Servo valve multi-loop feedback and energy recovery method and system

Abstract

The application relates to the technical field of automatic control, and discloses a servo valve multi-loop feedback and energy recovery method and system. The method comprises the following steps: differentially counting the polarity sequence of a current inner loop driving instruction and performing a ratio operation with a time constant of an accumulator liquid filling to obtain a recovery window quality coefficient; triggering equivalent time sequence rearrangement based on the coefficient; performing recovery control on a braking section counter electromotive force through a PWM inverter algorithm; and adaptively correcting rearrangement parameters according to actual recovery electric quantity and position error. The application significantly improves the mechanical energy to electric energy recovery efficiency under the braking condition while maintaining the positioning accuracy of a valve core.

Description

Technical Field

[0001] This application relates to the field of automatic control technology, and in particular to a method and system for multi-loop feedback and energy recovery of a servo valve. Background Technology

[0002] Electro-hydraulic servo control systems use liquid as the transmission medium and servo valves as the interface between the electrical control and hydraulic actuators. They control the speed and position of the hydraulic components by adjusting the valve opening. In the drive control of precision hydraulic servo valves, a three-level cascade closed-loop control architecture consisting of an outer position loop, a middle speed loop, and an inner current loop is widely used. Each loop operates with different control cycles and bandwidths, sequentially converting position deviation into speed commands, speed deviation into current commands, and current deviation into drive signals via hysteresis or PWM modulation. This, combined with feedback signals from multiple sensors, completes closed-loop control of the valve core displacement. Simultaneously, recyclable energy generated by the hydraulic system during actuator braking and deceleration, and the release of potential energy from the lifting mechanism, is typically recovered through accumulator charging or regenerative braking via a drive motor. This recovered energy is then released and reused when needed by the system after coordinated control, reducing overall system energy consumption.

[0003] However, existing servo valve multi-loop feedback control systems and energy recovery subsystems lack a unified coordination mechanism, operating independently on the time axis. At the multi-loop feedback control level, the hysteresis control strategy of the inner current loop only focuses on locking the actual current near the reference value, without any statistical analysis or management of the frequency distribution of drive signal polarity switching. Energy recovery triggering generally relies on the post-event detection of physical quantities such as pressure sensors or torque signals; disturbances must first act on the controlled object and generate observable deviations before compensation can be provided, resulting in an inherent control-physical time lag. At the energy recovery level, the accumulator or energy storage unit needs to meet the physical constraint of the shortest continuous filling time to complete one effective filling or charging cycle. Existing solutions do not consider the matching relationship between this constraint and the timing of the inner current loop output; there is no timing binding relationship between the energy recovery triggering timing and the control output cycle.

[0004] Under high-precision tracking conditions, the outer position loop continuously drives the middle speed loop and the inner current loop for high-frequency correction to eliminate minute errors. Consequently, the polarity of the drive signal in the inner current loop alternates frequently between braking and driving. Each continuous braking cycle is much shorter than the minimum time required for the accumulator or energy storage unit to complete effective filling, resulting in severe fragmentation of the braking window. The energy recovery device cannot complete effective filling or charging within each fragmented window before being interrupted by the next driving mode. Most of the mechanical energy generated by braking is lost as heat rather than recovered. The inherent logic of this contradiction lies in the fact that the pursuit of precision by the control system itself is the creator of fragmented recovery windows—the higher the control precision, the more frequent the polarity switching, the more severe the fragmentation, and the lower the recovery efficiency, forming an irreconcilable systemic paradox between control precision and energy recovery efficiency. Existing technologies have never incorporated the matching relationship between the polarity switching frequency of the inner current loop and the accumulator filling time constant into the design considerations of the control strategy, nor have they proposed a technical path for actively managing the timing arrangement of the inner current loop drive sequence without changing the macroscopic position tracking precision. Summary of the Invention

[0005] This application provides a method and system for multi-loop feedback and energy recovery in servo valves, which solves the technical problems of fragmentation of the energy recovery window caused by high-frequency polarity switching of the current inner loop in existing multi-loop feedback control systems of servo valves, and the inability to simultaneously achieve control accuracy and energy recovery efficiency. It significantly improves the efficiency of mechanical energy to electrical energy recovery under braking conditions without reducing the valve core positioning accuracy.

[0006] In a first aspect, this application provides a method for multi-loop feedback and energy recovery in a servo valve, the method comprising: Step S1: The multi-source feedback signals collected by the displacement sensor, velocity sensor and current sensor are sequentially passed through the position outer loop, velocity middle loop and current inner loop for cascade deviation adjustment to obtain the current inner loop drive command; Step S2: Perform differential statistics on the polarity sequence of the inner current loop drive command to obtain the quantized value of the braking drive switching frequency; perform a ratio calculation between the quantized value of the braking drive switching frequency and the accumulator charging time constant to obtain the recovery window quality coefficient; when the recovery window quality coefficient is lower than the shaping trigger threshold, perform equivalent timing rearrangement on the polarity sequence based on the position outer loop error margin to obtain the rearranged drive sequence and the set of braking segment start and end times; Step S3: Output the rearranged drive sequence to the inner current loop, and under the deceleration or reverse motion conditions corresponding to the start and end times of the braking segment, use the PWM inverter algorithm to control the mechanical energy to electrical energy recovery of the back electromotive force, and obtain the electrical energy recovery command. Step S4: Based on the actual recovered electricity and the maximum position error after the execution of the energy recovery command, adaptively correct the shaping trigger threshold and the minimum number of sampling points for continuous braking to obtain the corrected rearrangement parameters, and use the corrected rearrangement parameters for the next round of equivalent timing rearrangement.

[0007] Secondly, this application provides a servo valve multi-loop feedback and energy recovery system, the servo valve multi-loop feedback and energy recovery system comprising: The adjustment module is used to sequentially adjust the multi-source feedback signals collected by the displacement sensor, speed sensor and current sensor through the outer position loop, the middle speed loop and the inner current loop to obtain the drive command for the inner current loop. The calculation module is used to perform differential statistics on the polarity sequence of the inner current loop drive command to obtain the quantized value of the braking drive switching frequency; to perform a ratio calculation between the quantized value of the braking drive switching frequency and the accumulator charging time constant to obtain the recovery window quality coefficient; when the recovery window quality coefficient is lower than the shaping trigger threshold, to perform equivalent timing rearrangement of the polarity sequence based on the position outer loop error margin to obtain the rearranged drive sequence and the set of braking segment start and end times; The control module is used to output the rearranged drive sequence to the inner current loop, and under the deceleration or reverse motion conditions corresponding to the start and end times of the braking segment, to control the mechanical energy to electrical energy recovery of the back electromotive force through the PWM inverter algorithm to obtain the electrical energy recovery command. The correction module is used to adaptively correct the shaping trigger threshold and the minimum number of sampling points for continuous braking based on the actual recovered electricity and the maximum position error after the execution of the energy recovery command, so as to obtain the corrected rearrangement parameters, and use the corrected rearrangement parameters for the next round of equivalent timing rearrangement.

[0008] Thirdly, a servo valve multi-loop feedback and energy recovery device is provided, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor invokes the instructions in the memory to cause the servo valve multi-loop feedback and energy recovery device to perform the aforementioned servo valve multi-loop feedback and energy recovery method.

[0009] Fourthly, a computer-readable storage medium is provided, wherein instructions are stored therein, which, when executed on a computer, cause the computer to perform the aforementioned servo valve multi-loop feedback and energy recovery method.

