An online optimal control system for realizing optimal efficiency of a motor in a wide speed range

Abstract

The application discloses an online optimization control system for realizing optimal efficiency of a motor in a wide rotating speed range, and relates to the technical field of motor control. The application discloses an online optimization control system for realizing optimal efficiency of a motor in a wide rotating speed range, and relates to the technical field of motor control. The application discloses an online optimization control system for realizing optimal efficiency of a motor in a wide rotating speed range, and relates to the technical field of motor control. The application discloses an online optimization control system for realizing optimal efficiency of a motor in a wide rotating speed range, and relates to the technical field of motor control. The application discloses an online optimization control system for realizing optimal efficiency of a motor in a wide rotating speed range, and relates to the technical field of motor control. The application discloses an online optimization control system for realizing optimal efficiency of a motor in a wide rotating speed range, and relates to the technical field of motor control. The application discloses an online optimization control system for realizing optimal efficiency of a motor in a wide rotating speed range, and relates to the technical field of motor control. The application discloses an online optimization control system for realizing optimal efficiency of a motor in a wide rotating speed range, and relates to the technical field of motor control. The application discloses an online optimization control system for realizing optimal efficiency of a motor in a wide rotating speed range, and relates to the technical field of motor control. The application discloses an online optimization control system for realizing optimal efficiency of a motor in a wide rotating speed range, and relates to the technical field of motor control. The application discloses an online optimization control system for realizing optimal efficiency of a motor in a wide rotating speed range, and relates to the technical field of motor control. The application discloses an online optimization control system for realizing optimal efficiency of a motor in a wide rotating speed range, and relates to the technical field of motor control. The application discloses an online optimization control system for realizing optimal efficiency of a motor in a

Description

Technical Field

[0001] This invention relates to the field of motor control technology, and in particular to an online optimization control system for achieving optimal efficiency of a motor over a wide speed range. Background Technology

[0002] In automotive electric power steering systems or steer-by-wire road feel simulation systems, the steering wheel control motor is typically connected to the steering shaft or steering wheel via a reduction gear. To achieve precise torque sensing and transmission, a torsion bar, acting as an elastic element, is commonly found in the transmission link. When the vehicle is parking at low speeds, making fine-tuning adjustments at high speeds, or driving on rough surfaces, there is frequent torque interaction between the steering wheel and the wheels. This causes the torsion bar to undergo varying degrees of elastic torsional deformation during operation. This repeated torsion and return-to-center process is accompanied by the storage and release of elastic potential energy, resulting in significant dynamic elastic characteristics in the motor shaft end load. Consequently, the DC input power of the motor contains a reversible power fluctuation component closely related to the mechanical elastic deformation.

[0003] Existing online efficiency optimization control technologies for motors typically employ perturbation observation or gradient-based search strategies. These methods determine the direction of efficiency optimization by applying small trial signals to the motor control current and directly monitoring changes in DC bus input power or current. However, this conventional approach has significant limitations when applied to systems with elastic torsion bars. Because current technologies cannot effectively identify and eliminate reversible power fluctuations caused by the release of elastic potential energy from the torsion bar, the control algorithm is prone to misinterpreting the temporary decrease in input power due to the release of energy from the elastic element as an increase in motor efficiency, thus calculating incorrect optimization gradients. This misinterpretation leads to repeated jumps in the optimization direction, convergence point drift, and even deviations from the optimal efficiency operating point in dynamic conditions across a wide speed range, making it difficult to achieve stable and accurate energy efficiency optimization control. Summary of the Invention

[0004] The purpose of this invention is to solve the problem in the prior art where the power evaluation quantity is easily affected by the round-trip power fluctuation caused by the elastic energy storage and release of the torsion bar during the online optimization process, resulting in unstable optimization direction and drift of the optimal point under wide speed dynamic conditions. Therefore, this invention proposes an online optimization control system that realizes the optimal efficiency of the motor over a wide speed range.

[0005] To address the problems existing in the prior art, the present invention adopts the following technical solution:

[0006] An online optimization control system for achieving optimal efficiency of a motor over a wide speed range includes:

[0007] The data acquisition module is used to collect data on DC bus voltage, DC bus current, torsion bar input side angle, and torsion bar output side angle, and to calculate DC side input power and torsion bar torsion based on the collected data;

[0008] The energy storage index module is used to determine the torsion bar's round-trip energy storage power based on the torsion bar's torsion amount and its rate of change over time, and to calculate the percentage index of the torsion bar's round-trip energy storage power within a preset time window.

[0009] The residual evaluation module is used to calculate the projection correlation coefficient of the DC-side input power on the torsion bar reciprocating energy storage power, and to correct the DC-side input power based on the projection correlation coefficient to obtain the residual power evaluation quantity.

[0010] The optimization direction module is used to obtain the direct-axis current reference value, apply alternating perturbations to the direct-axis current reference value, and determine the optimization update direction based on the difference in the residual power evaluation quantity during the perturbation period.

[0011] The adaptive update module is used to generate adaptive weights based on the proportion index of the torsion bar reciprocating energy storage power, and to suppress the update amplitude of the optimization update direction according to the adaptive weights, so as to obtain the updated direct-axis current reference value.

[0012] The synthesis and control module is used to obtain the quadrature axis current reference value, determine the final current reference value based on the quadrature axis current reference value and the updated direct axis current reference value, and control the motor operation.

[0013] Preferably, the calculation of the DC-side input power and torsion bar torsion based on the collected data includes:

[0014] The DC bus voltage and DC bus current are multiplied to obtain the DC side input power.

[0015] The difference between the input side angle and the output side angle of the torsion bar is calculated, and the difference is periodically converted to obtain the continuous torsion bar torsion amount.

[0016] Preferably, the torsion bar reciprocating energy storage power is determined based on the torsion amount of the torsion bar and its rate of change over time, including:

[0017] The torsion bar torsion amount is differentially calculated to obtain the torsional change rate;

[0018] Obtain the equivalent torsional stiffness of the torsion bar, and multiply the equivalent torsional stiffness of the torsion bar, the current torsion amount of the torsion bar, and the rate of change of torsion to obtain the torsion bar's round-trip energy storage power.

