Intelligent unloading platform anti-pinch and follow-up control method

Abstract

The present application belongs to the technical field of fluid pressure control system, and particularly relates to an intelligent unloading platform anti-pinch and follow-up control method, which comprises the following steps: calculating the instantaneous output power of a hydraulic system in real time, and performing time-frequency domain joint analysis on the instantaneous output power to extract a composite risk index capable of distinguishing between instantaneous impact and persistent disturbance, and a follow-up energy instruction representing the lifting of a carriage; adopting an adaptive energy control strategy, the control law of which comprises a feedforward term for tracking follow-up and a feedback term for suppressing oscillation, wherein the damping gain of the feedback term is dynamically adjusted by the composite risk index; and converting the target energy into a hydraulic valve control signal through an online self-correcting inverse model. The present application realizes seamless transition from accurate follow-up to flexible suppression to emergency braking, and significantly improves the safety and smoothness of the platform.

Description

Technical Field

[0001] This invention relates to the field of fluid pressure control system technology. More specifically, this invention relates to an anti-pinch and follow-up control method for an intelligent unloading platform. Background Technology

[0002] A loading and unloading platform is a loading and unloading device driven by an electro-hydraulic system, featuring a liftable main platform and a tiltable ramp. It connects the warehouse platform to the truck bed, enabling seamless passage for forklifts and other handling equipment. During loading and unloading, the vehicle's suspension system may sink or rebound due to changes in cargo weight, resulting in a height difference between the truck bed floor and the platform. Therefore, the hydraulic system of the loading platform needs to be dynamically controlled to follow the real-time height changes of the truck bed floor.

[0003] Currently, a vehicle-mounted floating adjustment function is commonly used. This involves installing mechanical follow-up valves or floating oil circuits in the hydraulic system. Once the platform is placed on the truck bed, the main hydraulic cylinder is allowed to extend and retract freely within a certain range to passively adapt to the truck bed's undulations. Simultaneously, to ensure safety, existing technologies also incorporate anti-toe-pinch safety features, such as safety baffles installed on both sides of the platform. Related technologies, such as Chinese patent document CN218118209U, disclose a power unit for a fixed unloading platform. This patent discloses a method of connecting the power unit to the drive motor and oil tank via an oil circuit block, resulting in stable performance, simple operation, compact structure, low maintenance rate, and easy repair.

[0004] However, the existing follow-up control technology is passive and lagging. The platform will only move when a sufficiently large force is generated between the truck bed and the platform to overcome the static friction and damping of the hydraulic system. This results in the platform's following motion not being a smooth, continuous process, but rather a jerky, step-like movement, which is prone to causing bumps and impacts during loading and unloading. Secondly, the existing technology lacks effective identification and response to abnormal loads, and cannot distinguish between the slow descent of the vehicle and the violent impact of a forklift rapidly moving onto the bridge. Under impact, the platform may sink instantly, posing a safety hazard. Summary of the Invention

[0005] To address the aforementioned technical problems of poor anti-pinch and follow-up control performance of unloading platforms, this invention provides an intelligent anti-pinch and follow-up control method for unloading platforms, comprising:

[0006] The system acquires real-time pressure and displacement data of the main lifting hydraulic cylinder of the unloading platform; calculates the instantaneous output power of the hydraulic cylinder; performs time-frequency domain joint analysis on the instantaneous output power to extract a composite risk index characterizing system safety risk and a follow-up energy command characterizing the lifting of the cargo compartment; generates a target energy command using an adaptive energy control strategy, the strategy including a feedforward term for tracking the follow-up energy command and a feedback term whose gain is dynamically adjusted by the composite risk index to provide active damping; and converts the target energy command into a hydraulic valve control signal through a control inverse model and executes it.

[0007] This invention, through active calculation of system power flow and time-frequency domain analysis, can identify servo intentions and safety risks. Compared to existing technologies that cannot distinguish between normal loads and abnormal impacts, the composite risk index constructed in this invention can distinguish between instantaneous hard impacts and continuous soft disturbances. Compared to the rigid switching between different modes in traditional control strategies, this invention, through an energy control framework that smoothly adjusts active damping based on risk indices, achieves an adaptive transition from accurate servoing to flexible anti-pinch, improving the smoothness and safety of the unloading platform's operation.