[0010] In the technical solution provided in this application, the multi-source feedback signals collected by displacement sensors, velocity sensors, and current sensors are cascaded through position outer loop, velocity middle loop, and current inner loop to obtain the current inner loop drive command. Based on this, the polarity sequence of the current inner loop drive command is differentially statistically analyzed to obtain the quantized value of the braking drive switching frequency. This transforms the originally invisible polarity switching behavior inside the control system into a quantifiable and evaluable control domain statistic. This statistic does not depend on any new physical sensors and is extracted entirely from existing control signals, providing a new criterion dimension for energy recovery decisions without increasing hardware costs or system complexity. The technical feature of obtaining the recovery window quality coefficient by comparing the quantified value of the braking drive switching frequency with the accumulator filling time constant establishes a previously non-existent quantitative correlation between control domain statistics and hydraulic physical parameters in the specific application field of servo valve multi-loop feedback and energy recovery. This ratio calculation places the micro-switching behavior of the current inner loop and the macro-filling capacity of the accumulator on the same dimensionless scale for comparison. This enables the control system to perceive the fragmentation of the recovery window from its own output behavior for the first time, rather than passively waiting for physical sensors to provide feedback on disturbance information before responding. This fundamentally changes the existing architecture where the control subsystem and the energy recovery subsystem are isolated and operate independently.

[0011] When the quality coefficient of the recovery window is lower than the shaping trigger threshold, the polarity sequence is rearranged based on the position outer loop error margin to obtain the rearranged drive sequence and the set of start and end times of the braking segment. This technical feature enables a previously unproposed operation in the precision positioning control scenario of servo valves. Under the constraint of strictly maintaining the macro position tracking accuracy, the output timing of the current inner loop is actively rearranged to integrate the fragmented braking drive switching sequence into a continuous braking segment. This ensures that the duration of each braking segment meets the minimum time requirement for the accumulator or energy storage unit to complete effective liquid filling. Thus, the systemic bottleneck of recovery window fragmentation is eliminated at the control algorithm level, rather than passively adapting through hardware means such as increasing the accumulator volume or reducing the liquid filling time constant. The technology involves rearranging the drive sequence output to the inner current loop and using a PWM inverter algorithm to control the mechanical energy to electrical energy recovery from the back electromotive force under the deceleration or reverse motion conditions corresponding to the start and end times of the braking segment. This technology binds the control output and energy recovery execution to the same set of start and end times of the braking segment, ensuring that the braking and recovery cycles are precisely aligned on the time axis. Under the constraint of this binding relationship, the PWM inverter algorithm activates the power generation mode only within the continuous braking segment, avoiding the switching losses and electromagnetic interference caused by the frequent switching of the inverter arm between the electric and power generation modes under fragmented switching conditions. The technical feature of adaptively correcting the shaping trigger threshold and the minimum number of sampling points for continuous braking based on the actual recovered electricity and the maximum position error after the execution of the energy recovery command, and then using the corrected rearrangement parameters for the next round of equivalent timing rearrangement, uses energy recovery efficiency and position control accuracy as dual closed-loop feedback indicators to drive the online optimization of rearrangement parameters. This allows the shaping trigger threshold and the minimum number of sampling points for continuous braking to automatically approach their respective equilibrium values ​​with the dynamic changes in the working conditions, rather than relying on fixed parameters tuned offline. In practical applications where the working environment of the servo valve is variable and the load characteristics drift over time, it has continuous working condition adaptability, ensuring that control accuracy and energy recovery efficiency converge synchronously across the entire working condition range. Attached Figure Description

[0012] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a schematic diagram of one embodiment of the servo valve multi-loop feedback and energy recovery method in this application. Figure 2 This is a schematic diagram comparing the changes in accumulated recovered electricity over time before and after time sequence rearrangement in an embodiment of this application; Figure 3This is a schematic diagram illustrating the adaptive convergence of the shaping trigger threshold and the minimum number of continuous braking sampling points with the number of statistical window rounds in the embodiments of this application. Detailed Implementation

[0014] This application provides a method and system for multi-loop feedback and energy recovery in a servo valve. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0015] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the servo valve multi-loop feedback and energy recovery method in this application includes: Step S1: The multi-source feedback signals collected by the displacement sensor, velocity sensor and current sensor are sequentially passed through the position outer loop, velocity middle loop and current inner loop for cascade deviation adjustment to obtain the current inner loop drive command; Specifically, in the servo valve multi-loop feedback system, a displacement sensor is installed in the valve core stroke direction and collects the actual position of the valve core at a sampling interval of 1 to 2 milliseconds. A speed sensor collects the valve core movement speed at a sampling interval of 0.5 to 1 millisecond, and a current sensor collects the actual current of the drive coil at a sampling interval of 50 microseconds. The three signals constitute a multi-source feedback signal. The control cycles of the position outer loop, speed middle loop, and current inner loop are set to 2 milliseconds, 1 millisecond, and 50 microseconds, respectively. The three form a cascade structure. The position outer loop outputs the quantity based on the position deviation as the input command for the speed middle loop. The speed middle loop outputs the quantity based on the speed deviation as the input command for the current inner loop. The hysteresis controller of the current inner loop determines whether the deviation exceeds the hysteresis width (5 mA in this case) within a 50-microsecond cycle to determine whether the drive voltage should be flipped, and finally generates the current inner loop drive command. The so-called cascade deviation regulation refers to the fact that each loop only processes the difference between the feedback signal of its own level and the output command of the previous level. The bandwidth of each loop increases sequentially, with the current inner loop having the highest bandwidth and the fastest response, and the position outer loop having the lowest bandwidth and the strongest stability. This structure enables the valve core displacement control accuracy to reach the level of 0.02 mm.

[0016] Step S2: Perform differential statistics on the polarity sequence of the inner current loop drive command to obtain the quantized value of the braking drive switching frequency; perform a ratio calculation between the quantized value of the braking drive switching frequency and the accumulator charging time constant to obtain the recovery window quality coefficient; when the recovery window quality coefficient is lower than the shaping trigger threshold, perform equivalent timing rearrangement on the polarity sequence based on the position outer loop error margin to obtain the rearranged drive sequence and the set of braking segment start and end times; Specifically, the inner current drive command outputs a polarity value every 50 microsecond control cycle, forming a polarity sequence within a 10-millisecond statistical window (i.e., 200 sampling points). In this sequence, positive polarity indicates that the drive current outputs in the forward direction, driving the valve core to move towards the target, while negative polarity indicates that the current outputs in the reverse direction, generating braking force. The so-called braking drive switching frequency quantization value is the total number of times the polarity of adjacent points flips in these 200 sampling points. This value directly reflects the intensity of the back-and-forth switching between braking and drive in the current tracking task of the control system. The recovery window quality coefficient is the ratio of the average duration of a single continuous braking to the accumulator charging time constant. The accumulator charging time constant is determined by physical parameters such as the accumulator's rated volume, airbag pre-charge pressure, hydraulic oil bulk modulus, and charging pipeline flow resistance at the system's factory calibration. Its value varies under different accumulator pressures, generally ranging from 1.5 to 3 milliseconds, and is stored in the controller using a lookup table. The initial value of the reshaping trigger threshold is set to 0.6, meaning that equivalent timing rearrangement is triggered when the average braking window duration is less than 60% of the accumulator filling time constant. Physically, this threshold represents the critical proportion of effective filling of the accumulator within a single braking window. A value below this means that each braking window is too short, the accumulator is not fully filled, and the driving mode is interrupted, resulting in energy loss as heat. The core physical constraint of equivalent timing rearrangement is that the total number of negative and positive sampling points within the same statistical window remains unchanged before and after rearrangement. This ensures the conservation of the time integral of the electromagnetic force on the valve core within the window—the valve core's displacement trajectory on a macroscopic time scale does not deviate from the original trajectory due to rearrangement; the deviation only occurs at the microscopic sampling point arrangement level, and this microscopic deviation is strictly constrained by the position outer loop error margin.

