Semi-active suspension damping control method with vehicle speed and pavement grade fused with excitation intensity

Abstract

The invention relates to a semi-active suspension damping control method with vehicle speed and road surface grade fused excitation intensity. The method comprises the following steps that 1, acceleration and displacement signals and vehicle speed signals of corresponding point positions are obtained; 2, obtaining a pavement grade parameter reflecting the pavement unevenness level; 3, the equivalent road surface vibration excitation intensity borne by the suspension system under the current working condition is comprehensively represented; 4, determining a weight relationship between the comfort index and the stability index in damping control; 5, the target damping coefficient is made to be within the preset minimum damping and maximum damping range; and 6, real-time adjustment of damping of the shock absorber is achieved. The invention has the following beneficial effects: 1, damping adjustment can be adaptively adjusted along with the change of working conditions; 2, the problem of performance imbalance of a single control target under a complex working condition is avoided; 3, the effective utilization rate of the shock absorber is improved; 4, the method can be directly realized on an existing semi-active shock absorber hardware platform easily; and 5, the method has good universality and engineering popularization value.

Description

Technical Field

[0001] This invention relates to the field of vehicle suspension control technology, specifically to a semi-active suspension damping control method that integrates vehicle speed and road surface grade excitation intensity. Background Technology

[0002] The vehicle suspension system is a crucial component connecting the vehicle body and wheels. Its primary function is to isolate the vehicle body from vibrations caused by road unevenness, thereby improving ride comfort, handling stability, and driving safety. With increasing vehicle speeds and more complex road conditions, traditional passive suspensions, due to their fixed damping and stiffness parameters, struggle to simultaneously meet the demands of comfort and stability under varying speeds and road conditions, thus failing to satisfy the comprehensive performance requirements of modern vehicles. Semi-active suspensions, by adjusting the damping of the shock absorbers, achieve adaptive adjustment of vehicle vibration characteristics without introducing additional energy. They offer advantages such as low energy consumption, relatively simple structure, and high engineering feasibility, and are increasingly being researched and applied.

[0003] Currently, semi-active suspension damping control methods mainly include skyhook control, groundhook control, hybrid control, fuzzy control, and optimal control methods based on state feedback. Most existing control methods primarily adjust damping based on instantaneous vehicle state variables (such as sprung acceleration, suspension relative velocity, or displacement). Their control strategies implicitly assume relatively fixed vehicle operating conditions and do not explicitly consider the impact of vehicle speed variations on road excitation input characteristics. In actual driving, the spectral characteristics and energy distribution of the road excitation experienced by the vehicle change significantly with vehicle speed. Under different road surface grades and different speed combinations, the damping adjustment requirements of the suspension system vary significantly. However, existing semi-active suspension damping control methods are mostly based on fixed control logic or single-condition assumptions, making it difficult to simultaneously consider the coupled influence of vehicle speed and road surface grade on suspension performance requirements, resulting in limited control effectiveness under certain conditions. In existing technologies, vehicle speed information and road surface information are typically processed separately, or only used in control decisions through discrete rules. A unified fusion index that continuously and consistently reflects the vehicle's operating state and external road surface excitation characteristics has not yet been developed to guide the adaptive adjustment of semi-active suspension damping parameters. This limits the control strategy's responsiveness to complex operating conditions. Under complex road conditions and multiple vehicle speeds, using fixed values ​​or simple switching of damping parameters can easily lead to unreasonable trade-offs between comfort and handling stability in the suspension system, making it difficult to achieve optimized control across all operating conditions.

[0004] Existing technologies have not yet provided an effective method for continuously adjusting damping parameters based on comprehensive operating condition indices. After introducing vehicle speed and road surface grade information into semi-active suspension damping control, the system structure and control logic become more complex. Without systematic modeling and theoretical analysis, problems such as control instability or sensitivity to parameter uncertainties and external disturbances can easily arise. Therefore, this invention proposes a semi-active suspension damping control method based on a fusion index of vehicle speed and road surface grade. This technical solution uses vehicle operating state and road excitation information as a foundation, and by constructing a fusion index, adaptively adjusts the damping parameters of the semi-active suspension damper, thereby achieving dynamic optimization of suspension performance under different vehicle speeds and road surface grades, and improving the overall driving performance of the vehicle under complex operating conditions. Summary of the Invention

[0005] In summary, to overcome the shortcomings of the prior art, the technical problem to be solved by the present invention is to provide a semi-active suspension damping control method that integrates vehicle speed and road surface grade excitation intensity. It aims to solve the problem that the existing semi-active suspension damping control methods cannot simultaneously and comprehensively consider the changes in vehicle speed and the differences in road surface grade, resulting in the inaccurate characterization of the vibration excitation borne by the suspension system under different working conditions and the insufficient adaptability of damping adjustment.

[0006] The technical solution of this invention to solve the above-mentioned technical problems is as follows: a semi-active suspension damping control method that integrates vehicle speed and road surface grade excitation intensity, comprising the following steps:

[0007] Step 1: Collect the raw acceleration signal and vehicle speed signal through the signal sensing system, and define the relative motion parameters of the suspension after filtering and integration to obtain the suspension system state parameters and vehicle speed signal;

[0008] Step 2: Based on the suspension system state parameters obtained in Step 1, the road surface grade feature quantity is constructed after normalizing the acceleration. The relative road surface grade index is obtained by benchmarking against the reference working condition and then mapped to the discrete road surface grade to obtain the road surface grade feature quantity, the relative road surface grade index and the discrete road surface roughness grade parameter.

[0009] Step 3: Based on the road surface grade parameters obtained in Step 2 and the vehicle speed signal obtained in Step 1, the coupled excitation relationship between vehicle speed and road surface grade is derived. After determining the weight index through simulation fitting, a fusion index is constructed in multiplicative form to obtain a fusion excitation intensity index that can comprehensively characterize the equivalent vibration excitation intensity of the suspension.