[0008] Preferably, calculating the instantaneous output power of the hydraulic cylinder includes: multiplying the pressure difference between the rod chamber and the rodless chamber of the hydraulic cylinder by the effective area of ​​the piston to obtain the net output force, and then multiplying the net output force by the real-time movement speed of the piston.

[0009] Preferably, the instantaneous output power of the hydraulic cylinder satisfies the expression:

[0010] ;

[0011] In the formula, Indicates time Instantaneous output power, measured in watts. ; This represents the effective piston area of ​​the rodless chamber in a hydraulic cylinder, expressed in square meters. ; This represents the effective piston area of ​​the rod chamber in a hydraulic cylinder, expressed in square meters. ; , Indicates time The pressure in the rodless and rod chambers of a hydraulic cylinder, measured in Pascals. ; Indicates time The piston displacement, in meters. .

[0012] This invention utilizes the sensitivity of high-frequency energy to hard impacts such as collisions, and the sensitivity of mid-to-high-frequency energy variance to soft disturbances such as system oscillations and forklifts crossing bridges. By combining these two factors, it is possible to assess the type and level of risk currently faced by the system, providing a decision-making basis for the adaptive adjustment of subsequent control strategies.

[0013] Preferably, the time-frequency domain joint analysis of the instantaneous output power includes: performing wavelet packet transform on the instantaneous output power to decompose it into a preset high-frequency band and a mid-high-frequency band; the composite risk index is the sum of the standard deviations of the high-frequency band energy and the mid-high-frequency band energy.

[0014] This invention extracts the servo energy command representing the slow lifting and lowering intention of the carriage by low-pass filtering and integration of the power signal. The integrated energy is used as the feedforward command. Compared with directly using the noisy speed signal, it can provide a smoother and more stable tracking target, suppress noise interference, and lay the foundation for achieving smooth servo control.

[0015] Preferably, the acquisition of the follow-up energy command includes: performing low-pass filtering on the instantaneous output power to separate the low-frequency power component caused by the slow rise and fall of the carriage; and performing time integration on the low-frequency power component.

[0016] Preferably, the target energy command satisfies the expression:

[0017] ;

[0018] In the formula, Indicates time The target energy command is in joules. ; Indicates time The servo energy command, measured in joules. ; Indicates time The adaptive active damping gain, in Newton-seconds. ; Indicates time The actual speed of the platform, in meters per second. .

[0019] The feedforward term of this invention It actively compensates for the energy required for follow-up, while the feedback term This provides variable active damping. This feedforward + feedback energy control structure logically separates the tracking task and the stabilization task, facilitating subsequent adjustment... It provides a foundation for unifying different control behaviors and is structurally superior to traditional controllers.

[0020] Preferably, the adaptive active damping gain satisfies the expression:

[0021] ;

[0022] in, Indicates time The adaptive active damping gain, in Newton-seconds. ; This is the basic damping gain, measured in Newton-seconds. ; Indicates time The composite risk index, in watt-square seconds. ; and These are the historical mean and standard deviation of the composite risk index under normal operating conditions, in watt-square seconds. ; Represents the hyperbolic tangent function; This indicates a preset small value to avoid a denominator of 0.

[0023] Preferably, the control inverse model is a neural network model trained based on historical calibration data, used to establish the mapping relationship between target energy commands and hydraulic valve control signals.

[0024] Preferably, the method further includes: performing online adaptive correction on the control inverse model, including: calculating the energy tracking deviation between the actual output power of the hydraulic cylinder and the target energy command; and updating the network parameters of the neural network model online using a gradient descent algorithm based on the energy tracking deviation.

[0025] This invention calculates the energy tracking deviation in real time and uses this deviation to update the inverse model parameters online, enabling the controller to continuously learn and compensate for changes in system characteristics, thus ensuring optimal control performance throughout the entire equipment lifecycle.

[0026] Preferably, the method further includes an offline calibration step for acquiring historical data required to construct the control inverse model and determine the time-frequency domain analysis parameters, including: performing a step-by-step loading test in a stationary state of the platform to obtain static load-pressure-displacement characteristic data; performing a standard lifting and lowering cycle at different speeds under no-load conditions to obtain dynamic reference data; and applying a standard impact to the platform using a force control device to obtain typical collision event data.