[0017] Step S3: Output the rearranged drive sequence to the inner current loop, and at the start and end times of the braking segment, combine the corresponding deceleration or reverse motion conditions, and use the PWM inverter algorithm to control the mechanical energy to electrical energy recovery of the back electromotive force to obtain the electrical energy recovery command. Specifically, when the drive sequence is rearranged and output to the inner current loop, the hysteresis controller performs current tracking point by point, with the control logic being completely consistent with the original sequence, the only difference being the polarity order of the input sequence. Each pair of start and end times in the set of braking segment start and end times defines the boundary of a continuous braking segment. 0.2 milliseconds before the start time of each braking segment (this value is the hardware response time required for the inverter arm to switch from electric mode to generator mode, determined by the turn-off delay of the IGBT switch and the conduction time of the freewheeling diode), the controller sends an enable signal to the inverter arm, turning off the upper IGBT and turning on the lower IGBT, switching the drive loop to the generator topology. Back electromotive force (EMF) refers to the voltage induced by the drive coil cutting magnetic lines of force during deceleration or reverse movement of the valve core. This voltage is opposite in direction to the drive voltage. In electric mode, it is absorbed by the drive power supply as heat loss, while in generator mode, it is rectified into DC voltage by the freewheeling diode of the inverter arm. The recovered voltage is then adjusted to the charging voltage range of the energy storage unit through PWM chopping control. The recovered current is determined by the duty cycle of the inverter arm and the coil inductance. The specific meaning of the PWM inverter algorithm here is that during the power generation mode, the duty cycle of the inverter bridge arm is continuously adjusted with the same 50 microsecond control cycle as the inner current loop, so that the recovery current is kept within the upper limit of the charging current allowed by the energy storage unit, so as not to damage the energy storage unit due to overcurrent or waste the recovery window due to too low duty cycle.

[0018] Step S4: Based on the actual recovered electricity and the maximum position error after the execution of the energy recovery command, adaptively correct the shaping trigger threshold and the minimum number of sampling points for continuous braking to obtain the corrected rearrangement parameters, and use the corrected rearrangement parameters for the next round of equivalent timing rearrangement.

[0019] Specifically, the adaptive correction targets two parameters: the shaping trigger threshold and the minimum number of sampling points for continuous braking. The adjustment coefficient for the shaping trigger threshold is set to a downward adjustment coefficient of 0.95 and an upward adjustment coefficient of 1.05, meaning each correction increment is 5% of the current threshold. This magnitude is chosen because the rate of change of the accumulator charging time constant within the normal operating range generally does not exceed 8% of each statistical window. A 5% adjustment step size can track the target within 3 to 5 window periods without causing oscillations. The step size for the minimum number of sampling points for continuous braking is set to 2 sampling points, meaning each correction increases the minimum braking duration by 100 microseconds. The ratio of this step size to the accumulator charging time constant is within the range of 3% to 7%, ensuring sufficiently fine correction granularity without overshoot. When the maximum position error exceeds the specified position accuracy value for three consecutive statistical windows, the controller directly shuts down the timing shaping function and restores the original output sequence of the current inner loop until the maximum position error falls below 80% of the accuracy specification value before reactivation.

[0020] In one specific embodiment, step S1 includes: The actual position signal, actual speed signal and actual current signal of the valve core are collected by displacement sensor, speed sensor and current sensor respectively to form a multi-source feedback signal; The position deviation value is obtained by subtracting the position command from the actual position signal. After proportional-integral-derivative adjustment and amplitude limiting by the position outer loop controller, the speed command is output. Based on the deviation between the speed command and the actual speed signal, the speed feedforward component obtained by the derivative of the position command is superimposed, and after adaptive proportional-integral adjustment by the speed loop controller, the current command is output. The deviation between the current command and the actual current signal is determined by the hysteresis width. When the absolute value of the deviation exceeds the preset hysteresis width, the polarity of the drive voltage is flipped; otherwise, the current output is maintained, and the inner loop drive command of the current is obtained.

[0021] Specifically, after receiving the position deviation value, the outer loop controller linearly amplifies the deviation with a gain coefficient of 80 (in millimeters per second per millimeter deviation), the integral term eliminates the accumulated deviation with a time constant of 50 milliseconds, and the derivative term predicts and compensates for the deviation change trend with a time constant of 2 milliseconds. The three outputs are superimposed and then truncated to a range of ±300 millimeters per second to output the speed command. The reason for introducing a derivative term in the outer loop but not in the middle loop is that the outer loop control cycle is 2 milliseconds, which has a narrower bandwidth compared to the 50 microseconds of the inner current loop. The derivative term can effectively suppress position overshoot without amplifying high-frequency noise within this bandwidth. The middle loop control cycle is 1 millisecond, which already has a higher bandwidth; adding a derivative term would amplify the quantization noise of the speed sensor, causing glitches in the current command. The 300 millimeters per second limit is based on the maximum mechanical stroke speed of the servo valve spool. Speed ​​commands exceeding this value cannot be physically executed by the spool and would instead lead to integral saturation. The velocity feedforward component is the theoretical velocity value obtained by taking the first derivative of the position command with respect to time. This value is directly superimposed on the input of the velocity deviation adjustment. Its engineering purpose is that when the position command undergoes a step or ramp change, it does not need to wait for the position deviation to accumulate to a large enough size before generating a velocity command. Instead, the initial response direction and amplitude of the velocity loop are directly given by the rate of change of the position command itself, thereby shortening the time interval between the occurrence of the position deviation and the establishment of the velocity response.

[0022] In the adaptive proportional-integral (PI) regulation of the speed loop controller, the proportional gain is not a fixed constant, but switches between high and low gain ranges based on the absolute value of the current speed deviation: when the absolute value of the speed deviation is greater than 50 mm / s, the proportional gain is 3.5 (mA / mm / s deviation), corresponding to the need for rapid pullback from large deviations; when the absolute value of the speed deviation is less than 10 mm / s, the proportional gain drops to 1.2 to avoid current command oscillation caused by excessive gain when approaching the target speed; a linear interpolation transition is used between the two thresholds. The time constant of the integral term is fixed at 20 milliseconds, responsible for eliminating steady-state speed errors. After the current command output from the speed loop enters the inner current loop, the hysteresis width determination process involves subtracting the current command from the actual current signal collected by the current sensor to obtain the current deviation, and determining whether the absolute value of this deviation exceeds the preset hysteresis width of 5 mA—if it exceeds, the polarity of the drive voltage is reversed, i.e., if the current drive voltage is positive, it switches to negative, and if it is negative, it switches to positive; if it does not exceed, the current drive voltage polarity remains unchanged. The hysteresis width of 5 mA is chosen because the rated operating current of the servo valve drive coil is 1000 mA. 5 mA corresponds to 0.5% of the rated current. In engineering practice, this ratio can control the current ripple within the allowable range of the drive coil's heat dissipation without causing the switching frequency to rise above the IGBT power transistor's limit switching frequency due to an excessively narrow hysteresis. The toggle operation is performed once every 50 microseconds of the inner current loop control cycle, and the toggle result is the polarity value corresponding to that sampling point in the inner current loop drive command.