[0010] Step 4: Based on the fusion excitation intensity index obtained in Step 3, construct the comprehensive performance index of the suspension, analyze the risk law of comfort and stability in combination with random vibration theory, define and derive the dynamic weight function of the two, and obtain the comfort and stability weight coefficients that adapt to the fusion excitation intensity.

[0011] Step 5: Based on the weighting coefficients obtained in Step 4 and the suspension system state parameters obtained in Step 1, the optimal damping under the comfort and stability objectives is derived respectively. After fusion and physical constraint saturation processing, the final target damping coefficient that satisfies the physical constraints of the semi-active suspension is obtained.

[0012] Step 6: Based on the target damping coefficient obtained in Step 5, the target damping force is derived by combining the suspension relative velocity in Step 1. After multi-stage optimization, it is converted into a PWM drive signal and applied to the shock absorber. The control current and PWM drive signal that meet the hardware constraints are obtained, so as to realize the real-time adjustment of the shock absorber damping.

[0013] Based on the above technical solution, the present invention can be further improved as follows:

[0014] Furthermore, step 1 specifically involves:

[0015] Step 1.1, Signal Acquisition: Accelerometers are placed at the sprung and unsprung masses of the suspension to acquire the raw vertical acceleration signal of the sprung mass. Vertical acceleration signal of unsprung mass Vehicle speed signals are obtained through the vehicle's internal controller area network.

[0016] Step 1.2, Hardware RC Low-Pass Filter: Remove irrelevant high-frequency components from the original acceleration signal. The formula is:

[0017]

[0018] ;

[0019] Where R and C are the resistance and capacitance values ​​of the filter circuit, respectively, satisfying... , and The acceleration signal is after hardware filtering. The sampling period;

[0020] Step 1.3, Mean Filtering: Further smoothing the hardware-filtered signal, using the following formula:

[0021] ;

[0022] ;

[0023] in, At the current sampling time, The length of the sliding window;

[0024] Step 1.4, use trapezoidal integral to calculate velocity and displacement: Integrate the acceleration after mean filtering to obtain velocity, and then integrate the velocity to obtain displacement. The formula is:

[0025] ;

[0026] ;

[0027] in, The velocity at the previous moment, This represents the displacement at the previous moment;

[0028] Step 1.5, Define relative motion parameters: Define the relative velocity and displacement variables of the suspension as follows: The state parameters of the suspension system are formed by combining the filtered acceleration, velocity, and displacement.

[0029] Furthermore, step 2 specifically involves:

[0030] Step 2.1, Acceleration Normalization: Eliminate the differences in physical characteristics, statistical characteristics, and magnitude of acceleration between sprung and unsprung masses. The formula is:

[0031] ;

[0032] ;

[0033] in, This is the average value of the acceleration;

[0034] Step 2.2, Constructing pavement grade characteristic quantities: These are constructed from the equivalent energy of sprung and unsprung mass accelerations, and weighting coefficients are introduced to balance their contributions to pavement roughness characterization. The formula is as follows:

[0035] ;

[0036] in, , Based on statistical energy proportions: , ; For time windows;

[0037] Step 2.3, Determine the reference characteristic value: Select a reference working condition to obtain the suspension response reference characteristic value, which serves as the road surface grade benchmark. The formula is:

[0038] ;

[0039] Step 2.4, calculate the relative pavement grade index: representing the ratio of the current pavement excitation to the reference pavement excitation, the formula is:

[0040]

[0041] Step 2.5, Mapping Discrete Road Surface Levels: Using Empirical Thresholds The relative pavement grade index is mapped to a discrete pavement roughness grade using the following formula:

[0042]

[0043] Furthermore, step 3 specifically involves:

[0044] Step 3.1, Deriving the coupling excitation relationship between vehicle speed and road surface: Based on the ISO8608 road surface model, the road surface displacement power spectral density is... Combined with time-domain incentives The time-domain power spectral density is obtained as follows: The suspension system's response to road surface input is as follows: ,in Define the transfer function of the suspension system; define the system response energy. After variable substitution ,have to This proves the effective excitation strength of the suspension. It is determined by both the road surface grade and the vehicle speed;

[0045] Step 3.2, Assume the excitation intensity relationship: Based on engineering experience and a random road surface white noise model, assume the following relationship between excitation intensity and road surface grade and vehicle speed: ,in, For different excitation intensities at different points, the gain coefficient is... As the base for road surface grade, For vehicle speed, For road surface weighting index, Speed ​​weighting index;

[0046] Step 3.3, Fit the weighting index and construct the fusion index: Determine the weighting index through simulation fitting. , The values ​​are selected, and the fusion incentive intensity index is constructed in multiplicative form. The formula is as follows: .

[0047] Furthermore, step 4 specifically involves:

[0048] Step 4.1, Constructing a Comprehensive Performance Index: Taking into account both comfort and handling stability, a comprehensive performance index for the two-degree-of-freedom semi-active suspension system is constructed, with the following formula:

[0049] ;

[0050] in, To achieve the goal of ride comfort, the acceleration of sprung mass is suppressed; = Operational stability target: suppress wheel bounce;

[0051] Step 4.2, analyze the relationship between risk and fusion excitation intensity: the stability failure probability is the suspension response out-of-bounds probability, if the response approximates a Gaussian distribution: ,but: and Therefore, the stability risk function It exhibits an exponential change; the comfort risk is approximately linearly related to the intensity of the fusion incentive, i.e. ;

[0052] Step 4.3, define the stability weight function: using the proportion of stability failures as the stability weight, the formula is:

[0053]

[0054] in, The physical critical excitation is determined by the suspension travel limit and tire stiffness;

[0055] Step 4.4, Derive the comfort weight function: The comfort weight is the complement of the stability weight, and the formula is: .