[0027] The beneficial effects of this invention are as follows: by performing time-frequency domain analysis on the output power of the hydraulic system, a composite risk index that can distinguish between instantaneous impact and continuous oscillation is constructed. Based on this index, the active damping gain in the energy control law is smoothly and dynamically adjusted. With a unified control framework, adaptive switching from precise follow-up to flexible anti-pinch is achieved. Furthermore, the long-term accuracy of the control is ensured through an online self-correcting inverse model, thereby improving the safety and smoothness of the unloading platform. Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating an anti-pinch and follow-up control method for an intelligent unloading platform according to the present invention;

[0029] Figure 2 This is a schematic diagram illustrating the dynamic changes of the valve control signal and the adaptive damping gain. Detailed Implementation

[0030] This invention discloses an anti-pinch and follow-up control method for an intelligent unloading platform, referring to... Figure 1 This includes steps S1-S4:

[0031] S1: Obtain historical calibration data of the hydraulic system of the unloading platform and build a benchmark model library for condition assessment.

[0032] It should be noted that, in order to achieve efficient and safe control of the unloading platform, it is first necessary to establish a database that can accurately describe its normal behavior under various controlled operating conditions. This invention, through a series of standardized tests on the unloading platform in the offline phase, collects core hydraulic system data and, based on this, constructs a benchmark model library containing multiple benchmark models, providing a reference for subsequent online real-time anomaly detection and status assessment.

[0033] Specifically, this step is the offline calibration and modeling stage, the purpose of which is to obtain all the basic data used to build the subsequent analysis model, including:

[0034] With the platform stationary, by progressively loading standard counterweights, the correspondence between the first and second pressures in the rod-side and rodless chambers of the main lifting hydraulic cylinder and the platform displacement is recorded simultaneously, forming a static load-pressure-displacement dataset. It should be noted that the standard counterweights are a series of objects of known weight, such as multiple 500kg iron blocks, whose weight should cover the platform's rated load range. It should also be noted that the unit of pressure is Pascal. The unit of displacement is meters. .

[0035] Under standard operating conditions with no load and no obstructions, the platform executes standard lifting and lowering cycles at multiple different preset speeds. During this process, the corresponding first pressure time series, second pressure time series, and displacement time series are fully recorded at a preset sampling frequency, collectively forming a dynamic benchmark dataset. It should be noted that, for example, the preset speeds are 20%, 50%, and 80% of the maximum speed, and the preset sampling frequency is 1 kHz. The unit of the preset speed is meters per second. .

[0036] During the acquisition of dynamic benchmark datasets, a force control device simulates collision impacts on the platform's movement using multiple preset forces that meet safety standards, and records the corresponding pressure sequences to form a collision event dataset. For example, the preset forces are 50N and 100N.

[0037] Preferably, a benchmark model library is constructed based on historical calibration data, including:

[0038] Based on static characteristic data, a load mapping model is established through multinomial regression or neural network fitting, which can accurately calculate the static load on the platform according to real-time pressure and displacement.

[0039] Based on a dynamic benchmark dataset, a generative model, such as an autoencoder, is trained by extracting a multidimensional state vector containing pressure, displacement, and their derivatives to obtain a dynamic probabilistic model that can evaluate the normality of real-time states.

[0040] The pressure sequences in the collision event dataset are analyzed in the time and frequency domain using continuous wavelet transform to extract high-frequency energy features and construct a typical fingerprint database of collision events.

[0041] Thus far, historical calibration data, including static characteristic data, dynamic benchmark datasets, and typical event datasets, have been acquired during the offline phase, and a benchmark model library, including a load mapping model, a dynamic probability model, and a typical fingerprint database, has been constructed.

[0042] S2: Perform power flow and frequency domain joint analysis on the real-time data of the hydraulic system acquired online to obtain a composite state index characterizing the risk and motion intention of the hydraulic system.

[0043] It should be noted that during the online operation of the unloading platform, the safety status and follow-up requirements are implicit in the energy flow within the hydraulic system. Existing technologies typically analyze pressure or displacement signals separately, making it difficult to distinguish similar signal appearances caused by different physical events.