[0023] In one specific embodiment, differential statistics are performed on the polarity sequence of the current inner loop drive command to obtain the quantized value of the braking drive switching frequency, including: According to the control cycle of the inner current loop, the driving voltage direction of the inner current loop driving command is extracted point by point, and the positive driving is marked as positive polarity and the reverse braking is marked as negative polarity, thus forming a polarity sequence. Perform a pairwise difference operation on two adjacent sampling points in the polarity sequence. When the polarities of the two adjacent points are different, record a switching event. When the polarities of the two adjacent points are the same, do not record. This will give you a switching event sequence. The switching event sequence is summed within a preset statistical window to obtain the quantized value of the braking drive switching frequency.

[0024] Specifically, the polarity sequence is marked by the direction of the drive voltage output by the inner-loop hysteresis controller in each 50-microsecond control cycle, not the actual direction of the drive current. This distinction is crucial in engineering because the drive coil has inductance, and the reversal of the current direction is delayed by hundreds of microseconds compared to the reversal of the voltage direction. If the current direction is used as the marking basis, the polarity sequence will lag behind the actual decision moment of the controller, resulting in an underestimation of the switching frequency statistics. Positive polarity corresponds to the positive drive voltage output by the hysteresis controller, i.e., the direction in which the valve core moves towards the target position; negative polarity corresponds to the reverse drive voltage output, i.e., the direction in which the braking electromagnetic force is generated. Each control cycle corresponds to exactly one polarity mark value, and 200 consecutive control cycles constitute a polarity sequence of length 200. The specific operation rule for the pairwise differential operation is to take the absolute value of the difference between the polarity values ​​of two adjacent points and divide it by 2. When the previous point is positive and the next point is negative, the absolute value of the difference is 2, and dividing by 2 equals 1, recording a switching event; when the polarities of the two points are the same, the difference is 0, and no record is made. The switching event sequence is a Boolean array of length 199, where elements with a value of 1 mark the precise location where the polarity reversal occurs.

[0025] The preset statistical window duration is set to 10 milliseconds, corresponding to the number of sampling points for 200 current inner loop control cycles. The selection of 10 milliseconds is based on the intersection of two constraints: firstly, the control cycle of the servo valve position outer loop is 2 milliseconds, and the statistical window should cover at least 5 position outer loop cycles to reflect the complete trend of the valve core's movement state; 5 cycles correspond exactly to 10 milliseconds. Secondly, the typical range of the accumulator charging time constant is 1.5 to 3 milliseconds, and the statistical window should cover at least 3 charging time constants to ensure that the switching frequency statistics within the window have a resolution matching the accumulator charging capacity; 3 times the maximum charging time constant is 9 milliseconds, rounded up to 10 milliseconds. The sum of the elements with a value of 1 in the switching event sequence within this 10-millisecond window is the quantized value of the braking drive switching frequency, with dimensions in units of 'times', ranging from 0 to 199. 0 corresponds to continuous unidirectional valve core movement without any polarity reversal throughout the window, while 199 corresponds to the theoretical extreme condition where a polarity reversal occurs once per control cycle. In practical engineering, the quantization value usually falls between 40 and 120 in high-precision tracking scenarios of sinusoidal trajectories, and between 5 and 20 in step positioning scenarios.

[0026] In one specific embodiment, the ratio of the quantized value of the braking drive switching frequency to the accumulator filling time constant is calculated to obtain the recovery window quality coefficient, including: The total number of negative polarity sampling points is counted from the polarity sequence. Combined with the quantized value of the braking drive switching frequency and the current inner loop control cycle, the average duration of a single continuous braking is calculated. By referring to the pre-calibrated pressure-time constant lookup table based on the current pressure value of the accumulator, the accumulator filling time constant under the current operating conditions can be obtained; Dividing the average duration of a single continuous braking operation by the accumulator filling time constant yields the recovery window quality coefficient.

[0027] Specifically, the total number of negative polarity sampling points is obtained by scanning point by point in the polarity sequence constructed in the previous step and accumulating the number of sampling points with negative polarity values. This operation is completed within the same 10-millisecond statistical window as the differential statistics of the braking drive switching frequency quantization value, without the need for additional data acquisition. The calculation logic for the average duration of a single continuous braking is to divide the total number of negative polarity sampling points by the braking drive switching frequency quantization value. The quotient represents the average number of sampling points included in each continuous braking segment. This quotient is then multiplied by the current inner loop control cycle of 50 microseconds to convert the number of sampling points into time dimensions. To illustrate the operational relationship between these three, consider a specific data correlation: if there are 80 negative polarity sampling points out of 200 sampling points in the statistical window, and the differential statistics result in a braking drive switching frequency quantization value of 40 times, then the average number of sampling points included in each continuous braking is 80 divided by 40, which equals 2. Multiplying this by 50 microseconds gives an average duration of 100 microseconds for a single continuous braking. This 100 microseconds means that the valve core's movement in each braking direction is only maintained for 100 microseconds before being interrupted by the driving direction, while the shortest continuous time required for the accumulator to complete one effective filling is much greater than this value.

[0028] The pressure-time constant lookup table is generated through offline experimental calibration during the factory commissioning phase of the servo valve system. The calibration method involves opening the filling valve and recording the time it takes for the accumulator pressure to rise from its initial value to the target recovery pressure under different initial accumulator pressure conditions. This time is the filling time constant for the corresponding pressure point. The input variable of the lookup table is the current accumulator pressure value with an accuracy of 0.1 MPa, and the output variable is the filling time constant with an accuracy of 0.1 milliseconds. The data points in the table typically cover the entire operating pressure range of the accumulator, and linear interpolation is used between adjacent data points to obtain the time constant corresponding to the intermediate pressure value. The accumulator filling time constant is affected by four physical parameters: the rated volume of the accumulator, the pre-charge pressure of the air bladder, the bulk modulus of the hydraulic oil, and the flow resistance of the filling pipeline. Typical values ​​range from 1.5 to 3 milliseconds under different pressure conditions. Dividing the calculated average duration of a single continuous braking cycle by the accumulator filling time constant obtained from the lookup table yields a dimensionless ratio, which is the recovery window quality coefficient. Continuing with the previous data correlation, if the average duration of a single continuous braking event is 100 microseconds and the accumulator charging time constant under the current operating condition is 2 milliseconds (2000 microseconds), then the recovery window quality coefficient is 100 divided by 2000, which equals 0.05, far below 1. This indicates that the duration of the current braking window is only one-twentieth of the time required for the accumulator to complete effective charging. The accumulator cannot complete effective charging within each braking window, and most of the mechanical energy generated by braking is dissipated as heat rather than recovered. When the recovery window quality coefficient is equal to or greater than 1, it indicates that the duration of each continuous braking window has reached or exceeded the accumulator charging time constant, and the accumulator can complete effective charging within each braking window without triggering timing shaping.

