention lies in the active depth-keeping and heave compensation control method carried by the aforementioned modules.

[0112] like Figure 1 As shown, the method described in this embodiment includes the following steps:

[0113] S1: System Initialization and Parameter Configuration

[0114] The buoy control system executes the initialization process after power-on.

[0115] The S11 controller loads a preset target depth value. The target depth value is determined by the operation plan, typically ranging from 5 to 30 meters, and the actuator parameters include the response time constant. Maximum travel, maximum speed, and gain characteristics.

[0116] The S12 controller loads signal processing parameters, including the passband frequency range of the bandpass filter and the order of the prediction model. Typical values ​​range from 50 to 200, prediction time domain Typical values ​​range from 0.5 to 2.0 s, feedforward gain coefficient PID feedback control parameters , , Synchronous gain coefficient Among them, the feedforward gain coefficient Based on the transfer function characteristics of the lifting device, it is pre-calibrated through a system identification experiment; prediction in the time domain. With the actuator response time constant satisfy .

[0117] The S13 controller performs self-tests and communication handshakes on each sensor and communication module to confirm that each module is working properly.

[0118] The S14 controller initialization data buffer is used to store historical data on heave acceleration, heave displacement, and multi-buoy coordination.

[0119] Preferably, in step S12, the prediction time domain The value is determined based on the actuator response time constant. Adaptive setting: when the actuator responds quickly, a smaller value is selected. A larger value is selected when the actuator responds slowly. The value ensures that the prediction time domain matches the action time of the actuator.

[0120] Furthermore, the passband frequency range of the bandpass filter is determined based on the typical wave period of the operating sea area, with a lower passband limit. The upper limit of the passband corresponds to the longest wave period. Corresponding to the shortest wave cycle.

[0121] Specifically, the feedforward gain coefficient The calibration method is as follows: Apply a control command of known amplitude to the lifting device actuator, measure the actual response displacement of the actuator, calculate the transfer function gain between the control command and the response displacement, and use the reciprocal of this gain as... The initial value is then optimized through online fine-tuning.

[0122] like Figure 6 As shown, after the system initialization is completed, the controller displays the operating status of each buoy in real time through the human-machine interface. This interface provides users with global visual monitoring of the buoy array status, making it easy for operators to quickly grasp the depth control effect, communication status and system health of each buoy.

[0123] S2: Wave Disturbance Sensing and Acceleration Signal Acquisition

[0124] During the towing process, the controller operates at a sampling frequency. The Hz sensor continuously acquires triaxial acceleration data from the accelerometer IMU. Let the measured triaxial acceleration values ​​in the buoy carrier coordinate system be... , , .

[0125] S21 Coordinate Transformation: Since the buoy may experience roll and pitch during towing, and the accelerometer measurements are located in the carrier coordinate system, a coordinate transformation matrix is ​​required. Transform it to the navigation coordinate system. Let the buoy's roll angle be... The pitch angle is , heading angle is The attitude transformation matrix is ​​provided in real time by the IMU's attitude calculation module, and is then:

[0126]

[0127] The vertical acceleration in the navigation coordinate system is:

[0128]

[0129] in , , These are the triaxial acceleration measurements in the carrier coordinate system. The acceleration due to gravity; the vertical acceleration This refers to the heave acceleration mentioned in step S3. .

[0130] S22 Environmental Parameter Auxiliary Data Acquisition: Synchronous Acquisition of Wind Speed Wind direction, significant wave height Spectral peak period Environmental parameters, such as these, are used for subsequent sea state classification.

[0131] Preferably, the sampling frequency The system adaptively adjusts based on the wave characteristics of the operating area, when the spectral peak period... When the sampling frequency is small, increase the sampling frequency to ensure effective acquisition of short-period waves. When the sample size is large, the sampling frequency can be appropriately reduced to decrease the amount of data.

[0132] Furthermore, the attitude transformation matrix Based on the roll angle output in real time by the IMU at each sampling time Pitch angle Heading angle Recalculate to ensure the real-time performance and accuracy of coordinate transformation.

[0133] Specifically, the heave acceleration In the calculation, gravitational acceleration The local gravity value is taken, and when the buoy is tilted, the components of gravitational acceleration on each axis are automatically subtracted through the attitude transformation matrix.

[0134] S3: Heave Acceleration Signal Preprocessing and Heave Motion State Estimation

[0135] The controller's signal processing module processes the heave acceleration signal acquired in step S2. Perform digital signal processing. For example... Figure 3 As shown, the signal processing flow for this step is as follows: heave acceleration After bandpass filtering, velocity calculation using trapezoidal integration, displacement calculation using trapezoidal integration, and high-pass filtering to eliminate drift, a drift-free heave-displacement signal is obtained. .

[0136] S31 Bandpass Filter: A bandpass filter is used for... Filtering is performed to reduce the lower passband frequency. Hz corresponds to the lower limit of long-period wave frequency and the upper limit of passband frequency. Hz corresponds to the upper limit of short-period wave frequency. A bandpass filter removes high-frequency vibration noise and low-frequency zero-drift components, yielding a clean wave heave acceleration signal. Among them, high-frequency vibration noise is mainly caused by cable jitter, air gun excitation, etc., while low-frequency zero drift component is mainly caused by sensor temperature drift, integral drift, etc. Bandpass filtering of acceleration signal is a key preprocessing step for heave motion measurement.

[0137] S32 numerical integration to calculate heave velocity: This involves applying the filtered acceleration signal... The real-time heave velocity of the buoy was obtained by numerical integration using the trapezoidal integral method. :

[0138]

[0139] in The sampling interval is... ms, This represents the current sampling time.

[0140] S33 Numerical integration to calculate heave displacement: For velocity signals... The trapezoidal integral method was used again for numerical integration to obtain the real-time heave displacement of the buoy. :

[0141]

[0142] S34 high-pass filter eliminates integral drift: using cutoff frequency Hz high-pass filter Processing is performed to eliminate the DC drift component generated by long-term integration:

[0143]

[0144] in After high-pass filtering, a drift-free heave displacement signal is obtained. .

[0145] S35 State Estimation Update: Adjust the current heave / sag displacement. Store in historical data cache The entire state estimation process is executed in a loop with a period of 10ms.

[0146] Preferably, the bandpass filter is a fourth-order Butterworth filter, which has the maximum flat amplitude-frequency response characteristic in the passband, can effectively retain the acceleration signal energy in the wave frequency band, and has good stopband attenuation characteristics.

[0147] Furthermore, the initial value of the integral in the trapezoidal integral method... and The system is initialized to zero upon startup, and the integral result tends to stabilize after about 2 to 3 wave cycles.

[0148] Specifically, the cutoff frequency of the high-pass filter The Hz is much lower than the lower limit of the wave passband by 0.05Hz, ensuring that effective wave displacement signal components are not lost while eliminating integral drift.

[0149] S4: Heave displacement prediction based on recursive least squares

[0150] The controller's prediction module uses the historical heave displacement data obtained in step S3. A recursive least squares linear prediction model is used to predict the future. heave displacement at any moment .like Figure 3 As shown, after the historical displacement data is identified by the RLS model coefficients and predicted in multiple steps, the predicted values ​​of future heave and subsidence displacement are output.

[0151] S41 Prediction Model Construction: Heave displacement is described by a third-order linear model in the short time domain. :

[0152]

[0153] in These are the model coefficients. To predict residuals.

[0154] S42 model coefficient identification: utilizing historical data cache Data points, For a historical window of 0.5 to 2 seconds, the recursive least squares method is used to identify the model coefficients online. The recursive formula is as follows:

[0155]

[0156] in For the model coefficient vector, For the regression vector, Let covariance matrix be the variance matrix. Forgetting factor is used to assign higher weight to recent data in order to adapt to the time-varying nature of sea state.

[0157] S43 Multi-Step Prediction: Based on the Current Moment Identified model coefficients Regarding the future The heave displacement at time t is predicted. Let the number of prediction steps be . , s, The recursive formula for multi-step prediction is:

[0158]

[0159] in The unknown historical values ​​are derived from previous predicted values. Replacement. Ultimately obtained. .