[0044] It should be noted that changes in the output power of a hydraulic system can more fundamentally and sensitively reflect changes in external load. For example, when a platform collides with an obstacle, it will manifest as an instantaneous negative impact on the output power, while the slow lifting and lowering of the carriage will manifest as a stable, continuous positive or negative value in the output power.

[0045] It should be further explained that in order to accurately separate different event characteristics from complex power signals, this invention performs joint analysis of power flow and frequency domain. First, the instantaneous output power of the hydraulic system is calculated, and then time-frequency domain analysis is performed to extract composite risk indicators that can simultaneously characterize collision risk and instability risk, as well as indicators that characterize follow-up intention.

[0046] Specifically, based on the real-time acquired core status data, the instantaneous output power of the hydraulic system is calculated, including:

[0047] In each control cycle, the real-time first pressure time series, real-time second pressure time series, and real-time displacement time series of the main lifting hydraulic cylinder are acquired. It should be noted that the control cycle is the time interval between one complete calculation and decision-making operation performed by the control method of the present invention; for example, the control cycle is 50 milliseconds.

[0048] It should be noted that the output power of a hydraulic system is equal to the product of the output force of the hydraulic cylinder and the piston speed. Based on this logic, the instantaneous output power of the hydraulic system can be obtained.

[0049] The instantaneous output power of a hydraulic system satisfies the following expression:

[0050] ;

[0051] In the formula, Indicates time Instantaneous output power, measured in watts. ; This represents the effective piston area of ​​the rodless chamber in a hydraulic cylinder, expressed in square meters. ; This represents the effective piston area of ​​the rod chamber in a hydraulic cylinder, expressed in square meters. ; , Indicates time First pressure, second pressure, in Pascals ; This represents the displacement value at time t in a real-time displacement time series, in meters. ; This represents the value at time t, obtained by differentiating the real-time displacement time series with respect to time, in meters per second. It should be noted that, , The hydraulic cylinder design parameters are derived from the equipment data. Due to the presence of the piston rod, Area smaller Its value is the total area of ​​the piston minus the cross-sectional area of ​​the piston rod. It should be noted that the unit on the left side of the expression is watts. ,because , , Therefore, the unit on the right is Therefore, the units on the right side of the expression are consistent with the units and dimensions on the left side of the expression.

[0052] This represents the thrust generated on the piston by the pressure in the rodless chamber. This indicates the pulling force generated on the piston by the pressure in the rod chamber; It represents the net output force of the hydraulic cylinder. By calculating the resultant force generated by the pressure difference on both sides of the piston, it can accurately reflect the work done by the hydraulic system on the external load. This indicates the real-time movement speed of the hydraulic cylinder piston. This represents the instantaneous output power of the hydraulic system on the platform at time t, calculated by multiplying the net output force of the hydraulic cylinder by the real-time velocity of the piston. A larger value indicates a larger net output force of the hydraulic cylinder and a larger real-time velocity of the piston. The greater the instantaneous output power of the hydraulic system.

[0053] Preferably, time-frequency domain analysis is performed on the instantaneous output power to obtain a composite risk index, including:

[0054] It should be noted that unloading platforms experience both hard impacts from collisions and soft impacts and oscillations caused by rapid forklift movement or system instability. Hard impacts manifest as extremely high-frequency, energy-concentrated pulses in the power signal, while soft impacts or oscillations are characterized by a sustained increase in energy in the mid-to-high frequency range. This invention uses wavelet packet decomposition to break down the power signal into different frequency bands, and then calculates the energy of each frequency band to identify different types of risks.

[0055] The composite risk index satisfies the following expression:

[0056] ;

[0057] In the formula, Indicates time The composite risk index, measured in watt-square seconds. ; Indicates time The instantaneous output power of the hydraulic system, measured in watts. ; Indicates time The instantaneous output power of a hydraulic system is the energy in the preset high-frequency band HF, measured in watts per square second. ; Indicates time The instantaneous output power of a hydraulic system is the energy in the preset mid-to-high frequency (MHF) band, measured in watt-square seconds. ; This indicates the calculation of the variance function. The frequency ranges of the preset high-frequency band (HF) and the preset mid-high-frequency band (MHF) are determined by performing Fast Fourier Transform analysis on typical event datasets to find the characteristic frequency bands where the energy of collision events and load disturbance events is most concentrated, respectively. It should be noted that the first term on the right side of the expression... The unit is watt-square second. In the second item, The standard deviation is obtained by taking the square root of the variance of the energy, and the unit is watts squared seconds. Therefore, the units on both sides of the equals sign in the expression are consistent.