[0056] Furthermore, step 5 specifically involves:

[0057] Step 5.1, Derive the optimal damping for comfort: from the sprung mass dynamics formula Substitute , = ,have to: Find the instantaneous extremum of the comfort performance cost function: Solving for: To avoid This leads to singularities, so a smooth sign function is introduced. and small positive numbers After correction, saturation treatment is performed. The formula is: The saturation function is defined as:

[0058]

[0059] in, Small positive number, , These are the upper and lower limits of damping, respectively;

[0060] Step 5.2, Determine the optimal damping for stability: The tire dynamic load is defined as the fluctuating component: Then the stability performance index can be expressed in the form of variance: Taking its derivative, we get This indicates that the stability index is a monotonic function with respect to damping and has no internal minimum. Therefore, the optimal damping for stability is... ;

[0061] Step 5.3: Combining the weighted coefficients to fuse comfort and stability optimal damping, the initial target damping coefficient is obtained, as shown in the formula:

[0062]

[0063] in, .

[0064] Furthermore, step 6 specifically involves:

[0065] Step 6.1, Derivation of the target damping force: By combining the definition of linear damping with a smoothing factor to eliminate the mathematical singularity caused by the velocity dead zone, the formula is as follows:

[0066]

[0067] in, The target damping coefficient, The relative speed of the suspension. It is a smoothing factor;

[0068] Step 6.2, Bilinear interpolation to determine the initial control current: Prior to this, bench tests were conducted to obtain the initial control current of the vibration damper at different excitation speeds. and different control currents The damping force F generated is organized into a target current mapping table indexed by velocity and force. The controller retrieves real-time data. and Grid interval , Define the normalization factor , The initial target control current is obtained through bilinear interpolation, and the formula is as follows:

[0069] ;

[0070] in, The reference current value stored at the node in the table;

[0071] Step 6.3, Current slope limitation: To prevent sudden current changes from causing shock to the damper and damage to the electromagnetic coil, the formula is:

[0072] ;

[0073] in, , The preset maximum allowable current slope, To control the cycle, This represents the actual output current at the previous moment;

[0074] Step 6.4, Hysteresis Compensation: Superimpose a high-frequency, small sinusoidal signal to eliminate the hysteresis effect and static friction of the electromagnetic actuator. The formula is: ;in, The signal amplitude, This refers to the jitter frequency;

[0075] Step 6.5, converting PI closed-loop feedback to PWM signal: First, calculate the basic duty cycle, the formula is: ,

[0076] in, For the damper coil resistance, The equivalent resistance of the line. The on-board power supply provides real-time voltage; a PI closed-loop compensation system is introduced to compensate for coil resistance temperature drift and current deviation. The final duty cycle formula is:

[0077] ;

[0078] in, This is the proportionality coefficient. The integral coefficient is... This is the measured current;

[0079] Step 6.6, Power Constraint: Ensure the semi-active damper only dissipates energy and does not generate energy, and determine the power consumption state. :like Then the calculated result will be executed normally. ;like Forced settings The final PWM drive signal is applied to the semi-active damper to achieve real-time damping adjustment.

[0080] The beneficial effects of this invention are:

[0081] 1. This invention integrates vehicle speed information and road surface grade information to construct a fusion excitation intensity index that reflects the vibration excitation level under actual working conditions. This index is then used as the basis for damping control adjustment. This overcomes the problem of insufficient control targeting caused by relying solely on suspension state feedback and ignoring the influence of vehicle speed in existing semi-active suspension control methods. This allows damping adjustment to adaptively adjust with changes in working conditions.

[0082] 2. By dynamically determining the weight allocation of comfort and stability objectives in control based on the intensity of fusion excitation, this invention can focus on reducing sprung vibration and improving driving comfort under low excitation conditions, and focus on suppressing suspension travel and tire dynamic load and improving vehicle stability under high excitation conditions, effectively avoiding the problem of performance imbalance of a single control objective under complex conditions.

[0083] 3. This invention guides the damping level by integrating indicators to ensure that the damping always operates within a reasonable range that matches the current road surface and vehicle speed. This avoids the waste of control resources and performance degradation caused by traditional semi-active suspension frequently tending towards maximum or minimum damping under complex road conditions, and improves the effective utilization rate of the shock absorber.

[0084] 4. This invention explicitly introduces upper and lower limits of damping and current constraints in the damping calculation and control current generation process, ensuring that the control process does not inject additional energy into the system, which meets the physical characteristics and safety requirements of semi-active suspension and is easy to implement directly on existing semi-active damper hardware platforms.

[0085] 5. This invention adopts a modular control process, which sequentially realizes road level identification, vehicle speed normalization, excitation fusion, weight allocation and damping control. Each module is independent of each other but logically related. The parameters can be flexibly adjusted according to different vehicle models, different shock absorber types or different control objectives, and it has good versatility and engineering promotion value. Attached Figure Description

[0086] Figure 1 This is an overall flowchart of the present invention;

[0087] Figure 2 This is a diagram of a two-degree-of-freedom semi-active suspension model.

[0088] Figure 3 The power function fitting curves are shown for different road surface grades at a vehicle speed of 60 km / h.

[0089] Figure 4 The curves are fitted to the power function at different vehicle speeds on Class C roads.

[0090] Figure 5 The diagram shows the sprung acceleration at speeds of 40-90 km / h on Class C roads.

[0091] Figure 6 The diagram shows the tire dynamic load at speeds of 40-90 km / h on Class C road surfaces.

[0092] Figure 7 The sprung acceleration diagrams for various road surfaces at 60 km / h are shown.

[0093] Figure 8 The diagram shows the tire dynamic load on various road surfaces at 60 km / h.

[0094] Figure 9 The sprung acceleration diagrams are shown for varying speeds and road surface conditions.