[0160] S44 Prediction Confidence Assessment: Calculate the variance of the prediction residuals:

[0161]

[0162] As a quantitative indicator of prediction confidence. When the prediction residual variance When the sea state exceeds a preset threshold, it indicates a drastic change in sea state, causing the prediction model to temporarily fail and reducing the feedforward gain coefficient. Feedback control is prioritized to ensure system stability.

[0163] Preferably, the forgetting factor The value is adaptively adjusted based on the severity of sea state changes: a smaller value is taken during high sea states and periods of drastic wave changes, such as... This allows the algorithm to track wave changes faster; larger values ​​are used in low sea states and calm waves, such as... This ensures stable identification.

[0164] Furthermore, in the multi-step prediction, when the prediction steps... In each step of the prediction Unknown historical values ​​within Equal to the predicted value from the previous step Replacement, to achieve recursive prediction.

[0165] Specifically, the predicted residual variance The threshold is determined based on historical prediction error statistics, with a typical value being [value missing]. ,in This represents the variance of the heave displacement.

[0166] like Figure 7 As shown, during the execution of the heave displacement prediction step, the controller displays the prediction algorithm's running status and prediction results in real time through the human-machine interface. This interface provides users with visual monitoring of the RLS prediction algorithm's running status, making it easy for operators to view the predicted displacement, prediction confidence, historical displacement comparison, and model coefficient change trends in real time, and to evaluate the convergence and prediction accuracy of the prediction algorithm in a timely manner.

[0167] S5: Feedforward Compensation Calculation and Advance Driving

[0168] The feedforward controller of the controller predicts the future heave displacement based on step S4. Calculate the feedforward compensation control quantity and drive the lifting device actuator in advance. For example... Figure 4 As shown, the predicted displacement generates a feedforward compensation amount via a feedforward controller. , and feedback control quantity The superimposed output is sent to the actuator.

[0169] S51 Feedforward Compensation Calculation: The formula for calculating the feedforward compensation control quantity is as follows:

[0170]

[0171] in The feedforward gain coefficient is determined based on the transfer function characteristics of the lifting device, including gain, bandwidth, and phase lag. It is pre-calibrated through system identification experiments to convert the predicted heave displacement into control commands for the lifting device's actuators, such as motor speed and valve opening.

[0172] S52 control quantity limiting: To prevent the actuator from exceeding its physical limits, Amplitude limiting is applied:

[0173]

[0174] in and These are the maximum and minimum control command limits for the actuator, determined by the physical characteristics of the lifting device, namely maximum speed, maximum acceleration, and maximum stroke.

[0175] S53 advance drive output: This output adjusts the feedforward compensation amount after limiting. Before the actual arrival of the wave disturbance, i.e. The time signal is output to the lifting device's actuator via the control signal output terminal. The actuator, based on the control commands, drives the buoyancy adjustment mechanism (e.g., changing the airbag volume, adjusting the inner and outer oil bladder volumes), or the cable retraction / deployment mechanism (e.g., the winch unit) to perform reverse motion compensation.

[0176] When it is predicted that the buoy will move upward ( When the vibration source is in motion, the control device increases the load on the buoy or releases the cable to make the vibration source move downward relative to the buoy.

[0177] When it is predicted that the buoy will move downwards ( When the vibration source is in motion, control the lifting device to reduce the buoy load or retract the cable to make the vibration source move upward relative to the buoy.

[0178] Compensation range and predicted heave displacement Proportional.

[0179] S54 Actuator Status Monitoring: During actuator operation, the controller monitors the actuator's feedback signals in real time through the actuator's built-in encoder or pressure sensor, including position, speed, and pressure, to confirm that the actuator is acting according to the expected instructions. If an abnormal actuator response is detected, such as jamming, overspeeding, or overload, the safety protection mechanism is immediately triggered, i.e., output is stopped and an alarm is triggered.

[0180] Preferably, the feedforward gain coefficient The value of is determined through the following system identification experiment: A series of sinusoidal control commands with different frequencies and amplitudes are applied to the actuator, the response amplitude and phase of the actuator are measured, the frequency response curve of the actuator is plotted, and the gain and bandwidth of the actuator are determined based on the frequency response curve. It is the reciprocal of the DC gain of the actuator.

[0181] Furthermore, the control quantity limit value and The setting takes into account the physical constraints and operational safety margins of the actuator, with a typical value of 80% of the actuator's maximum capacity.

[0182] Specifically, when the lifting device is a cable winding and unwinding mechanism, the feedforward compensation amount Converted into speed command for the winch motor, positive Corresponding to the cable laying direction, negative Corresponding to the cable reeling direction.

[0183] S6: Deep Feedback Online Correction

[0184] The controller's feedback controller monitors the actual water penetration depth of the seismic source in real time using a depth sensor. , and target depth A comparison was made, and a PID control algorithm was used to correct the feedforward compensation online. For example... Figure 4 As shown, depth deviation The PID feedback controller generates feedback compensation. This is superimposed on the feedforward compensation amount.

[0185] S61 Depth Deviation Calculation: Calculate the depth deviation at the current moment:

[0186]

[0187] S62 PID feedback control quantity calculation: The feedback compensation quantity is calculated using the positional PID control algorithm.

[0188]

[0189] in , , These are the proportional, integral, and differential gain coefficients, determined by Ziegler-Nichols tuning. For feedback control cycle, ms, This is the sequence number of the current feedback control cycle. The integral term is used to eliminate static errors, such as buoyancy deviation and mean depth shift caused by ocean currents, while the derivative term is used to suppress drastic fluctuations in the rate of depth change.

[0190] S63 Integral Separation and Anti-Saturation Treatment: Integral separation strategy is adopted.

[0191] when At that time, forced Using only proportional and derivative control, where, This is the integration separation threshold, typically 0.5m;

[0192] when When this occurs, integral control is activated to eliminate static error.

[0193] At the same time, the integral term is limited to prevent overshoot and oscillation caused by integral saturation.

[0194] S64 Total Control Quantity Synthesis: The feedforward control quantity and the feedback control quantity are superimposed to obtain the total control quantity.

[0195]

[0196] Total control quantity After being limited, the signal is output to the actuator via the control signal output terminal.

[0197] S65 Dual Time Scale Control: Feedforward Control with Execution is performed in ms cycles, with feedback control. The ms represents the periodic execution time; both processes run in parallel, and the control quantities are superimposed for output.

[0198] Three-timescale coordination mechanism: such as Figure 2 As shown, the feedforward control cycle ms, feedback control cycle ms and the multi-buoy cooperative communication cycle in step S71 The three timescales form a collaborative architecture, and they satisfy the following conditions: The progressive relationship. The fastest update frequency of feedforward control ensures a rapid response and advance compensation to high-frequency wave disturbances with a cycle of 2-5 seconds. The intermediate update frequency of feedback control ensures effective correction of accumulated errors while avoiding high-frequency oscillations introduced by measurement noise; the slowest update frequency of cooperative control ensures the stability of array synchronous adjustment and prevents mutual interference caused by excessive communication.

[0199] Preferably, the PID parameters , , The Ziegler-Nichols tuning method is as follows: First, let , gradually increase Record the critical gain until the system output exhibits constant amplitude oscillation. and critical oscillation period Then calculate the PID parameters using the following formula: , , .

[0200] Furthermore, the integral separation threshold The setting is based on the target depth. The percentage is determined, typically 5% of the target depth, but not exceeding 1m.

[0201] Specifically, the ratio of the feedforward control cycle to the feedback control cycle... It was determined through optimization: if If it is too short, the calculation burden will be too heavy. If the time is too long, it will be unable to respond to high-frequency disturbances in a timely manner; if If it is too short, it will easily introduce measurement noise. If the time frame is too long, feedback and corrections will not be timely.