[0058] In the formula, More sensitive to instantaneous impact events such as collisions, and It is more sensitive to persistent disturbances such as forklifts going onto the bridge or system oscillations. Indicates time The standard deviation of the instantaneous output power of the hydraulic system in the preset mid-to-high frequency (MHF) band is used to unify the dimensions, therefore... and Adding them together gives the time. The composite risk index.

[0059] when When the threshold value exceeds the adaptive threshold determined by the dynamic probability model based on the current operating conditions, the intelligent controller of this invention determines that a safety risk exists. Specifically, the current real-time multidimensional state vector is input into the dynamic probability model to obtain the reconstruction error; the intelligent controller has pre-learned the mapping relationship between the reconstruction error and the fluctuation range of the composite risk index under normal operating conditions, and thus dynamically determines the threshold value based on the magnitude of the current reconstruction error.

[0060] It should be noted that the slow lifting and lowering of the carriage corresponds to a smooth and continuous energy input or output of the hydraulic system. Directly using speed as a feedforward signal is susceptible to noise interference, while energy, as an integral quantity, is not sensitive to instantaneous noise and can more stably reflect the overall responsiveness requirements.

[0061] Preferably, the instantaneous output power is low-pass filtered and integrated to obtain the servo energy command, including:

[0062] The instantaneous output power is low-pass filtered to separate the low-frequency power component caused by the slow rise and fall of the carriage; the low-frequency power component is integrated over time to obtain the follow-up energy command.

[0063] Thus, we have obtained a composite risk index characterizing system risk and a follow-up energy command characterizing follow-up demand.

[0064] S3: Adopt an adaptive energy control strategy based on risk level to generate hydraulic system control commands.

[0065] It should be noted that the control of the unloading platform needs to achieve an optimal balance between safety and performance. When the system faces different levels of risk, its control strategies should be fundamentally different. Traditional hybrid control strategies typically involve hard switching between several fixed modes, lacking smooth transitions and quantitative responses to risk levels. Composite risk indicators are continuously changing quantities that directly reflect the current risk level faced by the system. Therefore, this invention establishes an adaptive energy control strategy based on risk levels, using composite risk indicators as key adjustment factors that directly act on the control law. This allows for smooth adjustment of the controller's behavior according to the risk level, transitioning from precise servoing to flexible suppression, and then to emergency braking.

[0066] It should be noted that the behavior of an ideal unified control law should automatically adjust according to the risk level. When there is no risk, it should ensure that the output energy is exactly equal to the energy required for servoing. When risk occurs, it should actively provide damping to suppress oscillations, and the higher the risk level, the stronger the damping effect. Therefore, this invention designs a control law with energy as the control objective. This control law includes a feedforward term for tracking the servo energy command and a feedback term whose gain is dynamically adjusted by a composite risk index to provide active damping.

[0067] Specifically, an adaptive energy control strategy is used to generate hydraulic valve control signals, including:

[0068] It should be noted that the task of the control system is to output valve control signals to generate the desired target energy. Therefore, it is necessary to establish a model that can describe the inverse mapping relationship from the control signal to the energy generation process.

[0069] By learning from historical calibration data, an inverse model is obtained that converts the target energy into a valve control signal: for each segment of data in the dynamic benchmark dataset, the instantaneous output power of the hydraulic system is integrated to obtain the actual output energy; using the actual output energy as the input feature and the corresponding hydraulic valve control information as the output label of the model, a neural network model is trained to obtain the inverse model of the valve control signal.

[0070] The hydraulic valve control signal satisfies the expression:

[0071] ;

[0072] ;

[0073] In the formula, Indicates at time The hydraulic valve control signal output to the electro-hydraulic proportional valve is a dimensionless value. This represents the inverse model that converts the target energy into a valve control signal, and the output is a dimensionless numerical value. This indicates the target energy command, measured in joules. ; This represents the follow-up energy command, which, as a feedforward term, drives the platform to follow the movement of the carriage; the unit is joules. ; This indicates the platform's actual speed, measured in meters per second. ; This represents the adaptive active damping gain, measured in Newton-seconds. It should be noted that the left side... The unit is joule. The first item on the right The unit is joule. ; Second item The unit is Joule Therefore, the units on both sides of the equals sign in the expression are consistent.