[0019] Preferably, the proportion index of the torsion bar's round-trip energy storage power is calculated within a preset time window, including:

[0020] Within a preset time window, the root mean square value of the energy storage power of the torsion bar during round trip is calculated, and the output side angle of the torsion bar is differentially calculated to obtain the output side angular velocity.

[0021] Within a preset time window, the torsion bar torque is calculated based on the torsion bar equivalent torsional stiffness and torsion bar torsion amount. The torsion bar torque is then multiplied by the output side angular velocity to obtain the net power assist mechanical power.

[0022] Calculate the mean value of the net assist mechanical power and take its absolute value. Based on the ratio of the root mean square value of the torsion bar reciprocating energy storage power to the absolute value, obtain the proportion index of the torsion bar reciprocating energy storage power.

[0023] Preferably, the projection correlation coefficient of the DC-side input power on the torsion bar reciprocating energy storage power is calculated, and the DC-side input power is corrected based on the projection correlation coefficient, including:

[0024] Within a preset time window, the sum of the products of the DC input power and the torsion bar reciprocating energy storage power is taken as the first value; the sum of the squares of the torsion bar reciprocating energy storage power is taken as the second value.

[0025] The projection correlation coefficient is obtained based on the ratio of the first value to the second value.

[0026] The residual power is obtained by subtracting the product of the projection correlation coefficient and the torsion bar reciprocating energy storage power from the DC side input power.

[0027] The average value of the residual power within the preset time window is used as the residual power evaluation metric.

[0028] Preferably, alternating perturbations are applied to the direct-axis current reference value, and the optimization update direction is determined based on the difference in residual power evaluation during the perturbation period, including:

[0029] Positive and negative perturbations are alternately applied to the direct-axis current reference bias, and the residual power evaluation quantities during the application of positive perturbation and negative perturbation are calculated respectively.

[0030] The efficiency gradient is obtained by calculating the difference between the residual power evaluation during a positive disturbance and the residual power evaluation during a negative disturbance, and dividing the difference by twice the disturbance amplitude.

[0031] Preferably, adaptive weights are generated based on the proportion index of the torsion bar reciprocating energy storage power, and the update magnitude of the optimization update direction is suppressed according to the adaptive weights, including:

[0032] A function inversely proportional to the proportion of torsion bar reciprocating energy storage power is constructed to calculate adaptive weights;

[0033] The update magnitude is obtained by multiplying the efficiency gradient by the preset update coefficient.

[0034] The update amplitude is attenuated based on adaptive weights. The current direct-axis current reference value is superimposed with the attenuated update amplitude and combined with the current safety range limit to obtain the updated direct-axis current reference value.

[0035] Preferably, determining the final current reference value based on the quadrature-axis current reference value and the updated direct-axis current reference value includes:

[0036] The current vector magnitude is synthesized based on the updated direct-axis current reference value and quadrature-axis current reference value;

[0037] If the current vector amplitude exceeds the maximum allowable current limit of the motor, then calculate the ratio coefficient between the maximum current limit and the current vector amplitude;

[0038] The direct-axis current reference value and the quadrature-axis current reference value are simultaneously and proportionally reduced based on the scaling factor to obtain the final current reference value.

[0039] Compared with the prior art, the beneficial effects of the present invention are:

[0040] 1. This invention calculates the torsion bar's reciprocating energy storage power using the input and output angle data of the torsion bar, and further calculates the projection correlation coefficient of the DC-side input power onto this reciprocating energy storage power. This quantifies and deducts the reversible fluctuation component in the input power that is synchronized with the elastic deformation of the torsion bar, thus obtaining the residual power evaluation quantity. It separates the mechanical elastic energy storage release power, which is unrelated to efficiency, from the total input power, so that the evaluation index on which the optimization algorithm is based can truly reflect the loss changes and effective load power changes of the motor drive system. This effectively solves the problem of misjudging elastic release energy as efficiency improvement, which leads to errors in the optimization gradient calculation.

[0041] 2. This invention constructs a ratio index of torsion bar reciprocating energy storage power to quantify the strength of elastic energy storage interference relative to net power assist machinery, and generates an adaptive weight with an inverse relationship based on this. When the vehicle is driving on rough roads or performing frequent steering operations that cause severe torsion bar reciprocating energy storage interference, the system can automatically suppress the update amplitude of the direct axis current reference value through this weight, preventing repeated jumps in the optimization direction or convergence point drift caused by instantaneous high fluctuations. This enables the system to maintain the ability to track the optimal efficiency point under dynamic operating conditions in a wide speed range, while avoiding control oscillations caused by misjudgment, thus achieving stable and accurate energy efficiency optimization control. Attached Figure Description

[0042] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0043] Figure 1 This is a functional block diagram of an online optimization control system for achieving optimal efficiency of a motor over a wide speed range, according to the present invention.

[0044] Figure 2 This is a schematic diagram of the physical model and power composition of the torsion bar of the present invention. Detailed Implementation

[0045] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0046] Example: This example provides an online optimization control system for achieving optimal efficiency of a motor over a wide speed range. See [link to example]. Figure 1 Specifically, including:

[0047] The data acquisition module is used to collect data on DC bus voltage, DC bus current, torsion bar input side angle, and torsion bar output side angle, and to calculate DC side input power and torsion bar torsion based on the collected data;

[0048] The energy storage index module is used to determine the torsion bar's round-trip energy storage power based on the torsion bar's torsion amount and its rate of change over time, and to calculate the percentage index of the torsion bar's round-trip energy storage power within a preset time window.

[0049] The residual evaluation module is used to calculate the projection correlation coefficient of the DC-side input power on the torsion bar reciprocating energy storage power, and to correct the DC-side input power based on the projection correlation coefficient to obtain the residual power evaluation quantity.

[0050] The optimization direction module is used to obtain the direct-axis current reference value, apply alternating perturbations to the direct-axis current reference value, and determine the optimization update direction based on the difference in the residual power evaluation quantity during the perturbation period.