[0202] S7: Multi-buoy collaborative depth determination

[0203] In a multi-buoy towed array Each buoy shares data in real time via an underwater acoustic communication module, and the controller's co-controller achieves coordinated depth control. For example... Figure 5 As shown, each buoy sends prediction data to the main buoy via underwater acoustic communication. After calculating the average prediction deviation of the array, the main buoy sends the cooperative compensation amount back to each buoy.

[0204] S71 Buoy Data Sharing: Data sharing among buoys ( (based on communication cycle) MS transmits its heave displacement prediction data to the main buoy or shipboard central controller via the underwater acoustic communication module. Actual depth and prediction confidence .

[0205] S72 array average prediction deviation calculation: The main buoy or shipboard central controller collects prediction data from all buoys and calculates the array's average predicted depth deviation.

[0206]

[0207] Calculation of coordinated compensation for each buoy in S73:

[0208] First, a weighted average prediction bias is obtained by averaging the prediction confidence levels of each buoy:

[0209] Then, the coordinated feedforward compensation for each buoy is:

[0210] The first term is the absolute depth feedforward compensation for each buoy, and the second term is the synchronization control term. This is the synchronization gain coefficient; For the first The weight of each buoy, For the first The variance of the prediction residuals for each buoy.

[0211] S74 Synchronization Error Monitoring and Alarm: The controller monitors the depth difference between each buoy in real time. ( When the depth difference between any two buoys exceeds a preset threshold. Typical value is 5cm and continues to exceed When the alarm is triggered, the operator is notified via the radio communication module.

[0212] Preferably, the synchronization gain coefficient The value ranges from 0.3 to 1.0. A larger value is used when the array consistency requirement is high, and a smaller value is used when it is necessary to prioritize the absolute depth accuracy of a single buoy.

[0213] Furthermore, the underwater acoustic communication module adopts a code division multiple access or time division multiple access communication system, with each buoy transmitting data on its own time slot or code channel to avoid communication conflicts.

[0214] Specifically, when a communication interruption of a buoy is detected or the confidence level is predicted... If the depth remains too high, the buoy will automatically exit the collaborative mode and switch to an independent single buoy depth control mode, while other buoys continue to work together.

[0215] S8: Control Effect Evaluation and Adaptive Parameter Adjustment

[0216] The controller's parameter adaptive module continuously evaluates the depth control effect and adaptively adjusts the control parameters according to changes in sea state.

[0217] S81 control effectiveness evaluation: in each feedback control cycle ms, calculate the following performance metrics:

[0218] Root mean square error of depth: ,in To evaluate window length , corresponding to 10 seconds;

[0219] Peak depth error: ;

[0220] Depth error standard deviation: ,in .

[0221] S82 Sea State Classification: Based on the wave spectrum analysis results obtained in step S22, the effective wave height is determined. Spectral peak period and the standard deviation of heave displacement obtained in step S35 It automatically determines the current sea state level.

[0222] S83 Control Parameter Adaptive Adjustment: Based on sea state level and control effectiveness evaluation results, the parameter adaptive module automatically adjusts the following control parameters:

[0223] Prediction Time Domain Reduced at high sea states Increased in low sea states After adjustment, the data is sent back to the prediction module to change the number of prediction steps. ;

[0224] Feedforward gain coefficient The lifting device is finely adjusted online based on its actual response, and the adjustment is then sent back to the feedforward controller to change the feedforward compensation amount.

[0225] PID parameters , , The gain scheduling strategy is adopted, and different combinations of PID parameters are switched according to the sea state level. After adjustment, the data is sent back to the feedback controller to change the feedback compensation amount.

[0226] The passband range of a bandpass filter is determined by the spectral peak period. The lower and upper frequency limits of the filter passband are adaptively adjusted, and the adjusted values ​​are fed back to the signal processing module to change the filter transfer function. .

[0227] The above four adjustments are not independent parameter modifications, but rather adjustments made by the parameter adaptive module based on the same sea state determination results. , and the same set of control effectiveness evaluation indicators , , This involves coordinated adjustments to modules S3 through S6. The adjusted parameters take effect simultaneously within the same control cycle, ensuring that the four stages—signal preprocessing, heave prediction, feedforward control, and feedback control—re-achieve optimal matching under the new sea state conditions.

[0228] S84 Control Log Recording and Data Storage: The controller's data storage unit records key data from the entire operation process to local storage media at a frequency of no less than 10Hz for post-operation analysis, evaluation, and algorithm optimization. Key data includes heave acceleration. Heave displacement Predicted displacement Depth deviation Control quantity Feedback signals from implementing agencies, etc.

[0229] Preferably, the sea state level determination adopts a fuzzy logic method to determine the effective wave height. Spectral peak period and standard deviation of heave displacement As a fuzzy input variable, the output is the sea state level.

[0230] Furthermore, in the gain scheduling strategy, multiple sets of PID parameter combinations for different sea state levels are pre-stored. When the sea state level changes, the new PID parameters are smoothly transitioned through linear interpolation to avoid system shocks caused by parameter mutations.

[0231] Specifically, in the adaptive adjustment of the passband range of the bandpass filter, the lower limit of the passband... Take the peak frequency of the spectrum 0.5 times that, passband limit Take twice the peak frequency of the spectrum to ensure that the filter passband always covers the main wave energy frequency band.

[0232] Example 1: Single buoy cable retraction and deployment method for depth determination

[0233] This embodiment provides an active depth-determination and heave compensation method for marine seismic buoys, applicable to single-buoy towed operation scenarios. For example... Figure 1 As shown, this embodiment executes steps S1 to S6 and S8, and does not enable S7 multi-buoy collaborative depth determination.

[0234] In this embodiment, the lifting device adopts a cable winding and unwinding mechanism.

[0235] In step S1, the controller loads a preset target depth value. m. Actuator parameters include: maximum winch speed. rpm, the maximum speed at which the cable is retracted or extended. m / s, response time constant s. In signal processing parameters, prediction time domain Set to 1.0s, which is the response time constant of the actuator. Matching; Feedforward gain coefficient The system was calibrated to 0.8 through identification experiments; the PID parameters were determined by Ziegler-Nichols tuning. , , After performing self-tests on each sensor and communication module, the controller initializes the data buffer.

[0236] In step S2, the accelerometer sampling frequency is set to Hz, sampling interval The IMU attitude calculation module provides the roll angle in real time. Pitch angle Heading angle Through the attitude transformation matrix Transform the vehicle coordinate system acceleration to the navigation coordinate system, and subtract the gravitational acceleration. The heave acceleration is obtained after m / s². Simultaneously acquire wind speed, wind direction, and significant wave height. Spectral peak period Environmental parameters, etc.

[0237] Specifically, if the IMU measures the roll angle at a certain moment Pitch angle Heading angle The triaxial acceleration measurement value is m / s² m / s² m / s².

[0238] First, convert the angle to radians: rad, rad, rad.

[0239] Calculate the attitude transformation matrix Elements:

[0240]

[0241]

[0242]

[0243] Vertical acceleration:

[0244] m / s²

[0245] Obtain heave acceleration The m / s² value is approximately zero, indicating that the buoy is near its equilibrium position at that moment.

[0246] In step S3, the passband range of the bandpass filter is set to 0.05~0.5Hz, corresponding to a wave period of 2~20s. The cutoff frequency of the high-pass filter is set to... Hz, rad / s. Trapezoidal integration method based on sampling interval ms are incremented sequentially.

[0247] Suppose the filtered acceleration is at a certain sampling time The value is m / s², the previous moment m / s², velocity at the previous moment m / s, then the velocity at the current moment:

[0248]

[0249] If the displacement at the previous moment m, then the displacement at the current moment:

[0250]

[0251] After high-pass filtering with a cutoff frequency of 0.01Hz, a drift-free heave displacement signal is obtained. m. State estimation is performed cyclically with a 10ms period, recording the current heave and displacement. Store in historical data cache ,in .

[0252] In step S4, the order of the prediction model is determined. Historical Window Forgetting factor corresponding to 1 second of historical data RLS algorithm initial conditions: , Prediction in the time domain s, number of predicted steps .