[0095] Figure 10 This is a diagram showing the tire dynamic load under varying speeds and road surface conditions. Detailed Implementation

[0096] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0097] like Figure 1 As shown, the semi-active suspension damping control method that integrates vehicle speed and road surface grade excitation intensity includes the following steps:

[0098] Step 1: Acquire raw acceleration signals and vehicle speed signals through a signal sensing system. After filtering and integration, define the relative motion parameters of the suspension to obtain the suspension system state parameters and vehicle speed signals. Specifically:

[0099] Step 1.1: The signal sensing system of the present invention consists of two accelerometers, the layout of which is shown in the figure. Figure 2 As shown. Figure 2 As shown, the acceleration sensors of this invention are respectively installed on the sprung mass and unsprung mass of the suspension to measure the raw signal of the vertical acceleration of the sprung mass. Vertical acceleration signal of unsprung mass In one embodiment of the present invention, the vehicle speed signal can be obtained through the vehicle's internal controller area network (CAN) communication. During vehicle operation, the relevant control unit in the vehicle's electronic control system calculates the vehicle speed based on wheel speed, transmission system status, or vehicle dynamics information, and periodically sends the corresponding speed signal frame through the CAN bus.

[0100] Step 1.2: Due to factors such as sensor noise, local high-frequency vibrations of the structure, and electromagnetic interference, the acquired raw acceleration signal usually contains high-frequency components unrelated to the suspension system dynamics, requiring filtering of the acquired raw signal. A hardware RC low-pass filter preprocessing step is performed on the acquired raw signal, with the following formula:

[0101]

[0102]

[0103] Where R and C are the resistance and capacitance values ​​of the filter circuit, satisfying... , and The acceleration signal is after hardware filtering. The sampling period.

[0104] Step 1.3: Further smooth the hardware-filtered signal by taking the average of N consecutive sampling points. The smoothing filter formula is as follows: and ,in: At the current sampling time, The length of the sliding window is taken in the suspension control. =5.

[0105] Step 1.4: Calculate the spring acceleration after mean filtering. and unsprung acceleration The sprung velocity is obtained by integrating using the trapezoidal integral formula. and unsprung speed The formula is: ,in, : This represents the velocity at the previous moment.

[0106] After integration, the sprung speed and unsprung speed The sprung displacement is obtained by integrating the trapezoidal formula again. and unsprung displacement The formula is: ,in : represents the displacement at the previous moment.

[0107] Step 1.5: Define the relative velocity and displacement variables of the suspension as follows: .

[0108] Step 2: Based on the suspension system state parameters obtained in Step 1, road surface grade characteristic quantities are constructed after normalizing the acceleration. The relative road surface grade index is obtained by benchmarking against a reference working condition and mapped to the discrete road surface grade, resulting in road surface grade characteristic quantities, relative road surface grade index, and discrete road surface roughness grade parameters. Specifically:

[0109] Step 2.1: During vehicle operation, road surface unevenness is transmitted to the unsprung and sprung masses through the tire-suspension system, essentially manifesting as changes in system response energy and vibration intensity. Therefore, road surface grade can be equivalently characterized by the dynamic response characteristics of the suspension system to road input. Figure 2 It can be known that its dynamic equation can be expressed as:

[0110]

[0111] Road surface unevenness does not need to be measured directly; rather, its severity can be inferred from the system response.

[0112] This invention uses the characteristic values ​​of sprung mass acceleration and unsprung mass acceleration to characterize road surface roughness. However, these values ​​have different physical sources, statistical characteristics, and amplitude magnitudes, therefore they need to be normalized. The filtered sprung acceleration is obtained from step 1. and spring acceleration The data is normalized for reverse estimation of road surface roughness. The normalized data is as follows: , ,in and The acceleration data from 5 sampling points were used to calculate the average.

[0113] Step 2.2: The pavement grade characteristic quantity is jointly constructed from the equivalent energy of the accelerations of the sprung mass and the unsprung mass. To balance the contributions of the two types of responses to the pavement roughness characterization, weighting coefficients are introduced to weight different responses within a time window. Within this framework, road surface grade feature quantities are constructed in the following form: ,in Based on statistical energy proportions: , .

[0114] Step 2.3: Select a reference operating condition. In this invention, the reference vehicle speed is... Reference road surface: ISOC grade road surface. Under this reference working condition, the reference characteristic values ​​of the suspension response are obtained through simulation or experiment as a road surface grade benchmark.

[0115]

[0116] Step 2.4: Define the relative road surface grade index as the ratio of the equivalent excitation intensity sensed by the suspension under the current operating condition to the excitation intensity of the reference road surface: By discretizing the pavement grade based on its indicators, the corresponding pavement roughness grade can be obtained:

[0117]

[0118] in, This is an empirical threshold.

[0119] Step 3: Based on the road surface grade parameters obtained in Step 2 and the vehicle speed signal obtained in Step 1, the coupled excitation relationship between vehicle speed and road surface grade is derived. After determining the weighting index through simulation fitting, a fusion index is constructed in a multiplicative form to obtain a fusion excitation intensity index that can comprehensively characterize the equivalent vibration excitation intensity of the suspension. Specifically:

[0120] Step 3.1: In the ISO 8608 road surface model, the road surface displacement power spectral density is: This shows that road surface grade is a spatial statistic, not a temporal signal. However, in actual vehicle operation, the temporal excitation experienced by the vehicle and its speed v have the following relationship: Its time-domain power spectral density can be obtained from stochastic process theory: Meanwhile, since the equivalent excitation frequency is proportional to the vehicle speed: It can be seen that for the same road surface grade, the excitation spectrum is completely different at different vehicle speeds. From the above formula, the response of the suspension system to road input can be derived as follows: ,in The suspension system transfer function is highly sensitive near the vehicle's natural frequency. Substituting this into the road-vehicle speed relationship, we get: ,same ,different The energy distribution of the system response is completely different.

[0121] Define the system response energy: Substituting this into the above equation, we get: Substitute variables: ,have to Therefore, the effective excitation intensity that the suspension system withstands has the following relationship: The speed of the vehicle must be taken into account when constructing the excitation intensity function, rather than solely determined by the road surface grade.