[0051] The adaptive update module is used to generate adaptive weights based on the proportion index of the torsion bar reciprocating energy storage power, and to suppress the update amplitude of the optimization update direction according to the adaptive weights, so as to obtain the updated direct-axis current reference value.

[0052] The synthesis and control module is used to obtain the quadrature axis current reference value, determine the final current reference value based on the quadrature axis current reference value and the updated direct axis current reference value, and control the motor operation.

[0053] In embodiments of the present invention, data on DC bus voltage, DC bus current, torsion bar input side angle, and torsion bar output side angle are collected. Based on the collected data, the DC-side input power and torsion bar torsion are calculated, including:

[0054] Collect data on DC bus voltage, DC bus current, torsion bar input side angle, and torsion bar output side angle;

[0055] The DC-side input power is calculated based on the collected DC bus voltage and DC bus current.

[0056] The torsion amount of the torsion bar is determined based on the difference between the input side angle and the output side angle of the torsion bar.

[0057] Specifically, DC bus voltage refers to the potential difference between the two ends of the DC bus of the motor driver, used to characterize the voltage supply level of the DC bus to provide electrical energy to the inverter power stage; DC bus current refers to the current flowing through the DC bus circuit, used to characterize the current flux of the DC bus to the motor drive circuit; torsion bar input side angle refers to the angular position of the torsion bar at the input end relative to the reference direction, used to characterize the angular displacement input from the steering wheel side or upstream transmission side to the end of the torsion bar; torsion bar output side angle refers to the angular position of the torsion bar at the output end relative to the reference direction, used to characterize the angular displacement response of the torsion bar to the downstream mechanism or motor output side after elastic transmission; and torsion bar torsion is the degree of torsional deformation determined by the angular difference between the torsion bar input side angle and the torsion bar output side angle, used to describe the amount of elastic torsional deformation of the torsion bar due to force and as the basis for subsequent energy storage and release calculations.

[0058] Specifically, a voltage acquisition channel electrically connected to the positive and negative terminals of the DC bus and a current acquisition channel connected in series with the DC bus are set up in the motor controller. The voltage acquisition channel converts the DC bus voltage into a voltage signal that the controller can sample through resistor voltage division and isolation amplification. The current acquisition channel converts the DC bus current into a voltage signal proportional to the current through a shunt resistor or a Hall current sensor. The analog signals are synchronously sampled according to a preset sampling period, and the analog-to-digital conversion unit outputs digital DC bus voltage and DC bus current data. The acquisition of the torsion bar input and output angles is accomplished by angle sensors installed at the torsion bar input and output ends, respectively. The angle sensors can be magnetic encoders. An angle sensor, either a rotary transformer or a potentiometer, is fixed to the housing and coaxially connected to the end of the torsion bar to ensure that the angle changes synchronously with the end of the torsion bar. The controller obtains the original angle reading by pulse width measurement, serial bus reading, or by solving the excitation and feedback signals of the rotary transformer. In each sampling cycle, the controller performs zero-point alignment and direction consistency processing on the original angle readings on the input and output sides of the torsion bar. The sampling period of the acquisition process is preferably one to two milliseconds. This sampling period range can balance the real-time requirements of motor control and the performance capacity of the hardware acquisition module, avoiding the impact of data acquisition lag on control accuracy under dynamic conditions, and avoiding the increase of hardware computing load and energy consumption due to excessive sampling.

[0059] Specifically, the DC bus voltage and current data at the same sampling moment are used as the calculation input for the power input rate. Considering that the voltage and current together determine the instantaneous power when the DC bus supplies energy to the inverter power stage, the DC side input power is calculated according to the product relationship. The formula for calculating the DC side input power is as follows:

[0060]

[0061] in, Let be the DC-side input power at the k-th sampling time. Let be the DC bus voltage at the k-th sampling time. Let be the DC bus current at the k-th sampling time. The above formula uses the basic definition of circuit power to characterize the amount of electrical energy entering the drive system from the DC bus per unit time. To reduce the amplification effect of switching noise and sampling quantization error on power calculation, it is preferable to... and After performing first-order low-pass filtering, the results are substituted into the above formula. The filtering time constant is preferably one-tenth to one-fifth of the motor's electrical time constant to suppress carrier ripple without significantly introducing dynamic lag. In the torsion bar torsion calculation, the input and output angle data of the torsion bar at the same sampling time are used as the basis for measuring torsional deformation. The torsion bar torsion is determined based on the structural relationship reflecting the elastic torsional deformation of the torsion bar, using the difference in angular displacement between the two ends of the torsion bar. Considering the periodicity of the angle sensor output, it is preferable to perform minimum equivalent period conversion after the difference calculation to eliminate numerical jumps caused by zero-crossing. The torsion bar torsion is determined according to the following formula:

[0062]

[0063] in, Let be the torsion of the torsion bar at the k-th sampling time. Let the angle of the torsion bar input side be the angle at the k-th sampling time. Let be the torsion bar output angle at the k-th sampling time. This is a periodic conversion operator used to limit the angle difference within a preset minimum equivalent period range to maintain the continuity of the torsion bar torsion sequence. The minimum equivalent period range is preferably... or Based on the common output period and difference continuity requirements of angle sensors, the torsion bar torsion amount obtained by the above conversion can more realistically reflect the degree of elastic torsional deformation of the torsion bar during continuous steering, thus providing a stable input for subsequent calculation of the torsion bar round-trip energy storage power based on the torsion bar torsion amount and its rate of change.

[0064] In embodiments of the present invention, the torsion bar round-trip energy storage power is determined based on the torsion amount of the torsion bar and its rate of change over time, and the proportion index of the torsion bar round-trip energy storage power is calculated within a preset time window, including:

[0065] The torsion bar torsion amount is differentially calculated to obtain the torsional change rate;

[0066] Obtain the equivalent torsional stiffness of the torsion bar, and multiply the equivalent torsional stiffness of the torsion bar, the current torsion amount of the torsion bar, and the rate of change of torsion to obtain the torsion bar round-trip energy storage power.