[0253] Assuming the regression vector at the current moment Model coefficients at the previous time step Then the predicted value is:

[0254]

[0255] RLS recursively updates the model coefficients at the current time step. Then, the heave displacement in the next 1.0 seconds is predicted through multi-step recursion. In multi-step prediction, the unknown historical values ​​in the regression vector at each step are replaced by the predicted values ​​from the previous step, and this process is repeated for 100 steps to obtain the result. .

[0256] Assuming the multi-step prediction result is m represents an upward trend. Simultaneously, the variance of the predicted residuals is calculated. If the value is below the preset threshold of 0.05, it indicates that the prediction confidence is good.

[0257] In step S5, the feedforward compensation amount is calculated as follows:

[0258]

[0259] Control limit value The corresponding maximum cable winding and unwinding speed is 0.5 m / s. Corresponding to -0.5m / s, Within the limited range, therefore The feedforward compensation is output to the winch unit 1.0s before the actual arrival of the wave disturbance. The winch performs the cable laying action at a speed of 0.12m / s, causing the seismic source to move downward relative to the buoy, compensating for the expected upward displacement.

[0260] During the operation of the actuator, the controller monitors the winch speed and cable length in real time through the encoder to confirm that the actuator operates according to the expected instructions.

[0261] In step S6, the depth sensor monitors the actual water depth in real time. Assuming the current actual depth is 9.85m, the target depth is... m, depth deviation:

[0262]

[0263] Feedback control cycle ms, let the deviation of the previous time step be... m, integral term Its current value is 0.005.

[0264] Feedback compensation amount:

[0265]

[0266] Integral Separation Threshold m, current m m, enable integral control. Total control quantity:

[0267]

[0268] Total control quantity Exceed Therefore, the amplitude limit is The maximum cable laying speed is 0.5 m / s. Feedforward control is executed with a 10 ms cycle, and feedback control is executed with a 50 ms cycle; the two operate in parallel.

[0269] In step S8, the window length is evaluated. This corresponds to 10 seconds. Assume a depth deviation sequence of the most recent 200 sampling points. Sum of squares m², then:

[0270]

[0271] Depth peak error m, standard deviation of depth error m. Based on significant wave height m, spectral peak period The system determines that the current sea state is moderate. The parameter adaptive module fine-tunes the control parameters and then sends the data back to each module. Key data is recorded to local storage media at a frequency of no less than 10Hz.

[0272] This embodiment demonstrates the effectiveness of the present invention in single-buoy towing operations. By employing a cable deployment and retrieval mechanism, and through the parallel superposition of feedforward and feedback control with time-scale coordination, the depth fluctuation amplitude is reduced from the meter level to the centimeter level under the same wave conditions, achieving rapid suppression of high-frequency wave disturbances and precise elimination of static errors.

[0273] Example 2: Multi-buoy array collaborative depth determination example

[0274] This embodiment provides an active depth determination and heave compensation method for marine seismic buoys, applicable to multi-buoy towed array operations. Figure 1 and Figure 5 As shown, this embodiment executes the complete steps S1 to S8.

[0275] In this embodiment, the number of buoys Each buoy is equipped with an accelerometer (IMU), a depth sensor, and an underwater acoustic communication module. The lifting device uses a cable retrieval mechanism.

[0276] In step S1, each buoy is synchronously loaded with the same target depth value. m. The actuator parameters for each buoy are: maximum winch speed. rpm, the maximum speed at which the cable is retracted or extended. m / s, response time constant s; Signal processing parameters: prediction time domain s, feedforward gain coefficient PID parameters , , Synchronous gain coefficient Each buoy completes its self-check and data buffer initialization.

[0277] In step S2, each buoy is sampled at a frequency Hz continuously collects acceleration data, and after attitude transformation and gravity compensation, the heave acceleration is obtained. Taking buoy 1 as an example:

[0278] Suppose that at a certain moment, the IMU of buoy 1 measures the roll angle. Pitch angle Heading angle The triaxial acceleration measurement value is m / s² m / s² m / s².

[0279] Convert angles to radians: rad, rad, rad.

[0280] Calculate the attitude transformation matrix of buoy 1 The third row of elements:

[0281]

[0282]

[0283]

[0284] Vertical acceleration in the navigation coordinate system of buoy 1:

[0285] m / s²

[0286] Simultaneously acquire environmental parameters at the location of each buoy. Significant wave height at buoy 1. m, spectral peak period s; 2 buoys m, s; 3 buoys m, The wave parameters differ at different locations in the array.

[0287] In step S3, the passband of the bandpass filter for each buoy is set to 0.05~0.5Hz. Taking buoy 1 as an example, let the acceleration of buoy 1 after filtering at a certain sampling time be... m / s², the previous moment m / s², velocity at the previous moment m / s, then the velocity at the current moment:

[0288] m / s

[0289] Displacement at the previous moment m, then the displacement at the current moment:

[0290] m

[0291] After high-pass filtering with a cutoff frequency of 0.01Hz rad / s, to obtain a drift-free heave displacement signal. m. Store the historical data in the buffer of buoy 1 at 10ms intervals.

[0292] Buoy 2 and buoy 3 obtained their respective heave displacements according to the same signal processing procedure. and Because the wave phases differ at the locations of each buoy, their heave and displacement vary at the same moment. Let's assume the current heave and displacement of buoy 2 is... m, which represents the downward displacement of buoy 3. m means upward.

[0293] In step S4, each buoy uses the RLS algorithm, and the model order is... Historical Window Forgetting factor Perform multi-step prediction. Taking buoy 1 as an example:

[0294] The regression vector of buoy 1 at the current moment Model coefficients at the previous time step .

[0295] Predicted value: m

[0296] After RLS recursive update, we get Multi-step prediction yielded the heave displacement of buoy 1 in the next 1.0 seconds. m. Predicted residual variance .

[0297] Similarly, the RLS prediction results for buoy 2 and buoy 3 are respectively m, downward trend, predicted residual variance ; m, upward trend, predicting residual variance .

[0298] The reason for the difference in the confidence levels of the predictions from the different buoys is that the wave changes at the location of buoy 3 are more drastic. m, The high-frequency components of the heave motion are more numerous, resulting in a relatively large tracking error for the RLS prediction model.

[0299] In step S5, each buoy independently calculates its feedforward compensation. Taking buoy 1 as an example: This corresponds to a cable deployment speed of 0.096 m / s. Buoy 2: The corresponding retrieval speed is 0.064 m / s. Buoy 3: The corresponding cable laying speed is 0.040 m / s.

[0300] In step S6, each buoy independently performs online depth feedback correction. Let the actual depth deviation of each buoy at a certain moment be... m, actual depth 9.95m, m, actual depth 10.03m, m, actual depth 9.98m, feedback control cycle ms, let the deviations of each buoy at the previous time be respectively m、 m、 m, with cumulative values ​​of the integral terms of 0.002, 0.001, and 0.0015 respectively.

[0301] Buoy 1 feedback compensation amount:

[0302] Buoy 2 feedback compensation amount:

[0303] Buoy 3 feedback compensation amount:

[0304] Integral Separation Threshold m, , , All buoys are now controlled by integral control.

[0305] In step S71, each buoy... The ms periodically transmits its respective data to the main buoy, i.e., buoy 1, via the underwater acoustic communication module. , as well as .

[0306] In step S72, the main buoy calculates the unweighted array average prediction bias:

[0307]

[0308] In step S73, the weighted weights are first calculated based on the prediction confidence levels of each buoy:

[0309]

[0310] Total weight:

[0311] Weighted average prediction bias:

[0312]

[0313] The weighted average prediction bias of about 7.76 cm is higher than the unweighted average of 3 cm because the high-confidence buoy 1 has a stronger upward trend.

[0314] Then, the coordinated feedforward compensation for each buoy is calculated:

[0315] Buoy 1:

[0316] Buoy 2:

[0317] Buoy 3:

[0318] The total control quantity for each buoy is the sum of its respective feedforward compensation, coordinated synchronization control term, and feedback compensation.