[0029] In one specific embodiment, when the quality coefficient of the recovery window is lower than the shaping trigger threshold, the polarity sequence is rearranged in an equivalent time sequence based on the position outer loop error margin to obtain the rearranged driving sequence and the set of start and end times of the braking segment, including: Subtract the absolute value of the current position outer loop error from the position accuracy specification value specified by the process to obtain the position outer loop error margin; The minimum number of sampling points for continuous braking is obtained by rounding up the quotient of the accumulator charging time constant and the current inner loop control cycle. The total number of negative polarity sampling points is then divided by the minimum number of sampling points for continuous braking and rounded down to obtain the number of braking segments. According to the number of braking segments, the scattered negative polarity sampling points in the polarity sequence are grouped into continuous braking segments of equal length, and the positive polarity sampling points are grouped into continuous driving segments. They are arranged in an alternating order of braking segments first and driving segments last to form a rearranged driving sequence. The position deviation increment caused by the rearranged drive sequence is checked based on the valve core mass and hydraulic coefficient. When the position deviation increment exceeds the position outer loop error margin, the number of braking segments is reduced and rearranged until the position deviation increment falls within the position outer loop error margin range. The set of braking segment start and end times, consisting of the rearranged drive sequence and the start and end times of each continuous braking segment, is then output.

[0030] Specifically, the calculation of the position outer loop error margin is based on the position accuracy specification value stipulated by the process. This specification value is determined by the positioning accuracy requirements of the specific hydraulic actuator served by the servo valve. In high-precision hydraulic servo valve applications, a typical value is 0.05 mm, which is written into the controller register by the host computer during the system initialization phase. The current position outer loop error refers to the absolute value of the difference between the actual position collected by the displacement sensor at the end of the current statistical window and the position command. The difference obtained by subtracting this absolute value from 0.05 mm is the position outer loop error margin. Its physical meaning is the maximum additional position offset that can be introduced by the timing reordering operation under the current positioning error level without violating the process accuracy requirements.

[0031] The minimum number of sampling points for continuous braking is calculated by dividing the accumulator charging time constant by the current inner loop control cycle of 50 microseconds and rounding up. This rounding ensures that the duration of each continuous braking segment is strictly no less than the accumulator charging time constant. For example, with a 2-millisecond charging time constant, this value corresponds to 40 sampling points. The number of braking segments is obtained by dividing the total number of negative polarity sampling points by the minimum number of sampling points for continuous braking and rounding down. Rounding down means that if the total number of negative polarity sampling points is not divisible by the minimum number of sampling points for continuous braking, a small number of remaining negative polarity sampling points will not be able to form a complete braking segment. These remaining sampling points are not included in the aggregation; they are retained in their original positions and output according to their original polarity, without affecting the overall rearrangement logic. Taking a total of 80 negative polarity sampling points and a minimum of 40 sampling points for continuous braking as an example, there are 2 braking segments, each with exactly 40 sampling points, with no remainder; if the total number of negative polarity sampling points is 90, the number of braking segments is still 2 segments with 40 points each, and the remaining 10 negative polarity sampling points are scattered and retained in the gap between the driving segments and output according to their original polarity.

[0032] The construction rule for rearranging the drive sequence is to group the negative polarity sampling points, originally scattered among 200 sampling points, into several continuous blocks of equal length according to the number of braking segments. Positive polarity sampling points are similarly grouped into continuous blocks, and then spliced ​​together in an alternating order of braking segments first, followed by drive segments. The physical basis of equivalence lies in the fact that the total number of negative polarity sampling points and the total number of positive polarity sampling points remain unchanged throughout the entire statistical window before and after rearrangement. The magnitude of the electromagnetic force experienced by the valve core drive coil at each sampling point is determined by the current tracking accuracy of the hysteresis controller and is independent of the sampling point's position within the window. Therefore, a unchanged total number of negative polarity points means that the time integral of the electromagnetic force in the braking direction remains unchanged, and a unchanged total number of positive polarity points means that the time integral of the electromagnetic force in the driving direction remains unchanged. The position reached by the valve core at the end of the entire 10-millisecond window is consistent with the endpoint position before rearrangement.

[0033] However, at the instantaneous level within the window, due to the concentrated arrangement of the continuous braking segments, the valve core decelerates more during the braking segment than during the original dispersed braking, causing the instantaneous position of the valve core to deviate from the original trajectory, resulting in an incremental position deviation. This increment is verified using a simplified valve core dynamics model. The valve core mass is taken as 0.05 kg, a typical value for a servo valve core with a spring preload assembly, and the hydraulic coefficient is taken as 15 N / mm, a typical value for a medium-diameter servo valve at rated flow. The verification logic is to divide the impulse of the electromagnetic braking force during the duration of the continuous braking segment by the valve core mass to obtain the velocity change, and then multiply this by half the duration of the braking segment to obtain an approximate value for the position deviation increment.

[0034] When the approximate value exceeds the position outer loop error margin, the number of braking segments is reduced by 1, and the length of each segment is recalculated and rearranged with the new number of braking segments. The position deviation increment is checked again, and this process is iterated until the deviation increment falls within the error margin range. The termination condition of the iteration is that the number of braking segments is at least 1, that is, all negative polarity sampling points are gathered into one braking segment. If the deviation increment still exceeds the error margin at this time, the timing shaping is abandoned in this statistical window, and the original output of the current inner loop is maintained. This condition logically eliminates the possibility of an iteration dead loop. The rearranged drive sequence output after the check is passed is a polarity array of length 200. The set of start and end times of the braking segments is formed by multiplying the start index and end index of each continuous braking segment in the array by 50 microsecond control cycles to convert them into absolute times. Each pair of start and end times defines the time boundary of a continuous braking segment.

[0035] In one specific embodiment, step S3 includes: The rearranged drive sequence is output to the hysteresis controller point by point according to the current inner loop control cycle. The PWM duty cycle direction is determined according to the polarity value of each sampling point and the current tracking is maintained within the preset hysteresis width to obtain the rearranged valve core drive current. Before the start time of each braking segment in the set of braking segment start and end times arrives, a preset response compensation time is sent to the inverter bridge arm to send an inverter enable signal to switch the drive circuit of the current inner loop from electric mode to generator mode. During the effective period of the inverter enable signal, the back electromotive force generated by the valve core drive coil under deceleration or reverse motion conditions is rectified and inverted by the PWM inverter algorithm to obtain the recovery voltage and recovery current. The product of the recovered voltage and the recovered current is integrated over the duration of each braking segment to obtain the energy recovered in a single segment. The energy recovered in a single segment from all braking segments is then used to form an energy recovery command.

[0036] Specifically, when the rearranged drive sequence is output to the hysteresis controller, the controller reads the polarity value point by point according to the 50-microsecond control cycle and determines the direction of the PWM duty cycle for the current cycle accordingly. Positive polarity corresponds to the positive chopping mode in which the high level of the PWM signal drives the upper bridge arm IGBT to conduct and the low level drives the lower bridge arm IGBT to conduct. Negative polarity corresponds to the reverse chopping mode in which the upper and lower bridge arms conduct in opposite order. The specific value of the duty cycle is still adjusted in real time by the hysteresis controller according to whether the current deviation exceeds the 5 mA hysteresis width. The rearrangement operation only changes the polarity direction of each cycle and does not change the current tracking logic of the hysteresis controller itself. The inverter bridge arm consists of two IGBT power switches and two anti-parallel freewheeling diodes forming a half-bridge topology. In motoring mode, the upper and lower switches alternately conduct, driving current flows through the valve core drive coil to generate electromagnetic force. In generator mode, both the upper and lower switches are turned off. During deceleration or reverse movement of the valve core, the back electromotive force generated by the drive coil cutting the magnetic field of the permanent magnet forms a loop through the freewheeling diodes. The current flows from the coil through the lower bridge arm freewheeling diode to the positive terminal of the energy storage unit, and then from the negative terminal of the energy storage unit through the upper bridge arm freewheeling diode back to the coil, completing the conversion of mechanical energy into electrical energy.