[0122] Step 3.2: Based on engineering experience, it is known that on the same road surface, as the vehicle speed gradually decreases to zero, the road excitation experienced by the vehicle also gradually decreases until it reaches zero. Therefore, it can be inferred that the road excitation intensity experienced by the vehicle during driving is the product of the vehicle speed and the road surface grade, and based on the random road white noise model: The intensity of the excitation is related to the road surface grade and vehicle speed in the following way: Since road surface grade is the main factor affecting excitation intensity, we assume that the relationship between excitation intensity, road surface grade, and vehicle speed is as follows: ,in, : is the gain coefficient for the excitation intensity at different points. : This is the base value for road surface grade. : refers to vehicle speed. : is the road surface weight index, taken as By increasing the index, the influence of road surface grade is amplified, making it play a decisive role; For speed weighting index, take As speed increases, the excitation energy increases sublinearly.

[0123] Step 3.3: To determine the relationship of the coefficients in step 3.2, experiments were conducted on the same road surface at different vehicle speeds. The sprung and unsprung acceleration signals were collected as verification of the excitation intensity. A graphical relationship between excitation intensity and vehicle speed was plotted, and the speed weighting index was obtained by fitting this curve. Similarly, using the same speed on roads of different grades, the sprung and unsprung mass acceleration signals were collected as verification of the excitation intensity. An image of the excitation intensity versus vehicle speed was plotted, and the road weight index was obtained by fitting this curve. To save time, this invention uses simulation for verification, and the results are as follows: Figure 3 and Figure 4 As shown:

[0124]

[0125] Among them, the sprung mass Unsprung mass Suspension stiffness Tire stiffness Passive damping coefficient The values ​​obtained by fitting these values ​​are based on the above. =4.7, =0.5, Let (1, 2, 3, 4, 5, 6, 7, 8) be used. In order to comprehensively describe the relationship between the fusion excitation intensity, vehicle speed, and road surface grade, we can define... .

[0126] Step 4: Based on the fusion excitation intensity index obtained in Step 3, construct a comprehensive suspension performance index. Combine this with random vibration theory to analyze the risk patterns of comfort and stability, define and derive their dynamic weighting functions, and obtain the comfort and stability weighting coefficients that adaptively change with the fusion excitation intensity. Specifically:

[0127] Step 4.1: Construct the comprehensive performance index of the two-degree-of-freedom semi-active suspension system as follows: ,in, To achieve the goal of ride comfort, the acceleration of sprung mass is suppressed; = The objectives of operational stability and suppressing wheel wobbling are inherently mutually exclusive; therefore, the fusion excitation intensity obtained in step 3 needs to be considered. We will assign weight coefficients to these two performance metrics.

[0128] Step 4.2: In the study of a two-degree-of-freedom semi-active suspension system, comfort only needs to focus on the sprung mass acceleration. That's sufficient, but what stability truly concerns is whether the suspension will bottom out or top out and whether the tires will momentarily lift off the ground. Stability is a "probability problem," not a mean problem, and in a stochastic system, it can be represented as: In the theory of random vibrations, if the response is approximately Gaussian: The probability of exceeding the boundary is: , among them ,Right now and fusion incentive intensity It is proportional to the strength of the fusion excitation, therefore the risk function of stability can be obtained as it increases with the strength of the fusion excitation. If the risk incentive function changes exponentially, then it can be expressed as: This means that when the fusion incentive intensity When it's very small, the risk is almost zero; once... Approaching a certain critical value, the risk increases sharply; the risk to comfort is approximately linearly related to the intensity of the fusion incentive, i.e. .

[0129] Step 4.3: Redefine the stability weights rigorously, and define the stability failure function as follows: ,in The physical critical excitation determined by the suspension travel limit and tire stiffness can therefore be used as a stability weight. Since the risk to comfort can be approximated as linearly increasing: The stability weights after substitution are:

[0130]

[0131] Step 4.4: From the above formula, we can see that for small incentives: When the stability weight is almost zero, comfort dominates; when the fusion excitation intensity... At that time, the stability weight will rise rapidly, and as the incentive continues to increase until... hour, Stability is the dominant factor, which aligns with engineering control logic; therefore, the comfort weighting function is naturally: .

[0132] Step 5: Based on the weighting coefficients obtained in Step 4 and the suspension system state parameters obtained in Step 1, the optimal damping under the comfort and stability objectives is derived respectively. After fusion and physical constraint saturation processing, the final target damping coefficient satisfying the physical constraints of the semi-active suspension is obtained. Specifically:

[0133] Step 5.1: Comfort is measured by the acceleration of the sprung mass, derived from the sprung mass dynamics formula: Define the relative displacement variable of the suspension as: , = Then we have: Substituting this into the performance cost function, we get: Instantaneous optimal condition: in the current state and weight Given the damping coefficient Find the instantaneous extreme value: Substituting, we get: Set it to zero: Solving for: To avoid This leads to singularity, so a smooth sign function is introduced: Then the damping expression is modified to: ,in For small positive numbers, the damping coefficient must satisfy the following constraints: Ultimately, the optimal damping for comfort is defined as: ,Right now:

[0134]

[0135] Step 5.2: Tire dynamic load is defined as the fluctuating component: Then the stability performance index can be expressed in the form of variance: , due to Since it is a constant, the extremum problem is equivalent to:

[0136]

[0137] For the forces acting on the unsprung mass: Of these, only one is directly related to damping:

[0138] Therefore, we can know that when the mass on the spring... and unsprung mass During vibration, because the damping force is always opposite to the direction of the relative velocity, it can suppress the high-frequency, large-amplitude motion of the unsprung mass. The dynamic load on the tire is precisely caused by… Decision, and the emergence of stability indicators Taking the derivative, we get ,make Then there is ,when When the relative velocity is large (i.e., the tire is compressed or stretched), the suspension system will inevitably experience a large relative velocity, and the damping force and the relative velocity are in opposite directions. The CDC damper cannot inject active force; it can only reduce the amplitude of the relative motion. Therefore, we have... ,Right now This indicates that the stability index is a monotonic function with respect to damping and has no internal minimum. Therefore, we have... ,and Then the solution is: Then the final damping of the fusion control is: ,Right now:

[0139]

[0140] Right now:

[0141] Step 5.3: In summary, the final control law is:

[0142]

[0143] in:

[0144] It is a small positive number.