[0074] In the formula, As a feedforward term, the drive platform follows the movement of the carriage. It represents the feedback term used to provide active damping, which is dynamically adjusted by the composite risk index, and represents the work done by the damping force that is proportional to the velocity.

[0075] In the formula, the adaptive active damping gain satisfies the expression:

[0076] ;

[0077] in, Indicates time The adaptive active damping gain, in Newton-seconds. ; This is the basic damping gain, measured in Newton-seconds. ; Indicates time The composite risk index, in watt-square seconds. ; and These are the historical mean and standard deviation of the composite risk index under normal operating conditions, both in watt-square seconds. ; Represents the hyperbolic tangent function; To represent minute values ​​and avoid denominators of 0, for example... For example, For 1000 It should be noted that the unit on the left side of the expression is Newton-second. In the right side of the expression, No unit No unit The unit is Newton-second. Therefore, the units on both sides of the expression are consistent.

[0078] In the formula, This indicates that the composite risk index has been standardized. Indicates to Perform normalization processing to make It maps non-linearly between -1 and 1; Indicates to The numerical range is adjusted to 0 to 2, with the value corresponding to the state of low composite risk index being 1, while the value corresponding to the state of high composite risk index is close to 2. This indicates that the basic damping gain is adjusted using a composite risk index so that, under normal operating conditions, Smaller Maintain at a low value; when the risk increases, It will increase rapidly and nonlinearly, providing powerful active damping for the system.

[0079] It should be noted that, as Figure 2 This diagram illustrates the dynamic changes of the valve control signal and the adaptive damping gain. The adaptive active damping gain achieves smooth fusion of multiple modes through a unified framework. Under low-risk conditions, it acts as a precise energy follower controller, while under high-risk conditions, it automatically transforms into a powerful vibration suppressor, greatly improving the safety and robustness of the system.

[0080] At this point, a hydraulic valve control signal capable of adaptively adjusting behavior based on the risk level has been obtained.

[0081] S4: Execute control commands and correct the control inverse model online based on the energy tracking deviation.

[0082] It should be noted that the control law relies on the accuracy of the inverse model that converts the target energy into a valve port signal. This inverse model physically corresponds to the complex flow characteristics of the hydraulic valve and the energy conversion efficiency of the hydraulic cylinder, and it drifts with factors such as oil temperature, pressure, and wear. A fixed, offline-calibrated inverse model cannot adapt to this time-varying characteristic. The instantaneous power output of the hydraulic system can be calculated in real time. The deviation between this actual power and the target power is the most direct indicator of the accuracy of the inverse model. Therefore, this invention proposes an adaptive mechanism based on online model correction.

[0083] It should be noted that, in order to enable the inverse model to be automatically calibrated during use, considering that gradient descent is an effective way to iteratively optimize neural network parameters, we can construct an optimization problem with the square of the energy tracking deviation as the loss function, and update the network weights of the inverse model along the negative gradient direction.

[0084] Specifically, the hydraulic valve control signal is sent to the hydraulic valve for execution.

[0085] Preferably, based on the energy tracking deviation, the control inverse model is adaptively corrected online, including:

[0086] The difference between the actual output power and the target power is denoted as the energy tracking deviation.

[0087] Inverse model parameters The update satisfies the expression:

[0088] ;

[0089] ;

[0090] In the formula, , The parameters representing the inverse model for the k-th and (k-1)-th control cycles are unitless. The adaptive learning rate for the k-th control cycle is a unitless value. , This represents the normalized energy tracking error for the k-th and j-th control cycles, and is a unitless value. This represents the gradient of the inverse model output with respect to its parameters, and is a unitless value. This represents a preset time window, for example. It consists of 20 control cycles; This indicates the calculation of the variance function. It should be noted that this is an inverse model. parameters The update process expression has unitless values ​​on both sides of the equals sign, so the units on both sides of the expression are consistent.