[0067] Specifically, the torsional change rate is the rate at which the torsion of a torsion bar changes over time. It reflects whether the torsional deformation of the torsion bar is continuously increasing or decreasing, and the degree of drastic change. The equivalent torsional stiffness of the torsion bar is a parameter used to characterize the restoring torque capability corresponding to a unit torsional deformation within the elastic working range of the torsion bar. Its magnitude is determined by the torsion bar material, geometry, and equivalent installation conditions. The torsion bar reciprocating energy storage power is a characterization of the rate of change of elastic energy storage and release of the torsion bar per unit time. When the torsion of the torsion bar increases and the torsional change rate is in the same direction, it indicates that energy is input and stored in the elastic deformation of the torsion bar. When the torsion of the torsion bar decreases or the torsional change rate is in the opposite direction, it indicates that the stored energy is released and fed back to the transmission link or motor side. Thus, this power exhibits a reciprocating characteristic with alternating positive and negative values ​​depending on the operating conditions. It is used to characterize the impact of the torsion bar's elastic energy storage and release process on system power fluctuations and optimization evaluation quantities.

[0068] Specifically, after obtaining a continuous sequence of torsion bar torsional values, in order to characterize the rate of change of torsion bar torsional values ​​with time, the torsion bar torsional values ​​at adjacent sampling times are differentially calculated. The difference between the torsion bar torsional values ​​at adjacent sampling times is divided by the sampling period to convert it into the rate of change per unit time. Dividing the difference by the sampling period can obtain a rate of change consistent with the continuous time derivative under real-time calculation conditions.

[0069] Specifically, the equivalent torsional stiffness of the torsion bar is obtained and used to characterize the restoring torque capability corresponding to unit torsional deformation within the elastic working range. The equivalent torsional stiffness of the torsion bar is preferably obtained through calibration after assembly. During the calibration process, the steady-state torsion of the torsion bar is recorded under known loading torque conditions, and the ratio of the loading torque to the torsion bar torsion is taken as the equivalent torsional stiffness of the torsion bar. After obtaining the equivalent torsional stiffness of the torsion bar, the equivalent torsional stiffness of the torsion bar, the current torsion of the torsion bar, and the rate of change of torsion are multiplied to obtain the round-trip energy storage power of the torsion bar. Based on the rate of change of the elastic energy storage of the torsion bar with time, the restoring torque of the torsion bar within the elastic linear range is equal to the product of the equivalent torsional stiffness of the torsion bar and the torsion of the torsion bar. The instantaneous power of the torsion bar doing work is equal to the product of the restoring torque and the torsional angular velocity. The rate of change of torsion is a discrete approximation of the torsional angular velocity. Therefore, the value obtained by multiplying the three can characterize the intensity of the change in elastic energy storage and release of the torsion bar per unit time, and its sign changes with the direction of torsional change to reflect the round-trip characteristics of energy storage and release.

[0070] Within a preset time window, the root mean square value of the energy storage power of the torsion bar is calculated, and the output side angle of the torsion bar is differentially calculated to obtain the output side angular velocity. The torsion bar torque is calculated based on the equivalent torsional stiffness and torsion amount of the torsion bar, and the torsion bar torque is multiplied by the output side angular velocity to obtain the net power assist mechanical power.

[0071] Calculate the mean value of the net assist mechanical power and take the absolute value. Calculate the percentage index of the torsion bar reciprocating energy storage power based on the root mean square value of the torsion bar reciprocating energy storage power and the absolute value.

[0072] Specifically, net power assist refers to the instantaneous mechanical power measure of the torsion bar on the output side to the downstream steering mechanism or motor output side. It is the combined effect of the torsion bar's elastic recovery torque and the angular velocity on the output side, used to characterize the rate at which the torsion bar outputs mechanical energy to the downstream transmission link or absorbs mechanical energy from the downstream per unit time. When the steering assist direction is consistent, this power reflects the effective transmission level of assist energy. Under reverse steering or anti-torsion conditions, this power may change direction but still represents the mechanical power measure for that period. The torsion bar round-trip energy storage power ratio index is an indicator used to quantify the relative proportion of the round-trip energy storage power intensity caused by the torsion bar's elastic energy storage and release relative to the net power assist mechanical power measure. It is obtained by statistically analyzing the effective intensity of the torsion bar's round-trip energy storage power within a preset time window and comparing it with the statistical scale of the net power assist mechanical power. It is used to characterize the contribution of the torsion bar's elastic round-trip process to power fluctuations. When this ratio index is large, it indicates that the round-trip energy storage power fluctuation accounts for a higher proportion in the power composition and is more likely to interfere with power-based optimization evaluation. When this ratio index is small, it indicates that the net power assist mechanical power is dominant and the optimization evaluation can better reflect the actual loss changes.

[0073] Specifically, after continuously obtaining the torsion bar round-trip energy storage power sequence, torsion bar torsion quantity sequence, and torsion bar output side angle sequence, to ensure that the proportion index can reflect the stable statistical characteristics of the round-trip energy storage power intensity relative to the net power assist mechanical scale over a period of time, a preset time window is set for window statistics of the above sequences. The time window is preferably implemented with a fixed number of sampling points and consistent with the optimization perturbation rhythm; more preferably, fifty to two hundred sampling points are selected. The steering wheel control motor, under a millisecond-level control cycle, can cover more than one torsion bar round-trip energy exchange cycle within this window length without introducing excessive statistical lag. Within the time window, the torsion bar round-trip energy storage power at each sampling time within the window is squared, averaged, and the square root is taken to obtain the root mean square value, which is used to characterize the effective intensity of the round-trip energy storage power. The root mean square value is calculated according to the following formula:

[0074]

[0075] in, This represents the root mean square value of the torsion bar's round-trip energy storage power within the time window. The number of sampling points within the time window. For the first time window The reason for using the root mean square (RMS) of the torsion bar round-trip energy storage power at each sampling point is that the torsion bar round-trip energy storage power alternates between positive and negative values ​​with energy storage and release. Directly averaging the values ​​would cause them to cancel each other out and fail to reflect the intensity of the fluctuations. The RMS can retain the contribution of the power amplitude while avoiding the cancellation of positive and negative values.