[0145] Step 6: Based on the target damping coefficient obtained in Step 5, and combined with the suspension relative velocity from Step 1, the target damping force is derived, optimized through multiple stages, and converted into a PWM drive signal, which is then applied to the shock absorber. This yields the control current and PWM drive signal that satisfy the hardware constraints, enabling real-time adjustment of the shock absorber damping. Specifically:

[0146] Step 6.1: Based on the current relative speed of the suspension obtained in Step 1. According to the definition of linear damping: This formula will exist when To prevent mathematical singularities in the calculated damping force when it is close to 0 (velocity dead zone), a smoothing factor needs to be introduced. ,but:

[0147]

[0148] Derive the target damping force required under the current operating conditions. .

[0149] Step 6.2: To overcome the nonlinear damping characteristics of the semi-active damper at different motion speeds, this invention uses a combination of a preset MAP table and a bilinear interpolation algorithm to accurately determine the control current. Prior to this, bench tests were conducted to obtain the damper's performance at different excitation speeds. and different control currents The damping force F generated is organized into a target current mapping table indexed by velocity and force. The controller obtains the real-time speed. and target damping force Afterwards, search The table determines the grid interval to which the current operating condition belongs, satisfying... and Define the normalization factor:

[0150]

[0151] Then the target control current Solve using the following formula:

[0152] ,in, The reference current value stored at the node in the table.

[0153] Step 6.3: [The target current is then applied.] Before converting the signal into an actual drive signal, the physical limitations of the hardware system must be considered. Directly switching the current without constraints can cause the damper to experience enormous impact loads, and may even damage the electromagnetic coil. The electromagnetic coil has inductive reactance characteristics, meaning the current cannot change instantaneously (following...). To avoid misalignment between control commands and physical responses, this scheme introduces a slope limiter: Let... The actual output current at the previous moment, the control period is... The initial output at the current moment ,satisfy:

[0154]

[0155] in, , The preset maximum allowable current slope (unit: A / s).

[0156] Step 6.4: This ensures the continuity of the damping force, eliminates the "clunking" sound from the chassis caused by sudden current changes, and extends the life of the solenoid valve. Electromagnetic actuators exhibit hysteresis, meaning the damping force curves corresponding to current rise and fall do not coincide. To eliminate this nonlinear error, a high-frequency, small sinusoidal signal is superimposed on the output current. The dynamic signal causes the valve core to vibrate at a low frequency, converting static friction into dynamic friction. Among them, jitter frequency Usually taken It is much lower than the PWM carrier frequency but higher than the mechanical response frequency.

[0157] Step 6.4: Obtain the target current Next, it needs to be converted into a hardware-executable PWM signal. According to Ohm's law and the characteristics of the resistive-inductive load, the duty cycle in steady state is... The relationship with current is:

[0158]

[0159] in: : This represents the resistance of the damper coil; : is the equivalent resistance of the line; : This refers to the real-time voltage of the vehicle's power supply.

[0160] Step 6.5: Because the coil resistance increases with increasing temperature. Open-loop control alone cannot guarantee accuracy. Therefore, this step introduces current closed-loop feedback: let the measured current at the current moment be... Then the current deviation The final output duty cycle Proportional-integral (PI) control is used. This ensures that the magnetic field strength inside the actuator can accurately reach the level expected by the algorithm at any ambient temperature.

[0161] Step 6.6: During the calculation process, the physical boundary constraint formulas must be followed:

[0162] Semi-active dampers can only dissipate energy, not generate it. To prevent the damping force calculated by the algorithm from being in the same direction as the actual motion of the damper (i.e., becoming a "power source"), the following constraint logic must be implemented: Assume... For real-time relative velocity, The target damping force. The system needs to determine the power consumption status of both in real time. :

[0163] like Then the calculated result will be executed normally. ;

[0164] like The algorithm requires the shock absorber to do work (generate power), so the following settings are forced: .

[0165] Simulation experiment verification and analysis

[0166] To verify the effectiveness of the control strategy involved in this invention, a quarter-vehicle two-degree-of-freedom dynamic model was built in Simulink. A random excitation road surface white noise model was adopted. Considering that most vehicles operate on secondary cement roads and secondary asphalt roads, this simulation used B, C, D, and E level random road surface excitations, with vehicle speeds ranging from 40 to 90 kilometers per hour for verification.

[0167] First, under Class C random road excitation, the vehicle speed was uniformly accelerated from 40 km / h to 90 km / h, and the simulation time was 50 seconds. Passive damping, ceiling control, floor control, and acceleration control strategies were used as comparison objects to verify the advantages of the adaptive damping control strategy based on vehicle speed and road level fusion in this invention in reducing sprung acceleration and tire dynamic load when the vehicle speed changes. Figure 5 and Figure 6 For comparison of simulation results.

[0168] Table 1. Simulation results at different vehicle speeds on Class C road surfaces

[0169]

[0170] Table 1 shows that the algorithm of this invention improves the average sprung acceleration by 5.7% compared to the ceiling algorithm and by 15.6% compared to the passive algorithm; it also improves tire dynamic load by 5.3% compared to the ceiling algorithm and by 1.9% compared to the passive algorithm. Both improvements demonstrate that this algorithm can enhance stability and comfort on roads with varying speeds.