[0319] Total control quantity of buoy 1:

[0320] Total control quantity of buoy 2:

[0321] Total control quantity of buoy 3:

[0322] Control limit value , All three total control quantities mentioned above are within the limit range.

[0323] In step S74, the depth difference threshold cm, alarm duration The controller monitors the depth difference between each buoy in real time. Assuming the current... m has been running for 1 second without triggering an alarm; if it continues for more than 2 seconds, the operator will be notified via the radio communication module.

[0324] In step S8, each buoy independently performs control effect evaluation and parameter adaptive adjustment. Assume that the calculated parameters for each buoy are... The values ​​were 0.04m, 0.05m, and 0.045m, respectively, all below 0.06m, indicating good depth accuracy.

[0325] According to their respective and The adaptive adjustment results of the parameters of each buoy are slightly different:

[0326] Buoy 1 m, The sea state was determined to be moderate. The time was adjusted from 1.0s to 1.1s. The value was adjusted from 0.8 to 0.78.

[0327] Buoy 2 m, The sea state was determined to be moderate. Adjusted to 1.05s. Adjusted to 0.79;

[0328] Buoy 3 m, s is judged to be moderately hot. The time was adjusted from 1.0s to 0.9s. The value was adjusted from 0.8 to 0.82, and the passband of the bandpass filter was adjusted from 0.05~0.5Hz to 0.06~0.55Hz.

[0329] The adjusted parameters of each buoy are sent back to their respective S3, S4, S5, and S6 modules, and take effect simultaneously within the same control cycle.

[0330] This embodiment demonstrates the effectiveness of the present invention in multi-buoy towed array operation scenarios. By sharing the predicted displacement and prediction confidence of each buoy through underwater acoustic communication, calculating the average prediction deviation of the array and performing weighted processing, absolute depth control and relative consistency control are integrated into a unified control law. The mutual deviation of the depth of each buoy can be controlled within 5cm, ensuring that each seismic source is in the same depth plane.

[0331] Example 3: End-to-End Adaptive Depth Determination Implementation in Severe Sea State

[0332] This embodiment provides an active depth determination and heave compensation method for marine seismic detection buoys, which is applicable to operational scenarios under harsh sea conditions.

[0333] In this embodiment, the lifting device adopts a cable winding and unwinding mechanism.

[0334] In step S1, the controller loads the target depth value. m. The actuator parameters are: maximum winch speed. rpm, the maximum speed at which the cable is retracted or extended. m / s, response time constant The initial signal processing parameters are set as follows: prediction time domain. s, feedforward gain coefficient PID parameters , , The passband range of the bandpass filter is 0.05~0.5Hz. The controller completes self-test and data buffer initialization.

[0335] In step S2, the accelerometer... The IMU collects triaxial acceleration data at a sampling frequency of Hz. Let the current IMU measure the roll angle... Pitch angle Heading angle Triaxial acceleration m / s² m / s² m / s², with greater acceleration at high sea states. After attitude transformation matrix conversion:

[0336] rad, rad

[0337]

[0338]

[0339]

[0340] Vertical acceleration in the navigation coordinate system:

[0341] m / s²

[0342] Obtain heave acceleration m / s², a positive value indicates that the buoy has an upward acceleration.

[0343] Simultaneously acquire environmental parameters: wind speed m / s, approximately force 7 wind, significant wave height m, spectral peak period s.

[0344] In step S3, the initial bandpass filter passband is set to 0.05~0.5Hz. Perform filtering.

[0345] Suppose the acceleration after filtering at a certain sampling time m / s², the acceleration amplitude is larger in high sea states, the previous moment m / s², velocity at the previous moment m / s, sampling interval If ms, then the velocity at the current moment is:

[0346] m / s

[0347] Displacement at the previous moment m, then the displacement at the current moment:

[0348] m

[0349] After processing by a high-pass filter with a cutoff frequency of 0.01Hz, rad / s, to obtain a drift-free heave displacement signal. m is stored in the historical data cache area.

[0350] In step S4, the initial parameters of the RLS algorithm are... , , , Current regression vector Model coefficients at the previous time step Predicted value:

[0351] m

[0352] After RLS recursive update, we get Multi-step prediction At this point, the predicted residual variance... If the sea state exceeds the preset threshold of 0.05, it indicates drastic changes in sea state, causing the prediction model to temporarily fail and automatically reduce its speed. The value was reduced to 0.6, down from 0.8, prioritizing feedback control to ensure system stability.

[0353] In step S5, press Calculate the feedforward compensation and output it in advance. Control quantity limit value. The corresponding maximum cable winding and unwinding speed is 0.5 m / s. Corresponding to -0.5 m / s. The actuator performs reverse motion compensation according to the control command, and the controller monitors the actuator feedback signal in real time through the built-in encoder to confirm that the actuator acts according to the expected command.

[0354] In step S6, the depth sensor monitors the actual depth in real time. m, depth deviation m. Feedback control cycle ms, let the deviation of the previous time step be... m, cumulative integral value 0.008. Currently using the reduced... However, the PID parameters have not yet been switched, and the feedback compensation amount is still calculated based on the initial parameters:

[0355]

[0356] Integral Separation Threshold m, current m m, enable integral control. Total control quantity After amplitude limiting, the output is controlled by feedforward control with a 10ms cycle and feedback control with a 50ms cycle, and the two operate in parallel.

[0357] In step S82, based on the effective wave height m and spectral peak period s, determined to be high sea state, m or s.

[0358] In step S83, perform end-to-end coordinated adjustment:

[0359] Prediction Time Domain The time was reduced from 1.0s to 0.6s and sent back to S43, increasing the prediction steps. ;

[0360] Feedforward gain coefficient Adjusted from 0.8 to 0.9, and transmitted back to S51;

[0361] Switch the PID parameters to the high sea state gain combination: , , The data is then sent back to S62.

[0362] The passband range of the bandpass filter was adjusted from 0.05~0.5Hz to 0.08~0.8Hz and fed back to S31.

[0363] The above four adjustments will take effect simultaneously within the same control cycle.

[0364] After adjustment, the prediction time domain is shortened, which is beneficial for short-cycle waves. s, The prediction accuracy of Hz is improved; the feedforward gain is increased to cope with larger wave disturbances; and the PID parameters are optimized. Increase and enhance response speed Reduce and prevent integral saturation under high sea states. Increase and enhance damping; the filter passband of 0.08~0.8Hz covers the main wave energy frequency band of 0.08~0.4Hz.

[0365] In step S84, the critical data is recorded to the local storage medium at a frequency of not less than 10Hz.

[0366] This embodiment demonstrates the effectiveness of the invention under severe sea conditions. By coordinating and adjusting all parameters across the entire chain, including the prediction time domain, feedforward gain coefficient, PID parameters, and the passband range of the bandpass filter, the system maintains stable depth control performance even under high sea states, verifying the adaptability of the invention to severe sea conditions.

[0367] Example 4: Single Buoy Buoyancy Adjustment Type Depth Stabilization Example

[0368] This embodiment provides an active depth determination and heave compensation method for marine seismic exploration buoys, which is applicable to single buoy operation scenarios where depth needs to be adjusted by changing buoyancy.

[0369] In this embodiment, the lifting device adopts a buoyancy adjustment mechanism, including an outer oil bladder, an inner oil bladder, a high-pressure plunger pump, and an electromagnetic reversing valve.

[0370] In step S1, the controller loads the target depth value. m. Actuator parameters include: maximum flow rate of the high-pressure plunger pump. L / s m³ / s, maximum volume change of the external oil bladder L m³, switching frequency of electromagnetic directional valve Hz, response time constant s. Signal processing parameters: prediction time domain s, and Matching, feedforward gain coefficient PID parameters , , The controller completes self-testing and data buffer initialization.

[0371] In step S2, the accelerometer... The IMU collects triaxial acceleration data at a sampling frequency of Hz. Let the current IMU measure the roll angle... Pitch angle Heading angle Triaxial acceleration m / s² m / s² m / s². After attitude transformation matrix conversion:

[0372] rad, rad

[0373]

[0374]

[0375]

[0376] Vertical acceleration in the navigation coordinate system:

[0377] m / s²

[0378] Obtain heave acceleration m / s².