[0037] The preset response compensation time is set to 0.2 milliseconds, which is equivalent to 4 current inner loop control cycles. This value is determined by the sum of two hardware parameters: the turn-off delay of the IGBT switch and the forward recovery time of the freewheeling diode from cutoff to conduction. The typical value of the IGBT turn-off delay is 0.1 milliseconds, and the typical value of the freewheeling diode forward recovery time is 0.05 to 0.1 milliseconds. The sum of the two is taken as the upper limit of the engineering margin of 0.2 milliseconds. Sending the inverter enable signal 0.2 milliseconds before the start of each braking segment can ensure that the bridge arm has completed the topology switch from electric mode to generator mode when the braking segment officially begins.

[0038] The amplitude of the back electromotive force (EMF) is proportional to the valve core's movement speed, with the proportionality coefficient being the back EMF constant of the drive coil. The typical back EMF constant of the servo valve drive coil is 0.8 to 1.5 volts per meter per second. Under the condition of a maximum valve core stroke speed of 300 millimeters per second (0.3 meters per second), the peak back EMF is 0.24 to 0.45 volts. During the effective period of the inverter enable signal, the PWM inverter algorithm continuously adjusts the duty cycle of the lower bridge arm IGBT with a 50-microsecond cycle. This duty cycle determines the magnitude of the recovered current. A higher duty cycle results in a faster rate at which the magnetic field energy stored in the coil is released to the energy storage unit through the freewheeling diode, leading to a larger recovered current. However, the recovered current must not exceed the maximum allowable charging current of the energy storage unit. In this case, the energy storage unit uses a supercapacitor module with a maximum allowable charging current of 2 amps. The PWM inverter algorithm compares the current recovered current with the 2-amp upper limit in each control cycle. When the recovered current approaches the upper limit, the duty cycle is reduced; when it moves away from the upper limit, the duty cycle is increased, ensuring that the recovered current always fluctuates within the safe charging range of the energy storage unit.

[0039] The recovery voltage is determined by the current voltage of the energy storage unit. The operating voltage range of the supercapacitor module is 2.5 to 5 volts. The recovery current is determined by the PWM duty cycle and the coil inductance. The instantaneous recovery power for each control cycle is obtained by multiplying the recovery voltage by the recovery current. Multiplying the instantaneous recovery power by 50 microseconds of the control cycle yields the single-cycle recovery power. This single-segment recovery power is accumulated cycle by cycle over all control cycles of a braking segment. The sum of the single-segment recovery powers for all braking segments constitutes the energy recovery command. This command's data structure includes three fields: the sum of the recovery powers for all braking segments, an array of single-segment recovery powers for each braking segment, and an array of the actual start and end times for each braking segment. These three fields participate in the actual recovery efficiency calculation and position error correlation analysis in the next adaptive correction stage.

[0040] Figure 2 This is a schematic diagram comparing the changes in accumulated recovered electricity over time before and after time-series rearrangement in an embodiment of this application. Figure 2 As can be seen, the cumulative recovered energy curve after equivalent timing rearrangement exhibits a clear step-like upward trend. Each step corresponds to the process of the accumulator or energy storage unit centrally receiving and recovering energy during the period when the inverter arm is in power generation mode within a continuous braking segment. The flat sections between the steps correspond to the intervals where the inverter arm is closed and no energy is recovered within the continuous drive segment. The cumulative recovered energy reaches approximately 2500 millijoules over the entire 2000 milliseconds. In contrast, the cumulative recovered energy curve without rearrangement shows a slow, approximately linear increase, accumulating only about 500 millijoules over the entire process. The difference between the two is about 5 times, which directly reflects the phenomenon that a large amount of recoverable energy is lost as heat due to insufficient duration of the fragmented braking window and the accumulator filling time constant. It also demonstrates the technical effect of equivalent timing rearrangement in effectively eliminating this loss mechanism by aggregating fragmented braking segments into continuous braking segments.

[0041] In one specific embodiment, step S4 includes: The actual recovered amount is obtained by summing up the recovered amount of all braking segments in the energy recovery command, and then dividing the actual recovered amount of energy by the theoretical maximum recoverable amount of energy to obtain the actual recovery efficiency. Extract the extreme values ​​of the outer loop position error during the execution of the energy recovery command to obtain the maximum position error; When the actual recovery efficiency is lower than the preset recovery efficiency target value and the maximum position error does not exceed the position accuracy specification value specified by the process, the minimum number of continuous braking sampling points is increased by a preset step and the shaping trigger threshold is multiplied by a preset downward adjustment coefficient to obtain the corrected rearrangement parameters; when the maximum position error exceeds the position accuracy specification value, the shaping trigger threshold is multiplied by a preset upward adjustment coefficient and the number of braking segments is reduced to obtain the corrected rearrangement parameters. The corrected rearrangement parameters replace the currently stored integer trigger threshold and the minimum number of continuous braking sampling points, and are used for the next round of equivalent timing rearrangement.

[0042] Specifically, the actual recovered energy is obtained by accumulating the recovered energy from each braking segment in the energy recovery command, and the dimension of this value is joules. The calculation of the theoretical maximum recoverable energy is based on an ideal assumption—that the braking conditions corresponding to all negative polarity sampling points within the statistical window can be completely recovered, without any interruption in accumulator charging or bridge arm switching losses. The calculation method is to multiply the total number of negative polarity sampling points by the current inner loop control cycle of 50 microseconds to obtain the total braking time, and then multiply it by the average value of the back electromotive force under the braking condition and the rated value of the recovery current. The average back electromotive force is taken as the product of the coil back electromotive force constant and the average valve core speed within the current statistical window, and the rated value of the recovery current is taken as the maximum allowable charging current of the energy storage unit, 2 amperes. The actual recovery efficiency is equal to the actual recovered energy divided by the theoretical maximum recoverable energy, which is a dimensionless ratio between 0 and 1.

[0043] The preset recovery efficiency target value is set at 0.85, meaning that the actual recovered electricity is expected to reach 85% of the theoretical maximum value, with a 15% margin to cover the unavoidable energy loss caused by bridge arm switching losses, coil copper losses, and the forward voltage drop of the freewheeling diode. The extraction range of the maximum position error is strictly limited to the period during the execution of the energy recovery command, that is, from the start time of the first braking segment to the end time of the last braking segment. Within this interval, the maximum absolute value of the position deviation is selected from the point-by-point acquisitions at a 2-millisecond control cycle of the outer position loop.

[0044] The two correction rules have the following priority: position accuracy protection takes precedence over recovery efficiency improvement. At the end of each statistical window, the controller first checks whether the maximum position error exceeds the position accuracy specification of 0.05 mm. If it does, accuracy protection correction is performed regardless of the actual recovery efficiency level. Only if the position accuracy meets the standard will the controller further determine whether recovery efficiency needs to be improved. In accuracy protection correction, multiplying the shaping trigger threshold by a preset upward adjustment coefficient of 1.05 means that the trigger threshold is tightened by 5%. The next round requires a more severe degree of fragmentation to trigger timing shaping. At the same time, the number of braking segments is reduced, shortening the duration of each braking segment and reducing the deviation of the valve core's instantaneous trajectory from the rearrangement.