[0171] Secondly, at a vehicle speed of 60 km / h, random road surface excitations were applied sequentially from Class B, Class C, Class D, and Class E. The simulation time was 40 seconds, with 10 seconds for each road surface level. Passive damping, ceiling control, floor control, and acceleration control strategies were used as comparison objects to verify the advantages of the adaptive damping control strategy based on vehicle speed and road surface level fusion in reducing sprung acceleration and tire dynamic load when the road surface level changes. Figure 7 and Figure 8 For comparison of simulation results.

[0172] Table 2. Simulation results of road surfaces at 60km / h

[0173]

[0174] As shown in the table above, the algorithm of this invention improves the average sprung acceleration by 5.0% compared to the ceiling algorithm and by 16.2% compared to the passive algorithm. It also improves the tire dynamic load by 5.1% compared to the ceiling algorithm and by 3.5% compared to the passive algorithm. All of these improvements are better than the ceiling algorithm, indicating that this algorithm can improve stability and comfort at the same speed on different road surfaces.

[0175] Finally, under B-level random road surface excitation, the vehicle travels at 90 km / h for 10 seconds, then decelerates uniformly to 70 km / h for 5 seconds before entering a C-level random road surface. It maintains 70 km / h on the C-level random road surface for 10 seconds, then decelerates uniformly to 60 km / h for 5 seconds before entering a D-level random road surface. It maintains 60 km / h on the D-level random road surface for 10 seconds, then decelerates uniformly to 40 km / h for 10 seconds before entering an E-level random road surface. Finally, it maintains 40 km / h on the E-level random road surface for 10 seconds. Passive damping, roof control, ground control, and acceleration control strategies are used as comparisons to verify the advantages of the adaptive damping control strategy based on vehicle speed and road surface level fusion in reducing sprung acceleration and tire dynamic load when both vehicle speed and road surface level change. Figure 9 and Figure 10 For comparison of simulation results.

[0176] Table 3. Simulation results for speed and varying road surface grades

[0177]

[0178] As shown in the table above, the algorithm of this invention improves the average sprung acceleration by 5.2% compared to the ceiling algorithm and by 16% compared to the passive algorithm. It also improves the tire dynamic load by 6.4% compared to the ceiling algorithm and by 3.3% compared to the passive algorithm. All of these improvements demonstrate that the algorithm can enhance stability and comfort at the same speed on different road surfaces.

[0179] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A semi-active suspension damping control method that integrates vehicle speed and road surface grade excitation intensity, characterized in that, The steps include the following: Step 1: Collect the raw acceleration signal and vehicle speed signal through the signal sensing system, and define the relative motion parameters of the suspension after filtering and integration to obtain the suspension system state parameters and vehicle speed signal; Step 2: Based on the suspension system state parameters obtained in Step 1, the road surface grade feature quantity is constructed after normalizing the acceleration. The relative road surface grade index is obtained by benchmarking against the reference working condition and then mapped to the discrete road surface grade to obtain the road surface grade feature quantity, the relative road surface grade index and the discrete road surface roughness grade parameter. Step 3: Based on the road surface grade parameters obtained in Step 2 and the vehicle speed signal obtained in Step 1, the coupled excitation relationship between vehicle speed and road surface grade is derived. After determining the weight index through simulation fitting, a fusion index is constructed in multiplicative form to obtain a fusion excitation intensity index that can comprehensively characterize the equivalent vibration excitation intensity of the suspension. Step 4: Based on the fusion excitation intensity index obtained in Step 3, construct the comprehensive performance index of the suspension, analyze the risk law of comfort and stability in combination with random vibration theory, define and derive the dynamic weight function of the two, and obtain the comfort and stability weight coefficients that adapt to the fusion excitation intensity. Step 5: Based on the weighting coefficients obtained in Step 4 and the suspension system state parameters obtained in Step 1, the optimal damping under the comfort and stability objectives is derived respectively. After fusion and physical constraint saturation processing, the final target damping coefficient that satisfies the physical constraints of the semi-active suspension is obtained. Step 6: Based on the target damping coefficient obtained in Step 5, the target damping force is derived by combining the suspension relative velocity in Step 1. After multi-stage optimization, it is converted into a PWM drive signal and applied to the shock absorber. The control current and PWM drive signal that meet the hardware constraints are obtained, so as to realize the real-time adjustment of the shock absorber damping.

2. The semi-active suspension damping control method based on the fusion of vehicle speed and road surface grade excitation intensity as described in claim 1, characterized in that, Step 1 is as follows: Step 1.1, Signal Acquisition: Accelerometers are placed at the sprung and unsprung masses of the suspension to acquire the raw vertical acceleration signal of the sprung mass. Vertical acceleration signal of unsprung mass Vehicle speed signals are obtained through the vehicle's internal controller area network. Step 1.2, Hardware RC Low-Pass Filter: Remove irrelevant high-frequency components from the original acceleration signal. The formula is: ； Where R and C are the resistance and capacitance values ​​of the filter circuit, respectively, satisfying... , and The acceleration signal is after hardware filtering. The sampling period; Step 1.3, Mean Filtering: Further smoothing the hardware-filtered signal, using the following formula: ； ； in, At the current sampling time, The length of the sliding window; Step 1.4, use trapezoidal integral to calculate velocity and displacement: Integrate the acceleration after mean filtering to obtain velocity, and then integrate the velocity to obtain displacement. The formula is: ； ； in, The velocity at the previous moment, This represents the displacement at the previous moment; Step 1.5, Define relative motion parameters: Define the relative velocity and displacement variables of the suspension as follows: The state parameters of the suspension system are formed by combining the filtered acceleration, velocity, and displacement.