[0379] Acquire environmental parameters simultaneously: effective wave height m, spectral peak period s.

[0380] In step S3, the passband of the bandpass filter is 0.05~0.5Hz. Perform filtering.

[0381] Let the filtered acceleration be... m / s², the previous moment m / s², velocity at the previous moment m / s, sampling interval If ms, then the velocity at the current moment is:

[0382] m / s

[0383] Displacement at the previous moment m, then the displacement at the current moment:

[0384] m

[0385] After processing by a high-pass filter with a cutoff frequency of 0.01Hz, rad / s, to obtain a drift-free heave displacement signal. m is stored in the historical data cache area.

[0386] In step S4, the RLS algorithm , Regression vector Model coefficients at the previous time step Predicted value:

[0387] m

[0388] After RLS recursive update, we get Multi-step prediction yields the heave displacement in the next 0.8 seconds. m represents the upward trend. Predicting residual variance. If the value is below the preset threshold of 0.05, it indicates that the prediction confidence is good.

[0389] In step S5, the feedforward compensation amount is:

[0390]

[0391] The controller will Convert to pump speed command. Pump maximum flow rate. L / s, corresponding control quantity Therefore, the current pump speed is at the maximum flow rate. pump flow rate L / s m³ / s. Due to The buoy will move upward, and the electromagnetic reversing valve will switch to the "oil injection" position to inject hydraulic oil into the outer oil bladder, increasing the drainage volume and buoyancy, causing the vibration source to move downward relative to the buoy.

[0392] Control quantity limit: Corresponding maximum flow L / s, control quantity 1.0, Corresponding to -2L / s, the control quantity is -1.0. Within the limited range.

[0393] During the actuator's operation, the controller monitors the hydraulic system pressure in real time via a pressure sensor and the actual output flow of the pump via a flow meter to confirm that the actuator is acting according to the expected commands. If an abnormal pressure or excessive flow deviation is detected, the safety protection mechanism is immediately triggered.

[0394] Furthermore, to generate a 100N change in buoyancy, the volume of water displaced must be changed. m³ L is close to the maximum volume change of the outer oil bladder. L.

[0395] In step S6, the depth sensor monitors the actual depth in real time. m, depth deviation m. Feedback control cycle ms, let the deviation of the previous time step be... m, cumulative value of integral term 0.003.

[0396] Feedback compensation amount:

[0397]

[0398] Integral Separation Threshold m, Integral control is enabled. Total control quantity:

[0399]

[0400] After limiting The flow rate did not exceed the maximum flow control limit of 1.0, corresponding to the pump flow rate. The speed (L / s) is within a safe range. Feedforward control executes with a 10ms cycle, and feedback control executes with a 50ms cycle; both operate in parallel.

[0401] In step S8, the window length is evaluated. This corresponds to 10 seconds. Calculation m, m, m, with good depth-determining accuracy.

[0402] Based on significant wave height m, spectral peak period The system determines that the current sea state is moderate, and the parameter adaptive module fine-tunes the control parameters: The time was adjusted from 0.8s to 0.85s. The value was adjusted from 1.2 to 1.18, while the PID parameters remained unchanged. The adjustment was then sent back to each module, taking effect simultaneously within the same control cycle. Key data was recorded to local storage at a frequency of at least 10Hz.

[0403] This embodiment demonstrates the versatility of the present invention for different actuator types. When using a buoyancy adjustment mechanism, the feedforward compensation is converted into a flow command for the high-pressure plunger pump through a flow conversion coefficient. The control logic is completely consistent with the cable reeling and releasing mechanism, and the depth-holding accuracy also reaches the centimeter level, verifying the adaptability of the method of the present invention to different actuators.

[0404] Example 5: Multi-buoy weighted collaborative depth determination example

[0405] This embodiment provides an active depth determination and heave compensation method for marine seismic exploration buoys, which is applicable to multi-buoy array operation scenarios where the prediction confidence of each buoy is different.

[0406] In this embodiment, the number of buoys Each buoy is equipped with an accelerometer IMU, a depth sensor, and an underwater acoustic communication module. The lifting device uses a cable retrieval mechanism.

[0407] In step S1, each buoy is synchronously loaded with the same target depth value. m. Signal processing parameters for each buoy: s, , , , , , , Each buoy completes its self-check and data buffer initialization.

[0408] In step S2, each buoy... Acceleration data is independently acquired at a sampling frequency of Hz. Taking buoy 3 as an example, a low-confidence buoy, suppose its IMU measures the roll angle at a certain moment. Pitch angle Heading angle The large roll and pitch indicate that the wave disturbance at this buoy is more intense. The triaxial acceleration measurement value is... m / s² m / s² m / s².

[0409] Convert angles to radians: rad, rad.

[0410] Calculate the third row elements of the attitude transformation matrix of buoy 3:

[0411] Vertical acceleration in the buoy 3 navigation coordinate system:

[0412] m / s²

[0413] Simultaneously acquire environmental parameters at the location of each buoy: Buoy 1 m, s; 2 buoys m, s; 3 buoys m, The waves were most intense at buoy 3.

[0414] In step S3, the passband of the bandpass filter for each buoy is set to 0.05~0.5Hz. Taking buoy 3 as an example, let the filtered acceleration... m / s², with a large amplitude, at the previous moment m / s², velocity at the previous moment m / s, then:

[0415] m / s

[0416] Displacement at the previous moment m, then:

[0417] m

[0418] After high-pass filtering m is stored in the historical data cache area.

[0419] In step S4, each buoy undergoes multi-step prediction using the RLS algorithm. In this embodiment, the prediction confidence levels of each buoy differ significantly:

[0420] Buoy 1: m, High confidence level, the location has relatively stable waves.

[0421] Buoy 2: m, Medium confidence level

[0422] Buoy 3: m, Low confidence level, the location is experiencing the most intense waves, and the IMU signal quality is degraded.

[0423] Reasons for the differences in predicted displacement among the buoys: 3 locations on each buoy m, The waves are most violent, with more high-frequency components of heave and sag, resulting in relatively large tracking errors for the RLS prediction model.

[0424] In step S5, each buoy independently calculates the feedforward compensation amount:

[0425] Buoy 1:

[0426] Buoy 2:

[0427] Buoy 3:

[0428] In step S6, each buoy independently performs online depth feedback correction. Let the actual depth deviation of each buoy at a certain moment be... m、 m、 m, feedback control period Let ms be the deviations of each buoy at the previous moment, which are 0.04m, -0.02m, and 0.03m respectively, and the cumulative values ​​of the integral term be 0.002, 0.001, and 0.0018 respectively.

[0429] Feedback compensation amount for each buoy:

[0430]

[0431] In step S71, each buoy sends its predicted displacement, actual depth, and prediction confidence to the main buoy, buoy 1, at a period of 100ms. .

[0432] In step S72, the main buoy calculates the unweighted array average prediction bias:

[0433] m

[0434] In step S73, the main buoy calculates the weights based on the variance of the prediction residuals of each buoy:

[0435] , ,

[0436] Total weight:

[0437] Weighted average prediction bias:

[0438] m

[0439] The unweighted average prediction bias is 0.03m. Using a weighted strategy, the high-confidence buoy has a higher weighting. Low confidence buoy 3-weight ratio This improves the robustness of collaborative control.

[0440] The calculation of the coordinated compensation for each buoy uses a weighted average prediction deviation of 0.0776m.

[0441] Buoy 1:

[0442] Buoy 2:

[0443] Buoy 3:

[0444] If weighted collaboration is not used, and an unweighted average of 0.03m is used, then:

[0445] Buoy 1:

[0446] Buoy 2:

[0447] Buoy 3:

[0448] It can be seen that after weighted coordination, the coordination compensation amount of low confidence buoy 3 decreased from 0.052 to 0.023, which reduced the dependence on low quality data.

[0449] The total control quantity for each buoy is the sum of its respective feedforward compensation, coordinated synchronization control term, and feedback compensation.