[0076] Specifically, the output-side angular velocity is obtained by differentiating the torsion bar output-side angle and dividing by the sampling period. The output-side angular velocity is determined according to the following formula:

[0077]

[0078] in, For the first The output side angular velocity at each sampling time For the first The torsion bar output side angle at each sampling time For the first The torsion bar output side angle at each sampling time The reason for using a differential approximation for the sampling period is that the output-side angular velocity is the rate of change of the output-side angle with respect to time, and the angle data is obtained through discrete sampling. Within the time window, the torsion bar torque is calculated based on the equivalent torsional stiffness and torsion amount of the torsion bar to characterize the magnitude of the elastic recovery torque of the torsion bar. The torsion bar torque is calculated according to the following formula:

[0079]

[0080] in, For the first Torque at each sampling time, The equivalent torsional stiffness of the torsion bar. For the first The torsion of the torsion bar at each sampling time is approximately proportional to the recovery torque of the torsion bar within the small deformation elastic range. After obtaining the torsion bar torque and the output side angular velocity, the two are multiplied to obtain the net assist mechanical power, which characterizes the mechanical power output by the torsion bar to the downstream transmission link or absorbed from the downstream. The net assist mechanical power is calculated according to the following formula:

[0081]

[0082] in, For the first The net assist mechanical power at each sampling moment is given. The instantaneous mechanical power of the rotating system is equal to the product of torque and angular velocity. Considering that the net assist mechanical power may vary positively or negatively under different operating conditions, the average net assist mechanical power within the time window is calculated and the absolute value is taken to obtain a power metric, which is used for comparison with the round-trip energy storage power intensity. The power metric is calculated according to the following formula:

[0083]

[0084] in, The absolute value of the net assist mechanical power within the time window is used because the percentage index is used to represent the ratio of strength to scale without distinguishing the power direction. Based on the above two types of scale quantities, the percentage index of the torsion bar round-trip energy storage power is obtained by dividing the root mean square value of the torsion bar round-trip energy storage power by the absolute value of the net assist mechanical power. The percentage index is calculated according to the following formula:

[0085]

[0086] in, This represents the percentage of energy stored by the torsion bar during round trip. To prevent extremely small positive numbers from diverging in value due to the denominator being close to zero, The preferred value is one-thousandth to one-ten-thousandth of the average net assist mechanical power scale under typical assist conditions. This avoids numerical instability caused by an excessively small denominator at low speeds or near-zero assist, and does not cause significant bias in the proportion calculation under normal operating conditions. This results in an index that can stably characterize the proportion of torsion bar reciprocating energy storage power intensity relative to the net assist mechanical power scale, which can then be used for subsequent adaptive weight generation.

[0087] In an embodiment of the present invention, the projection correlation coefficient of the DC-side input power on the torsion bar reciprocating energy storage power is calculated, and the DC-side input power is corrected based on the projection correlation coefficient, including:

[0088] Within a preset time window, the projection correlation coefficient is calculated based on the DC side input power and the torsion bar reciprocating energy storage power.

[0089] The residual power is obtained by subtracting the product of the projection correlation coefficient and the torsion bar reciprocating energy storage power from the DC side input power.

[0090] The average value of the residual power within the preset time window is used as the residual power evaluation metric.

[0091] Specifically, the projection correlation coefficient refers to a coefficient used within a preset time window to measure the degree of synchronization between the DC-side input power and the torsion bar reciprocating energy storage power in terms of their changing trends, and to give the proportion of the intensity of the synchronization component. It is obtained by accumulating the point-by-point product of the two within the window and normalizing it with the energy scale of the torsion bar reciprocating energy storage power itself, thereby reflecting what proportion of the DC-side input power can be regarded as a fluctuation component that moves in phase with the torsion bar reciprocating energy storage power. The residual power refers to the remaining power component obtained after subtracting the fluctuation component of the DC-side input power that changes synchronously with the torsion bar reciprocating energy storage power according to the projection correlation coefficient. This remaining power component is more likely to reflect the actual loss of the motor drive system and the change in the effective power of the load, and is less likely to be disturbed by the reversible power fluctuations caused by the torsion bar elastic reciprocating energy storage. The residual power evaluation quantity refers to the evaluation index used for online optimization obtained by statistically summarizing the residual power within a preset time window. It is preferable to use the average value of the residual power within the window as the evaluation quantity to balance instantaneous noise and power pulsation, so that the evaluation quantity can be used as a basis for comparing the efficiency changes before and after the perturbation and to determine the optimization update direction and update magnitude.

[0092] Specifically, after continuously obtaining the DC-side input power sequence and the torsion bar reciprocating energy storage power sequence, in order to remove the fluctuation component synchronized with the torsion bar elastic reciprocating energy storage from the DC-side input power without introducing additional sensors, window statistics are performed on the two power sequences within the aforementioned preset time window. Within this time window, the DC-side input power and the torsion bar reciprocating energy storage power at each sampling point are multiplied point by point and the products are accumulated. The accumulated sum is used as the first value to characterize the synchronous change intensity of the two within the window. At the same time, the torsion bar reciprocating energy storage power at each sampling point is squared and accumulated. The sum of the squares is used as the second value to characterize the energy scale of the torsion bar reciprocating energy storage power within the window. The ratio of the first value to the second value is determined as the projection correlation coefficient. The meaning of this coefficient is the scaling factor corresponding to projecting the DC-side input power along the fluctuation direction of the torsion bar reciprocating energy storage power, thereby estimating the magnitude of the fluctuation component in the DC-side input power that moves in phase with the torsion bar reciprocating energy storage power.