[0045] The minimum number of sampling points for continuous braking in the recovery efficiency improvement correction is increased by a preset step of 2 sampling points, or 100 microseconds. This extends the minimum duration of each braking segment by 100 microseconds to improve the integrity of the accumulator's filling. Multiplying the shaping trigger threshold by a preset downward adjustment coefficient of 0.95 means that the trigger threshold is relaxed by 5%, allowing the system to enter the timing shaping mode and expand the coverage of energy recovery even under less fragmented operating conditions. The step of 2 sampling points is based on the fact that the ratio of this increment to the accumulator filling time constant (1.5 to 3 milliseconds) is in the range of 3% to 7%. The single correction amplitude is small enough to avoid abrupt increases in braking segment duration between adjacent wheels, which would cause fluctuations in position accuracy.

[0046] The symmetrical setting of the upward adjustment coefficient of 1.05 and the downward adjustment coefficient of 0.95 ensures that the trigger threshold adjusts towards the opposite direction at the same rate when the position accuracy exceeds the limit or the recovery efficiency is insufficient. After several rounds of correction, it tends to a dynamic equilibrium value. The corrected rearranged parameters are written to the controller register at the end of the current statistical window to overwrite the original value. The write operation occurs during the interval before the start of the next statistical window, i.e., within the register refresh time slot between two adjacent statistical windows. The width of this time slot is 50 microseconds, which is sufficient to complete the register writing of the two parameters without causing read / write competition with the currently executed control operation. As a safety fallback measure under extreme conditions, when the maximum position error exceeds the position accuracy specification value for three consecutive statistical windows, the controller disables the timing shaping function as a whole and restores the original output of the current inner loop until the maximum position error falls back to 80% of the accuracy specification value, i.e., below 0.04 mm, before reactivating the timing shaping function.

[0047] Figure 3 This is a schematic diagram illustrating the adaptive convergence of the shaping trigger threshold and the minimum number of continuous braking sampling points with the number of statistical window rounds in an embodiment of this application. Figure 3As can be seen, the reshaping trigger threshold, starting from an initial value of 0.60, gradually decreased within the first 25 rounds of the statistical window because the actual recovery efficiency was consistently lower than the preset recovery efficiency target value of 0.85 and the maximum position error never exceeded the position accuracy specification value of 0.05 mm. It was multiplied by a preset downgrade coefficient of 0.95 each round, converging to around 0.19 and stabilizing near the 25th round. The minimum number of sampling points for continuous braking, starting from an initial value of 40, gradually increased at a preset step size of 2 points per round, converging to approximately 86 points and stabilizing near the 25th round. Both curves simultaneously entered a stable region after the 25th round, indicating that the actual recovery efficiency and the maximum position error simultaneously reached their respective target constraints near this round. The adaptive correction process completed parameter optimization and entered a dynamic equilibrium state, and the corrected rearranged parameters no longer changed significantly.

[0048] The above describes the servo valve multi-loop feedback and energy recovery method in the embodiments of this application. The following describes the servo valve multi-loop feedback and energy recovery system in the embodiments of this application. One embodiment of the servo valve multi-loop feedback and energy recovery system in the embodiments of this application includes: The adjustment module is used to sequentially adjust the multi-source feedback signals collected by the displacement sensor, speed sensor and current sensor through the outer position loop, the middle speed loop and the inner current loop to obtain the drive command for the inner current loop. The calculation module is used to perform differential statistics on the polarity sequence of the inner current loop drive command to obtain the quantized value of the braking drive switching frequency; to perform a ratio calculation between the quantized value of the braking drive switching frequency and the accumulator charging time constant to obtain the recovery window quality coefficient; when the recovery window quality coefficient is lower than the shaping trigger threshold, to perform equivalent timing rearrangement of the polarity sequence based on the position outer loop error margin to obtain the rearranged drive sequence and the set of braking segment start and end times; The control module is used to output the rearranged drive sequence to the inner current loop, and under the deceleration or reverse motion conditions corresponding to the start and end times of the braking segment, to control the mechanical energy to electrical energy recovery of the back electromotive force through the PWM inverter algorithm to obtain the electrical energy recovery command. The correction module is used to adaptively correct the shaping trigger threshold and the minimum number of sampling points for continuous braking based on the actual recovered electricity and the maximum position error after the execution of the energy recovery command, so as to obtain the corrected rearrangement parameters, and use the corrected rearrangement parameters for the next round of equivalent timing rearrangement.

[0049] This invention also provides a servo valve multi-loop feedback and energy recovery device, which can be a server. The servo valve multi-loop feedback and energy recovery device includes a processor, memory, display screen, input device, network interface, and database connected via a system bus. The processor, designed as a computer, provides computing and control capabilities. The memory of the servo valve multi-loop feedback and energy recovery device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the servo valve multi-loop feedback and energy recovery device stores the data corresponding to this embodiment. The network interface of the servo valve multi-loop feedback and energy recovery device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements the above-described method.

[0050] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the servo valve multi-loop feedback and energy recovery method.

[0051] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0052] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a servo valve multi-loop feedback and energy recovery device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0053] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for multi-loop feedback and energy recovery in a servo valve, characterized in that, The method includes: Step S1: The multi-source feedback signals collected by the displacement sensor, velocity sensor and current sensor are sequentially passed through the position outer loop, velocity middle loop and current inner loop for cascade deviation adjustment to obtain the current inner loop drive command; Step S2: Perform differential statistics on the polarity sequence of the inner current loop drive command to obtain the quantized value of the braking drive switching frequency. Specifically, the driving voltage direction is extracted point-by-point from the inner current loop drive command according to the control cycle of the inner current loop, with positive drive marked as positive polarity and reverse braking marked as negative polarity, forming the polarity sequence. A pairwise differential operation is performed on adjacent sampling points in the polarity sequence. When the polarities of adjacent points are different, a switching event is recorded; when the polarities of adjacent points are the same, no event is recorded, resulting in a switching event sequence. The switching event sequence is accumulated and summed within a preset statistical window to obtain the quantized value of the braking drive switching frequency. The quantized value of the braking drive switching frequency is compared with the accumulator charging time constant to obtain the recovery window quality coefficient. When the recovery window quality coefficient is lower than the shaping trigger threshold, the polarity sequence is rearranged based on the position outer loop error margin to obtain the rearranged drive sequence and the set of braking segment start and end times. Step S3: Output the rearranged drive sequence to the inner current loop, and under the deceleration or reverse motion conditions corresponding to the start and end times of the braking segment, use the PWM inverter algorithm to control the mechanical energy to electrical energy recovery of the back electromotive force, and obtain the electrical energy recovery command. Step S4: Based on the actual recovered electricity and the maximum position error after the execution of the energy recovery command, adaptively correct the shaping trigger threshold and the minimum number of sampling points for continuous braking to obtain the corrected rearrangement parameters, and use the corrected rearrangement parameters for the next round of equivalent timing rearrangement.