[0450] Total control quantity of buoy 1:

[0451] Total control quantity of buoy 2:

[0452] Total control quantity of buoy 3:

[0453] Control limit value , All three total control quantities mentioned above are within the limit range.

[0454] In step S74, the depth difference threshold cm. Assume m, 4cm < 5cm If the value is 6cm > 5cm and remains so for 1.5s, but the 2s alarm condition has not yet been met, continue monitoring.

[0455] If the variance of the prediction residual of a certain buoy If it consistently exceeds the threshold of 0.16, the corresponding weight... This indicates that the buoy's prediction data has low reliability, and the system automatically excludes it from the cooperative control, while the remaining buoys continue to cooperate.

[0456] In step S8, each buoy independently performs control effect evaluation and parameter adaptive adjustment. Assume that the calculated parameters for each buoy are... The depths are 0.04m, 0.05m, and 0.055m respectively. (Bus 3) The depth m is slightly higher than that of other buoys, but still meets the depth-keeping accuracy requirements. Each buoy independently adjusts its control parameters based on its own sea state parameters.

[0457] This embodiment uses a weighted collaborative strategy to maintain the performance of array collaborative control even when the quality of some buoy data deteriorates.

[0458] This embodiment demonstrates the robustness of the present invention in scenarios where there are differences in prediction confidence among buoys. Through a weighted collaborative strategy, the weight of low-confidence buoys is automatically reduced, and their collaborative compensation amount is reduced from an unweighted 0.052 to 0.023, reducing the dependence on low-quality data and maintaining the performance of array collaborative control even when the data quality of some buoys deteriorates.

[0459] Example 6: Shallow Water Operation Depth Measurement Example

[0460] This embodiment provides an active depth determination and heave compensation method for marine seismic detection buoys, applicable to shallow water areas with depths less than 20m. Figure 1 As shown, this embodiment executes steps S1 to S6 and S8, and does not enable S7 multi-buoy collaborative depth determination.

[0461] In this embodiment, the lifting device adopts a cable winding and unwinding mechanism. The wave period in shallow water is usually short, i.e., 2~4s, and the wave amplitude is relatively small, but the stroke of the actuator is limited. The water depth in shallow water is shallow, and the adjustable range of the vibration source is small.

[0462] In step S1, the controller loads the target depth value. m. The actuator parameters are: maximum winch speed. rpm, the maximum speed at which the cable is retracted or extended. m / s, response time constant In shallow water, wave periods are short, requiring faster response times. Signal processing parameters: prediction time domain. s, and Matching, feedforward gain coefficient The PID parameters were determined by Ziegler-Nichols tuning. , , The maximum stroke of the actuator is limited to 3 meters, and the control amplitude is limited. The corresponding maximum travel distance is 3m. Corresponding to -3m. After performing self-tests on each sensor and communication module, the controller initializes the data buffer.

[0463] In step S2, the accelerometer sampling frequency is set to Hz, sampling interval ms. Assume the current IMU measures the roll angle. Pitch angle Heading angle The triaxial acceleration measurement value is m / s² m / s² m / s².

[0464] First, convert the angle to radians: rad, rad, rad.

[0465] Calculate the attitude transformation matrix Elements:

[0466]

[0467]

[0468]

[0469] Vertical acceleration:

[0470] m / s²

[0471] Obtain heave acceleration m / s².

[0472] Simultaneous acquisition of environmental parameters: typical spectral peak periods in shallow water areas s, significant wave height m.

[0473] In step S3, the passband range of the bandpass filter is adjusted to 0.1~0.8Hz to accommodate short-period waves in shallow water. The cutoff frequency of the high-pass filter is set to... Hz, rad / s.

[0474] Suppose the filtered acceleration is at a certain sampling time The value is m / s², in shallow water the wave height is smaller and the acceleration amplitude is smaller, the previous moment m / s², velocity at the previous moment m / s, sampling interval If ms, then the velocity at the current moment is:

[0475]

[0476] If the displacement at the previous moment m, then the displacement at the current moment:

[0477]

[0478] After high-pass filtering with a cutoff frequency of 0.01Hz, a drift-free heave displacement signal is obtained. m. State estimation is performed cyclically with a 10ms period, recording the current heave and displacement. Store in historical data cache ,in In shallow water, the wave period is short, and the historical window is shortened to 0.5s.

[0479] In step S4, the order of the prediction model is determined. Historical Window For 0.5s of historical data, the forgetting factor In shallow water, wave changes are rapid, requiring faster tracking speeds. RLS algorithm initial conditions: , Prediction in the time domain s, number of predicted steps .

[0480] Assuming the regression vector at the current moment Model coefficients at the previous time step Then the predicted value is:

[0481]

[0482] RLS recursively updates the model coefficients at the current time step. Then, the heave displacement in the next 0.5 seconds is predicted through multi-step recursion. In the multi-step prediction, the unknown historical value in the regression vector of each step is replaced by the predicted value of the previous step. After 50 recursive steps, the result is obtained. .

[0483] Assuming the multi-step prediction result is m represents an upward trend. Simultaneously, the variance of the predicted residuals is calculated. If the value is below the preset threshold of 0.05, it indicates that the prediction confidence is good.

[0484] In step S5, the feedforward compensation amount is calculated as follows:

[0485]

[0486] Control limit value The corresponding maximum cable winding and unwinding speed is 0.5 m / s. Corresponding to -0.5m / s, Within the limited range, therefore The feedforward compensation is output to the winch unit 0.5 seconds before the actual arrival of the wave disturbance, and the winch... The cable-laying action is performed at a speed of m / s, causing the seismic source to move downward relative to the buoy, compensating for the expected upward displacement.

[0487] During the actuator's operation, the controller monitors the winch speed and cable length in real time via the encoder to confirm that the actuator is acting according to the expected instructions. If an abnormal response from the actuator is detected, the safety protection mechanism is immediately triggered.

[0488] In step S6, the depth sensor monitors the actual water depth in real time. Assuming the current actual depth is 4.97m, the target depth is... m, depth deviation:

[0489]

[0490] Feedback control cycle ms, let the deviation of the previous time step be... m, integral term Its current value is 0.001.

[0491] Feedback compensation amount:

[0492]

[0493] Integral Separation Threshold m, in shallow water areas the target depth is small, and the threshold is correspondingly reduced. m m, enable integral control. Total control quantity:

[0494]

[0495] Total control quantity exist Within the range, therefore after limiting the amplitude Feedforward control executes with a 10ms cycle, and feedback control executes with a 50ms cycle; both operate in parallel.

[0496] In step S8, the window length is evaluated. This corresponds to 10 seconds. Assume a depth deviation sequence of the most recent 200 sampling points. Sum of squares m², then:

[0497]

[0498] Depth peak error m, standard deviation of depth error m. Based on significant wave height m, spectral peak period The current sea state is determined to be moderate.

[0499] The parameter adaptive module adjusts the control parameters according to the sea state: the current sea state is moderate. The time was adjusted from 0.5s to 0.6s. The value was adjusted from 0.6 to 0.65. The adjusted value was then sent back to each module and took effect simultaneously within the same control cycle.

[0500] If sea conditions deteriorate m or s, the parameter adaptation module will Reduce to 0.3s, which is the lower limit. Increase to 0.8, the upper limit, to ensure the trip is not exceeded. If sea conditions improve, m, It can be increased to 0.8s, which is the upper limit. Reduce to 0.4, which is the lower limit. Critical data is recorded to local storage media at a frequency of no less than 10Hz.

[0501] This embodiment demonstrates the adaptability of the present invention in shallow water operation scenarios. By configuring parameters for shallow water conditions, shortening the prediction time domain, adjusting the filter passband, and narrowing the control amplitude, effective depth control is achieved under travel-limited conditions, verifying the adaptability of the present invention to different operating environments.