[0093] Specifically, after obtaining the projection correlation coefficient, a desynchronization component operation is performed on the DC-side input power at each sampling point within the time window. The residual power is obtained by subtracting the product of the projection correlation coefficient and the torsion bar round-trip energy storage power from the DC-side input power. The residual power is calculated according to the following formula:

[0094]

[0095] in, Let be the residual power at the k-th sampling time. Let be the DC-side input power at the k-th sampling time. Let be the torsion bar reciprocating energy storage power at the k-th sampling time. The reason for using linear projection subtraction in the above formula for the projection correlation coefficient is that the part of the fluctuation of DC input power that is synchronized with the torsion bar reciprocating energy storage power can be approximated by the proportional scaling of the torsion bar reciprocating energy storage power. The residual power obtained after subtracting this part can more centrally reflect the changes in motor drive loss and effective load power and is not easily disturbed by the reversible reciprocating energy storage fluctuation. In order to make the residual power stable when used for online optimization and reduce the impact of instantaneous noise, the residual power is averaged within the same preset time window and the average value is used as the residual power evaluation quantity. The statistical scope of the residual power evaluation quantity is consistent with the calculation scope of the projection correlation coefficient, so it can be directly used to compare the evaluation quantities before and after the perturbation to determine the optimization update direction and to achieve update amplitude suppression with adaptive weights.

[0096] In an embodiment of the present invention, a direct-axis current reference value is obtained, and alternating perturbations are applied to the direct-axis current reference value. An optimization update direction is determined based on the difference in residual power evaluation values ​​during the perturbation period, including:

[0097] Positive and negative perturbations are applied alternately to the direct-axis current reference value. The residual power evaluation quantity during the period of applying the positive perturbation and the residual power evaluation quantity during the period of applying the negative perturbation are calculated respectively. The difference between the two is calculated and the difference is divided by the perturbation amplitude range to obtain the efficiency gradient.

[0098] Specifically, the efficiency gradient refers to a quantity used to characterize the trend and sensitivity of the residual power evaluation quantity to small changes in the direct-axis current reference value during the application of positive and negative perturbations. It is obtained by comparing the difference between the residual power evaluation quantity during the positive perturbation period and the residual power evaluation quantity during the negative perturbation period, and normalizing it by combining the equivalent change amplitude between the two perturbations. This gives whether increasing or decreasing the direct-axis current reference value near the current operating condition and the current control point will increase or decrease the residual power evaluation quantity, and the strength of the change. The sign of the efficiency gradient value is used to indicate the optimization update direction, and the absolute value of the efficiency gradient value is used to reflect the sensitivity of the evaluation quantity to the direct-axis current reference value and as the basis for adjusting the update step size, so that online optimization can gradually update the direct-axis current reference value in a direction that is conducive to reducing the residual power evaluation quantity without relying on an external efficiency model.

[0099] Specifically, when entering the online optimization cycle, the reference value of the direct-axis current used to control the motor flux linkage component is read from the reference generation unit of the motor current control loop and used as the optimization benchmark input. The direct-axis current reference value is preferably the direct-axis current reference value that has been effective in the previous control cycle or the previous optimization cycle, so as to ensure that the optimization is carried out near the current operating condition and does not introduce abrupt changes. The direct-axis current reference value is used as the optimization center point, and a small disturbance with symmetrical amplitude is constructed near the center point to form a comparable observation of the residual power evaluation quantity. The disturbance amplitude is preset by the controller and remains unchanged in the optimization cycle. The disturbance amplitude is preferably one-thousandth to one-hundredth of the allowable range of the direct-axis current reference value. This amplitude can make the residual power evaluation quantity produce a distinguishable change in the noise background without causing obvious torque fluctuations and perceptible vibrations in the steering wheel control motor.

[0100] Specifically, the direct-axis current reference value and the positive disturbance amplitude are superimposed as the positive disturbance control quantity, and a preset evaluation time window length is maintained to ensure that the statistical caliber of the residual power evaluation quantity is consistent with the aforementioned projection desynchronization component processing. The evaluation time window length is preferably consistent with the time window used for calculating the projection correlation coefficient. After the evaluation window expires, the direct-axis current reference value and the negative disturbance amplitude are superimposed as the negative disturbance control quantity, and the same evaluation time window length is maintained. During the maintenance period, the residual power evaluation quantity during the negative disturbance period is output in the same residual power calculation and window averaging manner. Since the two disturbances only show symmetrical changes in the direct-axis current reference value and the evaluation window length is consistent, the difference between the two evaluation quantities can be attributed to the influence of the change in the direct-axis current reference value on the system loss and effective power.

[0101] Specifically, the difference between the residual power evaluation quantity during the positive disturbance period and the residual power evaluation quantity during the negative disturbance period is calculated, and the difference is divided by twice the disturbance amplitude to obtain the efficiency gradient value. Here, twice the disturbance amplitude corresponds to the equivalent change between the positive disturbance control quantity and the negative disturbance control quantity. Using this normalization, the difference in evaluation quantity can be converted into the change amplitude of evaluation quantity caused by the change of unit direct-axis current reference value. The sign of the efficiency gradient value directly indicates whether the direct-axis current reference value should be increased or decreased to reduce the residual power evaluation quantity. Based on this, the optimization update direction is determined and input is provided for subsequent update amplitude suppression combined with adaptive weights.

[0102] In an embodiment of the present invention, an adaptive weight is generated based on the proportion index of the torsion bar reciprocating energy storage power, and the update magnitude of the optimization update direction is suppressed according to the adaptive weight, including:

[0103] A function inversely proportional to the proportion of torsion bar reciprocating energy storage power is constructed to calculate adaptive weights;

[0104] The update magnitude is calculated based on the efficiency gradient value;

[0105] The update amplitude is attenuated based on adaptive weights. The updated direct-axis current reference value is obtained by superimposing the attenuated update amplitude on the direct-axis current reference value and combining it with the current safety range limit.

[0106] Specifically, to reduce the aggressiveness of optimization updates and avoid excessive oscillation of the direct-axis current reference value under conditions of strong interference from the torsion bar's elastic reciprocating energy storage, a function inversely proportional to the proportion index of the torsion bar's reciprocating energy storage power is constructed to calculate the adaptive weight. The adaptive weight is preferably determined by the following formula:

[0107]

[0108] in, For adaptive weights, The ratio of the torsion bar reciprocating energy storage power is used because when the ratio increases, it indicates that the reciprocating energy storage power fluctuation accounts for a higher proportion in the power composition and the residual power evaluation is more susceptible to disturbances. In this case, by making the weight monotonically decrease as the ratio increases, the update amplitude can be actively suppressed to improve the optimization stability. When the ratio is small, the weight is close to one to maintain the optimization response speed. At the same time, this form has a constant positive denominator and low computational cost, making it suitable for real-time implementation. To avoid excessive update amplitude under extremely low ratios, the lower limit of the adaptive weight is preferably set to 0.1 to 0.3 to ensure that the optimization can still proceed slowly under strong disturbances and to prevent the convergence from stalling due to excessively small weights. The upper limit of the adaptive weight is preferably set to one to maintain no amplification effect.