[0091] It should be noted that this online adaptive correction mechanism enables the controller's inverse model to continuously and automatically adapt to changes in the physical characteristics of the hydraulic system during use, ensuring that the upper-level energy control can be converted into valve action under any working condition, thus achieving optimal performance and maintenance-free operation throughout the entire life cycle of the control system.

[0092] Thus, the present invention completes the anti-pinch and follow-up control of the unloading platform.

[0093] While various embodiments of the invention have been shown and described in this specification, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention.

Claims

1. A method for anti-pinch and follow-up control of an intelligent unloading platform, characterized in that, include: Obtain real-time pressure and displacement data of the main lifting hydraulic cylinder of the unloading platform; Calculate the instantaneous output power of the hydraulic cylinder; The instantaneous output power is subjected to joint time-frequency domain analysis, including: wavelet packet transform of the instantaneous output power, decomposing it into preset high-frequency and mid-high-frequency bands; the composite risk index is the sum of the standard deviations of the high-frequency energy and the mid-high-frequency energy, in order to extract the composite risk index characterizing the system safety risk and the follow-up energy command characterizing the car lifting. The follow-up energy command includes: performing low-pass filtering on the instantaneous output power to separate the low-frequency power component caused by the slow rise and fall of the carriage; and performing time integration on the low-frequency power component. An adaptive energy control strategy is employed to generate the target energy command, including: In the formula, For a moment The target energy command is in joules. ; For a moment The servo energy command, measured in joules. ; For a moment The adaptive active damping gain, in Newton-seconds. ; For a moment The actual speed of the platform, in meters per second. The strategy includes a feedforward term for tracking servo energy commands and a feedback term whose gain is dynamically adjusted by a composite risk index to provide active damping. Adaptive active damping gain includes: ;in, This is the basic damping gain, measured in Newton-seconds. ; For a moment The composite risk index, in watt-square seconds. ; and These are the historical mean and standard deviation of the composite risk index under normal operating conditions, in watt-square seconds. ; It is the hyperbolic tangent function; To preset a tiny value, avoid the denominator being 0; The target energy command is converted into a hydraulic valve control signal through the control inverse model and then executed.

2. The intelligent unloading platform anti-pinch and follow-up control method according to claim 1, characterized in that, Calculating the instantaneous output power of the hydraulic cylinder includes: multiplying the pressure difference between the rod chamber and the rodless chamber of the hydraulic cylinder by the effective area of ​​the piston to obtain the net output force, and then multiplying the net output force by the real-time movement speed of the piston.

3. The intelligent unloading platform anti-pinch and follow-up control method according to claim 2, characterized in that, The instantaneous output power of the hydraulic cylinder satisfies the expression: ； In the formula, Indicates time Instantaneous output power, measured in watts. ; This represents the effective piston area of ​​the rodless chamber in a hydraulic cylinder, expressed in square meters. ; This represents the effective piston area of ​​the rod chamber in a hydraulic cylinder, expressed in square meters. ; , Indicates time The pressure in the rodless and rod chambers of a hydraulic cylinder, measured in Pascals. ; Indicates time The piston displacement, in meters. .

4. The intelligent unloading platform anti-pinch and follow-up control method according to claim 1, characterized in that, The control inverse model is a neural network model trained based on historical calibration data, used to establish the mapping relationship between target energy commands and hydraulic valve control signals.

5. The intelligent unloading platform anti-pinch and follow-up control method according to claim 4, characterized in that, The method further includes: performing online adaptive correction on the control inverse model, including: calculating the energy tracking deviation between the actual output power of the hydraulic cylinder and the target energy command; and updating the network parameters of the neural network model online using a gradient descent algorithm based on the energy tracking deviation.

6. The intelligent unloading platform anti-pinch and follow-up control method according to claim 1, characterized in that, The method also includes an offline calibration step for acquiring historical data required to construct the control inverse model and determine the time-frequency domain analysis parameters, including: performing a step-by-step loading test in a stationary state of the platform to obtain static load-pressure-displacement characteristic data; performing a standard lifting and lowering cycle at different speeds under no-load conditions to obtain dynamic reference data; and applying a standard impact to the platform using a force control device to obtain typical collision event data.

Images