3. The semi-active suspension damping control method based on the fusion of vehicle speed and road surface grade excitation intensity according to claim 2, characterized in that, Step 2 is as follows: Step 2.1, Acceleration Normalization: Eliminate the differences in physical characteristics, statistical characteristics, and magnitude of acceleration between sprung and unsprung masses. The formula is: ； ； in, This is the average value of the acceleration; Step 2.2, Constructing pavement grade characteristic quantities: These are constructed from the equivalent energy of sprung and unsprung mass accelerations, and weighting coefficients are introduced to balance their contributions to pavement roughness characterization. The formula is as follows: ； in, , Based on statistical energy proportions: , ; For time windows; Step 2.3, Determine the reference characteristic value: Select a reference working condition to obtain the suspension response reference characteristic value, which serves as the road surface grade benchmark. The formula is: ； Step 2.4, calculate the relative pavement grade index: representing the ratio of the current pavement excitation to the reference pavement excitation, the formula is: Step 2.5, Mapping Discrete Road Surface Levels: Using Empirical Thresholds The relative pavement grade index is mapped to a discrete pavement roughness grade using the following formula:

4. The semi-active suspension damping control method based on the fusion of vehicle speed and road surface grade excitation intensity according to claim 3, characterized in that, Step 3 specifically involves: Step 3.1, Deriving the coupling excitation relationship between vehicle speed and road surface: Based on the ISO8608 road surface model, the road surface displacement power spectral density is... Combined with time-domain incentives The time-domain power spectral density is obtained as follows: The suspension system's response to road surface input is as follows: ,in Define the transfer function of the suspension system; define the system response energy. After variable substitution ,have to This proves the effective excitation strength of the suspension. It is determined by both the road surface grade and the vehicle speed; Step 3.2, Assume the excitation intensity relationship: Based on engineering experience and a random road surface white noise model, assume the following relationship between excitation intensity and road surface grade and vehicle speed: ,in, For different excitation intensities at different points, the gain coefficient is... As the base for road surface grade, For vehicle speed, For road surface weighting index, Speed ​​weighting index; Step 3.3, Fit the weighting index and construct the fusion index: Determine the weighting index through simulation fitting. , The values ​​are selected, and the fusion incentive intensity index is constructed in multiplicative form. The formula is as follows: .

5. The semi-active suspension damping control method based on the fusion of vehicle speed and road surface grade excitation intensity according to claim 4, characterized in that, Step 4 specifically involves: Step 4.1, Constructing a Comprehensive Performance Index: Taking into account both comfort and handling stability, a comprehensive performance index for the two-degree-of-freedom semi-active suspension system is constructed, with the following formula: ； in, To achieve the goal of ride comfort, the acceleration of sprung mass is suppressed; = Operational stability target: suppress wheel bounce; Step 4.2, analyze the relationship between risk and fusion excitation intensity: the stability failure probability is the suspension response out-of-bounds probability, if the response approximates a Gaussian distribution: ,but: and Therefore, the stability risk function It exhibits an exponential change; the comfort risk is approximately linearly related to the intensity of the fusion incentive, i.e. ; Step 4.3, define the stability weight function: using the proportion of stability failures as the stability weight, the formula is: in, The physical critical excitation is determined by the suspension travel limit and tire stiffness; Step 4.4, Derive the comfort weight function: The comfort weight is the complement of the stability weight, and the formula is: .

6. The semi-active suspension damping control method based on the fusion of vehicle speed and road surface grade excitation intensity according to claim 5, characterized in that, Step 5 specifically involves: Step 5.1, Derive the optimal damping for comfort: from the sprung mass dynamics formula Substitute , = ,have to: Find the instantaneous extremum of the comfort performance cost function: Solving for: To avoid This leads to singularities, so a smooth sign function is introduced. and small positive numbers After correction, saturation treatment is performed. The formula is: The saturation function is defined as: in, Small positive number, , These are the upper and lower limits of damping, respectively; Step 5.2, Determine the optimal damping for stability: The tire dynamic load is defined as the fluctuating component: Then the stability performance index can be expressed in the form of variance: Taking its derivative, we get This indicates that the stability index is a monotonic function with respect to damping and has no internal minimum. Therefore, the optimal damping for stability is... ; Step 5.3: Combining the weighted coefficients to fuse comfort and stability optimal damping, the initial target damping coefficient is obtained, as shown in the formula: in, .

7. The semi-active suspension damping control method based on the fusion of vehicle speed and road surface grade excitation intensity according to claim 6, characterized in that, Step 6 specifically involves: Step 6.1, Derivation of the target damping force: By combining the definition of linear damping with a smoothing factor to eliminate the mathematical singularity caused by the velocity dead zone, the formula is as follows: in, The target damping coefficient, The relative speed of the suspension. It is a smoothing factor; Step 6.2, Bilinear interpolation to determine the initial control current: Prior to this, bench tests were conducted to obtain the initial control current of the vibration damper at different excitation speeds. and different control currents The damping force F generated is organized into a target current mapping table indexed by velocity and force. The controller retrieves real-time data. and Grid interval , Define the normalization factor , The initial target control current is obtained through bilinear interpolation, and the formula is as follows: ； in, The reference current value stored at the node in the table; Step 6.3, Current slope limitation: To prevent sudden current changes from causing shock to the damper and damage to the electromagnetic coil, the formula is: ； in, , The preset maximum allowable current slope, To control the cycle, This represents the actual output current at the previous moment; Step 6.4, Hysteresis Compensation: Superimpose a high-frequency, small sinusoidal signal to eliminate the hysteresis effect and static friction of the electromagnetic actuator. The formula is: ;in, The signal amplitude, This refers to the jitter frequency; Step 6.5, converting PI closed-loop feedback to PWM signal: First, calculate the basic duty cycle, the formula is: , in, For the damper coil resistance, The equivalent resistance of the line. The on-board power supply provides real-time voltage; a PI closed-loop compensation system is introduced to compensate for coil resistance temperature drift and current deviation. The final duty cycle formula is: ； in, This is the proportionality coefficient. The integral coefficient is... This is the measured current; Step 6.6, Power Constraint: Ensure the semi-active damper only dissipates energy and does not generate energy, and determine the power consumption state. :like Then the calculated result will be executed normally. ;like Forced settings The final PWM drive signal is applied to the semi-active damper to achieve real-time damping adjustment.

Images