[0109] Specifically, in the update amplitude calculation section, an undiminished update amplitude is generated based on the efficiency gradient value. The update amplitude is preferably obtained by multiplying a preset update coefficient by the efficiency gradient value. The update coefficient is preferably a percentage of the amplitude corresponding to one-thousandth to one-hundredth of the rated direct-axis current range of the motor. The value is chosen to match the aforementioned disturbance amplitude level, ensuring that the change in the direct-axis current reference value within each optimization cycle does not exceed the range that the current control loop can smoothly track, and avoiding any perceptible impact on the assist feel. After obtaining the undiminished update amplitude, an adaptive weight is used to attenuate the update amplitude to obtain a attenuated update amplitude. The direct-axis current reference value is then superimposed with the attenuated update amplitude to obtain a candidate updated direct-axis current reference value. The superposition direction is determined by the sign of the efficiency gradient, ensuring that the update follows a direction that is conducive to reducing the residual power evaluation quantity. Specifically, this is implemented as follows:

[0110]

[0111] In the formula, This is the updated direct-axis current reference value. This is the current direct-axis current reference value. For adaptive weights, For update range.

[0112] Specifically, to meet the current safety range limit, the upper and lower limits of the direct-axis current reference value are preset and saturation constraints are applied to the candidate values. The upper and lower limits are preferably determined by the maximum allowable current limit of the motor and the current amplitude of the quadrature-axis current reference value, so that the amplitude of the current vector synthesized by the direct-axis current reference value and the quadrature-axis current reference value does not exceed the maximum current limit. This ensures the effectiveness of the optimization update while avoiding the triggering of overcurrent protection or the degradation of control performance caused by current loop saturation. Finally, the updated direct-axis current reference value that meets the safety range is output for subsequent motor current control.

[0113] In an embodiment of the present invention, obtaining a quadrature-axis current reference value, determining a final current reference value based on the quadrature-axis current reference value and an updated direct-axis current reference value, and controlling the motor operation include:

[0114] The current vector magnitude is synthesized based on the updated direct-axis current reference value and quadrature-axis current reference value;

[0115] If the current vector amplitude exceeds the maximum allowable current limit of the motor, then calculate the ratio coefficient between the maximum current limit and the current vector amplitude;

[0116] Based on the scaling factor, the direct-axis current reference value and the quadrature-axis current reference value are simultaneously and proportionally reduced to obtain the final current reference value;

[0117] Specifically, after completing the online update of the direct-axis current reference value, the quadrature-axis current reference value is used as the direct execution quantity to meet the power assist requirement. The quadrature-axis current reference value is determined based on the power assist requirement, which is preferably determined by the target power assist torque given by the steering wheel input signal and vehicle speed and other operating conditions. The target power assist torque is converted into the quadrature-axis current reference value based on the relationship between the motor torque constant and the current coordinate transformation. A change rate limit is applied to the quadrature-axis current reference value to avoid power assist jitter caused by sudden current changes. The change rate limit is preferably determined according to the motor current loop bandwidth and is further preferably selected so that the quadrature-axis current reference value can complete the full-scale change within ten milliseconds to fifty milliseconds.

[0118] Specifically, after obtaining the updated direct-axis current reference value and quadrature-axis current reference value, the magnitude of the current vector synthesized from the two is calculated to characterize the extent to which the current reference current occupies the thermal constraints of the inverter and motor. The magnitude of the current vector is calculated using the following formula:

[0119]

[0120] In the formula, Let the magnitude of the current vector be the value at the k-th sampling time. This is the updated direct-axis current reference value at the k-th sampling time. The reference value of the quadrature-axis current at the k-th sampling time is calculated based on the fact that the Euclidean norm of the orthogonal component synthesis can characterize the actual magnitude of the current vector and directly correspond to the current constraints of the power device and winding.

[0121] The current vector amplitude is compared with a preset maximum current limit. The maximum current limit is preferably determined by the motor's rated current, the inverter's continuous current capability, and thermal constraints. More preferably, it is the rated current with a short-term overload factor, preferably between 1.2 and 1.5. This value is chosen because the power steering system has short-term peak demand and is subject to winding temperature rise and device junction temperature constraints. When the current vector amplitude does not exceed the maximum current limit, the updated direct-axis current reference value and quadrature-axis current reference value are directly output as the final current reference value. When the current vector amplitude exceeds the maximum current limit, the ratio of the maximum current limit to the current vector amplitude is calculated. This ratio is calculated using the following formula:

[0122]

[0123] In the formula, The scaling factor is the value at the k-th sampling time. This represents the maximum allowable current limit for the motor. Let be the magnitude of the current vector at the k-th sampling time.

[0124] Specifically, by multiplying the direct-axis current reference value and the quadrature-axis current reference value by a proportional coefficient, they are reduced synchronously and proportionally, so that the amplitude of the reduced current vector returns to within the maximum current limit. This satisfies the current constraint and maintains the consistency of the current vector direction without changing the relative proportional relationship between the direct-axis and quadrature-axis current reference values. The controller inputs the final direct-axis current reference value and the final quadrature-axis current reference value to the current closed-loop regulator, and generates a voltage command by combining the sampled actual current components. This command is then modulated to drive the inverter to output the phase voltage, thereby realizing the current vector control of the steering wheel control motor and balancing the power assist requirement with efficiency optimization across the entire speed range.