2. The servo valve multi-loop feedback and energy recovery method according to claim 1, characterized in that, Step S1 includes: The actual position signal, actual speed signal and actual current signal of the valve core are collected by displacement sensor, speed sensor and current sensor respectively to form the multi-source feedback signal; The position deviation value is obtained by subtracting the position command from the actual position signal. After proportional-integral-derivative adjustment and amplitude limiting by the position outer loop controller, the speed command is output. Based on the deviation between the speed command and the actual speed signal, a speed feedforward component obtained by differentiating the position command is superimposed, and after adaptive proportional-integral adjustment by the speed loop controller, a current command is output. The deviation between the current command and the actual current signal is determined by the hysteresis width. When the absolute value of the deviation exceeds the preset hysteresis width, the polarity of the drive voltage is flipped; otherwise, the current output is maintained, thus obtaining the inner loop drive command for the current.

3. The servo valve multi-loop feedback and energy recovery method according to claim 2, characterized in that, The step of calculating the ratio of the quantized value of the braking drive switching frequency to the accumulator filling time constant to obtain the recovery window quality coefficient includes: The total number of negative polarity sampling points is counted from the polarity sequence, and the average duration of a single continuous braking is calculated by combining the quantized value of the braking drive switching frequency with the current inner loop control cycle. By referring to the pre-calibrated pressure-time constant lookup table based on the current pressure value of the accumulator, the accumulator filling time constant under the current operating conditions can be obtained; The recovery window quality coefficient is obtained by dividing the average duration of the single continuous braking by the accumulator filling time constant.

4. The servo valve multi-loop feedback and energy recovery method according to claim 3, characterized in that, When the quality coefficient of the recovery window is lower than the shaping trigger threshold, the polarity sequence is rearranged in an equivalent time sequence based on the position outer loop error margin to obtain a rearranged driving sequence and a set of braking segment start and end times, including: Subtract the absolute value of the current position outer loop error from the position accuracy specification value specified by the process to obtain the position outer loop error margin; The minimum number of sampling points for continuous braking is obtained by rounding up the quotient of the accumulator filling time constant and the current inner loop control cycle, and the number of braking segments is obtained by dividing the total number of negative polarity sampling points by the minimum number of sampling points for continuous braking and rounding down. According to the number of braking segments, the scattered negative polarity sampling points in the polarity sequence are grouped into continuous braking segments of equal length, and the positive polarity sampling points are grouped into continuous driving segments. They are arranged in an alternating order of braking segments first and driving segments last to form the rearranged driving sequence. The position deviation increment caused by the rearranged drive sequence is checked based on the valve core mass and hydraulic coefficient. When the position deviation increment exceeds the position outer loop error margin, the number of braking segments is reduced and re-arranged until the position deviation increment falls within the position outer loop error margin range. The set of braking segment start and end times, consisting of the rearranged drive sequence and the start and end times of each continuous braking segment, is then output.

5. The servo valve multi-loop feedback and energy recovery method according to claim 1, characterized in that, Step S3 includes: The rearranged drive sequence is output to the hysteresis controller point by point according to the current inner loop control cycle. The PWM duty cycle direction is determined according to the polarity value of each sampling point and the current tracking is maintained within the preset hysteresis width to obtain the rearranged valve core drive current. Before the start time of each braking segment in the set of braking segment start and end times arrives, a preset response compensation time is sent to the inverter bridge arm to send an inverter enable signal to switch the drive circuit of the current inner loop from electric mode to power generation mode. During the effective period of the inverter enable signal, the back electromotive force generated by the valve core drive coil under deceleration or reverse motion conditions is rectified and inverted by the PWM inverter algorithm to obtain the recovery voltage and recovery current. The product of the recovered voltage and the recovered current is integrated over the duration of each braking segment to obtain the single-segment recovered energy. The single-segment recovered energy of all braking segments is summed to form the energy recovery command.

6. The servo valve multi-loop feedback and energy recovery method according to claim 1, characterized in that, Step S4 includes: The actual recovered power is obtained by summing up the recovered power of all braking segments in the energy recovery command, and then dividing the actual recovered power by the theoretical maximum recoverable power to obtain the actual recovery efficiency. Extract the extreme values ​​of the outer loop position error during the execution of the energy recovery command to obtain the maximum position error; When the actual recovery efficiency is lower than the preset recovery efficiency target value and the maximum position error does not exceed the position accuracy specification value specified by the process, the minimum number of continuous braking sampling points is increased by a preset step and the shaping trigger threshold is multiplied by a preset downward adjustment coefficient to obtain the corrected rearrangement parameters; when the maximum position error exceeds the position accuracy specification value, the shaping trigger threshold is multiplied by a preset upward adjustment coefficient and the number of braking segments is reduced to obtain the corrected rearrangement parameters; the priority relationship between the two correction rules is that position accuracy protection takes precedence over recovery efficiency improvement. At the end of each statistical window, it is first determined whether the maximum position error exceeds the position accuracy specification value. If it does, accuracy protection correction is performed regardless of the actual recovery efficiency level. Only when the position accuracy meets the standard is it further determined whether the recovery efficiency needs to be improved. The corrected rearrangement parameters replace the currently stored integer trigger threshold and the minimum number of continuous braking sampling points for the next round of equivalent timing rearrangement.

7. A servo valve multi-loop feedback and energy recovery system, characterized in that, For implementing the servo valve multi-loop feedback and energy recovery method as described in any one of claims 1-6, the servo valve multi-loop feedback and energy recovery system comprises: The adjustment module is used to sequentially adjust the multi-source feedback signals collected by the displacement sensor, speed sensor and current sensor through the outer position loop, the middle speed loop and the inner current loop to obtain the drive command for the inner current loop. The calculation module is used to perform differential statistics on the polarity sequence of the current inner loop drive command to obtain the quantized value of the braking drive switching frequency. Specifically, the driving voltage direction of the current inner loop drive command is extracted point-by-point according to the control cycle of the current inner loop, with positive drive marked as positive polarity and reverse braking marked as negative polarity, forming the polarity sequence. Pairwise differential operations are performed on adjacent sampling points in the polarity sequence. When the polarities of adjacent points are different, a switching event is recorded; when the polarities of adjacent points are the same, no event is recorded, resulting in a switching event sequence. The switching event sequence is accumulated and summed within a preset statistical window to obtain the quantized value of the braking drive switching frequency. The quantized value of the braking drive switching frequency is compared with the accumulator charging time constant to obtain the recovery window quality coefficient. When the recovery window quality coefficient is lower than the shaping trigger threshold, the polarity sequence is rearranged based on the position outer loop error margin to obtain the rearranged drive sequence and the set of braking segment start and end times. The control module is used to output the rearranged drive sequence to the inner current loop, and under the deceleration or reverse motion conditions corresponding to the start and end times of the braking segment, to control the mechanical energy to electrical energy recovery of the back electromotive force through the PWM inverter algorithm to obtain the electrical energy recovery command. The correction module is used to adaptively correct the shaping trigger threshold and the minimum number of sampling points for continuous braking based on the actual recovered electricity and the maximum position error after the execution of the energy recovery command, so as to obtain the corrected rearrangement parameters, and use the corrected rearrangement parameters for the next round of equivalent timing rearrangement.

8. A servo valve multi-loop feedback and energy recovery device, characterized in that, The system includes a memory and a processor, the memory storing a computer program that can run on the processor, and the processor executing the computer program to implement the servo valve multi-loop feedback and energy recovery method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run by the processor, it causes the processor to execute the servo valve multi-loop feedback and energy recovery method as described in any one of claims 1 to 6.

Images