[0502] It should be noted that the above embodiments are only used to explain and illustrate the present invention, and are not intended to limit the present invention. The scope of protection of the present invention is defined by the claims. The specification and drawings are only for explanation and illustration. Any modifications, equivalent substitutions, improvements, etc., made by those skilled in the art based on the content described in the specification and drawings without departing from the concept of the present invention should fall within the scope of protection of the present invention.

Claims

1. A method for active depth determination and heave compensation of a marine seismic buoy, characterized in that, Includes the following steps: S1: System initialization and parameter configuration, the controller loads the preset target depth value and lifting device actuator parameters, and establishes a data cache area; S2: Wave disturbance sensing and acceleration signal acquisition. The controller senses the buoy's heave motion in real time through the accelerometer (IMU) and acquires triaxial acceleration measurements in the buoy's coordinate system. , , After attitude transformation matrix Perform coordinate transformation and subtract gravity components Then obtain the heave acceleration in the navigation coordinate system. Simultaneously collect effective wave height Spectral peak period Marine environmental parameters; S3: Heave acceleration signal preprocessing and heave motion state estimation. The controller uses a bandpass filter to process the heave acceleration signal. Bandpass filtering is performed, followed by double numerical integration to obtain the heave velocity and heave displacement. Then, a high-pass filter is used to eliminate high-pass drift and extract the pure wave heave displacement signal. ; S4: Heave displacement prediction. The controller uses the historical wave heave displacement data and employs the recursive least squares method to identify the coefficients of the third-order linear prediction model online, predicting future heave displacements in the time domain. heave displacement And calculate the variance of the predicted residuals. As a quantitative indicator of prediction confidence; S5: Feedforward compensation calculation and advance drive; the controller calculates the predicted heave displacement. and feedforward gain coefficient Calculate the feedforward compensation control quantity Leading the arrival of wave disturbances The time output is sent to the lifting device actuator for reverse compensation; S6: Depth feedback online correction; the controller measures the depth of the vibration source in the water using a depth sensor mounted on the buoy. Calculate depth deviation The PID control algorithm is used to calculate the feedback compensation amount. and feedforward compensation amount Superimposed as total control quantity The feedforward control operates at a first preset period. The feedback control is executed at a second preset period. Execution, satisfaction ; S7: Multi-buoy collaborative depth determination, where each buoy in the multi-buoy array coordinates its communication cycle via underwater acoustic communication. Share their respective predicted heave and sag displacements and prediction confidence The average prediction bias of the main buoy or shipboard central controller's computational array The weighted average prediction bias is obtained by weighting the prediction confidence levels of each buoy. Each buoy calculates the collaborative feedforward compensation amount according to a unified control law that integrates absolute depth control and synchronization control. The controller integrates absolute depth control and relative consistency control into a single control law; it monitors the depth difference between each buoy in real time and triggers an alarm when the difference exceeds the limit. S8: Control effect evaluation and parameter adaptive adjustment; the controller evaluates the depth control effect in real time and calculates the root mean square error of depth. Peak depth error and depth error standard deviation According to the significant wave height Spectral peak period and standard deviation of heave displacement A comprehensive assessment of the sea state level was conducted, along with predictions for the time domain. Feedforward gain coefficient PID parameters , , The passband range of the bandpass filter is adjusted in a coordinated manner. In steps S2 to S8: the acceleration acquisition result of S2 provides data input for S3; the displacement estimation result of S3 provides prediction model input for S4; the prediction result of S4 provides feedforward compensation basis for S5; the feedforward compensation amount of S5 and the feedback compensation amount of S6 are output in a superimposed manner; S7 superimposes synchronous control incremental correction on the feedforward compensation amount of S5; the adaptive adjustment result of S8 is fed back to S3, S4, S5 and S6 respectively.

2. The method according to claim 1, characterized in that, Step S2 specifically includes: S21: The coordinate transformation is performed through the attitude transformation matrix. The attitude transformation matrix is ​​achieved by the buoy's roll angle. Pitch angle Heading angle Real-time configuration, vertical acceleration in the navigation coordinate system The calculation formula is: in , , These are the triaxial acceleration measurements in the carrier coordinate system. The acceleration due to gravity; the vertical acceleration This refers to the heave acceleration mentioned in step S3. ; S22: The controller synchronously acquires marine environmental parameters such as wind speed, wind direction, significant wave height, and spectral peak period for subsequent sea state level determination.

3. The method according to claim 1, characterized in that, Step S3 specifically includes: S31: Lower limit of the passband frequency of the bandpass filter The lowest frequency in the wave band, and the upper limit of the passband frequency. The highest frequency in the wave band, satisfying ; S32: The second numerical integration adopts the trapezoidal integration method, and the heave velocity and heave displacement are obtained by integrating the filtered acceleration signal in sequence. S33: The cutoff frequency of the high-pass filter satisfy .

4. The method according to claim 1, characterized in that, Step S4 specifically includes: S41: The prediction model is a third-order linear model. in For heave displacement, These are the model coefficients. To predict residuals; S42: The coefficients of the prediction model are identified online using the recursive least squares method, and the recursive formula is: in , , Let covariance matrix be the variance matrix. Forgetting factor, ; S43: When predicting residual variance When the preset threshold is exceeded, the controller reduces the feedforward gain coefficient. It prioritizes feedback control.

5. The method according to claim 1, characterized in that, Step S5 specifically includes: S51: The calculation formula for the feedforward compensation control quantity is: in This is the feedforward gain coefficient. For the predicted future heave and sag displacement, For prediction in the time domain; S52: The lifting device actuator is a cable winding and unwinding mechanism or a buoyancy adjustment mechanism; S53: The way the feedforward compensation amount is converted into actuator control commands depends on the type of actuator. When the actuator is a cable take-up and release mechanism, the controller will... The controller converts the speed command to the winch motor. When the actuator is a buoyancy adjustment mechanism, the controller will... Convert to flow command for high-pressure plunger pump.

6. The method according to claim 1, characterized in that, Step S6 specifically includes: S61: The depth deviation The calculation formula is: S62: The PID feedback control quantity is: in , , These are the proportional, integral, and differential gain coefficients, respectively. For feedback control cycle, This is the sequence number of the current feedback control cycle; S63: The feedforward control operates at a first preset period. The feedback control is executed at a second preset period. Execution, satisfaction ; S64: The total control quantity For feedforward compensation amount With feedback compensation amount Superposition: S65: The feedforward control cycle Feedback control cycle Cooperative communication cycle with multiple buoys To form a three-timescale collaborative architecture that satisfies .

7. The method according to claim 1, characterized in that, Step S7 specifically includes: S71: Each buoy sends its own heave displacement prediction data, actual depth and prediction confidence level to the main buoy or shipboard central controller via underwater acoustic communication in a coordinated communication cycle. S72: The main buoy or shipboard central controller collects prediction data from all buoys and calculates the array's average prediction bias. in For the number of buoys, For the first Predicted heave displacement of a single buoy; S73: Calculation of the coordinated compensation amount for each buoy, and obtaining the weighted average prediction bias by weighting the prediction confidence levels of each buoy: in For the first The weight of each buoy, For the first The variance of the prediction residuals for each buoy; The coordinated feedforward compensation of each buoy is calculated according to a unified control law that integrates absolute depth control and synchronization control: in For the first The amount of coordinated feedforward compensation for each buoy This is the feedforward gain coefficient. is the synchronization gain coefficient; the first term on the right side of the equation is the absolute depth control term, and the second term is the synchronization control term; S74: The controller monitors the depth difference between each buoy in real time. When the depth difference between any two buoys exceeds the preset threshold and continues for more than the preset time, an alarm is triggered.

8. The method according to claim 1, characterized in that, Step S8 specifically includes: S81: The control effect evaluation index includes the deep root mean square error. Peak depth error and depth error standard deviation ; S82: The sea state level is based on the significant wave height. Spectral peak period and standard deviation of heave displacement Comprehensive judgment; S83: The coordinated adjustment includes the controller simultaneously adjusting the prediction time domain. Feedforward gain coefficient PID parameters , , The passband range of the bandpass filter is adjusted, and the adjusted parameters take effect simultaneously within the same control cycle.

Images