[0125] like Figure 2 As shown, Figure 2 This diagram illustrates the physical model and power composition of a torsion bar. It shows the angular relationship and power composition between the input and output sides of the torsion bar. The steering wheel side corresponds to the input angle, and the actuator side corresponds to the output angle. The torsion amount is determined by the difference between the input and output angles. The equivalent torsional stiffness of the torsion bar characterizes its ability to recover torque within its elastic working range. Regarding power relationships, the DC input power, as the source of power for the motor drive system, is divided into two parts: net mechanical power related to effective power delivery and elastic potential energy related to the torsion bar's elastic energy storage. The net mechanical power serves as a measure to reflect the power level of the net power output. The elastic potential energy increases or decreases with the torsion state of the torsion bar, and its rate of change forms the torsion bar's reciprocating energy storage power. This reciprocating energy storage power represents the reversible power fluctuation caused by reciprocating energy storage, characterizing the impact of the torsion bar's elastic energy storage and release on the input power fluctuation.

[0126] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An online optimization control system for achieving optimal efficiency of a motor over a wide speed range, characterized in that, include: The data acquisition module is used to collect data on DC bus voltage, DC bus current, torsion bar input side angle, and torsion bar output side angle, and to calculate DC side input power and torsion bar torsion based on the collected data; The energy storage index module is used to determine the torsion bar's round-trip energy storage power based on the torsion bar's torsion amount and its rate of change over time, and to calculate the percentage index of the torsion bar's round-trip energy storage power within a preset time window. The residual evaluation module is used to calculate the projection correlation coefficient of the DC-side input power on the torsion bar reciprocating energy storage power, and to correct the DC-side input power based on the projection correlation coefficient to obtain the residual power evaluation quantity. The optimization direction module is used to obtain the direct-axis current reference value, apply alternating perturbations to the direct-axis current reference value, and determine the optimization update direction based on the difference in the residual power evaluation quantity during the perturbation period. The adaptive update module is used to generate adaptive weights based on the proportion index of the torsion bar reciprocating energy storage power, and to suppress the update amplitude of the optimization update direction according to the adaptive weights, so as to obtain the updated direct-axis current reference value. The synthesis and control module is used to acquire the quadrature-axis current reference value, determine the final current reference value based on the quadrature-axis current reference value and the updated direct-axis current reference value, and control the motor operation. The DC-side input power and torsion bar torsion are calculated based on the collected data, including: The DC bus voltage and DC bus current are multiplied to obtain the DC side input power. Calculate the difference between the input side angle and the output side angle of the torsion bar, and perform periodic conversion processing on the difference to obtain the continuous torsion bar torsion amount; The torsion bar's reciprocating energy storage power is determined based on its torsion amount and its rate of change over time, including: The torsion bar torsion amount is differentially calculated to obtain the torsional change rate; Obtain the equivalent torsional stiffness of the torsion bar, and multiply the equivalent torsional stiffness of the torsion bar, the current torsion amount of the torsion bar, and the rate of change of torsion to obtain the torsion bar round-trip energy storage power. Calculate the percentage index of the torsion bar reciprocating energy storage power within a preset time window, including: Within a preset time window, the root mean square value of the energy storage power of the torsion bar during round trip is calculated, and the output side angle of the torsion bar is differentially calculated to obtain the output side angular velocity. Within a preset time window, the torsion bar torque is calculated based on the torsion bar equivalent torsional stiffness and torsion bar torsion amount. The torsion bar torque is then multiplied by the output side angular velocity to obtain the net power assist mechanical power. Calculate the mean value of the net assist mechanical power and take the absolute value. Based on the ratio of the root mean square value of the torsion bar reciprocating energy storage power to the absolute value, obtain the proportion index of the torsion bar reciprocating energy storage power. Calculate the projection correlation coefficient of the DC-side input power onto the torsion bar reciprocating energy storage power, and correct the DC-side input power based on the projection correlation coefficient, including: Within a preset time window, the sum of the products of the DC input power and the torsion bar reciprocating energy storage power is taken as the first value; the sum of the squares of the torsion bar reciprocating energy storage power is taken as the second value. The projection correlation coefficient is obtained based on the ratio of the first value to the second value. The residual power is obtained by subtracting the product of the projection correlation coefficient and the torsion bar reciprocating energy storage power from the DC side input power. The average value of the residual power within the preset time window is used as the residual power evaluation metric. Alternating perturbations are applied to the direct-axis current reference value, and the optimization update direction is determined based on the differences in the residual power evaluation during the perturbation period, including: Positive and negative perturbations are alternately applied to the direct-axis current reference bias, and the residual power evaluation quantities during the application of positive perturbation and negative perturbation are calculated respectively. The efficiency gradient is obtained by calculating the difference between the residual power evaluation during a positive disturbance and the residual power evaluation during a negative disturbance, and dividing the difference by twice the disturbance amplitude.

2. The online optimization control system for achieving optimal efficiency of a motor over a wide speed range according to claim 1, characterized in that, Adaptive weights are generated based on the proportion index of torsion bar reciprocating energy storage power. The update magnitude of the optimization update direction is suppressed based on these adaptive weights, including: A function inversely proportional to the proportion of torsion bar reciprocating energy storage power is constructed to calculate adaptive weights; The update magnitude is obtained by multiplying the efficiency gradient by the preset update coefficient. The update amplitude is attenuated based on adaptive weights. The current direct-axis current reference value is superimposed with the attenuated update amplitude and combined with the current safety range limit to obtain the updated direct-axis current reference value.

3. The online optimization control system for achieving optimal efficiency of a motor over a wide speed range according to claim 1, characterized in that, The final current reference value is determined based on the quadrature-axis current reference value and the updated direct-axis current reference value, including: The current vector magnitude is synthesized based on the updated direct-axis current reference value and quadrature-axis current reference value; If the current vector amplitude exceeds the maximum allowable current limit of the motor, then calculate the ratio coefficient between the maximum current limit and the current vector amplitude; The direct-axis current reference value and the quadrature-axis current reference value are simultaneously and proportionally reduced based on the scaling factor to obtain the final current reference value.

4. The online optimization control system for achieving optimal efficiency of a motor over a wide speed range according to claim 1, characterized in that, The periodic conversion process limits the angle difference to a preset minimum equivalent period range to eliminate discontinuous value jumps caused by the zero-crossing flip of the angle sensor.

Images