CN111553021B - Design method of active suspension system based on cascade disturbance observer - Google Patents
Design method of active suspension system based on cascade disturbance observer Download PDFInfo
- Publication number
- CN111553021B CN111553021B CN202010340486.9A CN202010340486A CN111553021B CN 111553021 B CN111553021 B CN 111553021B CN 202010340486 A CN202010340486 A CN 202010340486A CN 111553021 B CN111553021 B CN 111553021B
- Authority
- CN
- China
- Prior art keywords
- observer
- active suspension
- suspension system
- disturbance
- sprung mass
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 239000000725 suspension Substances 0.000 title claims abstract description 117
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000013461 design Methods 0.000 title claims abstract description 30
- 230000009467 reduction Effects 0.000 claims abstract description 8
- 230000001133 acceleration Effects 0.000 claims description 36
- 230000005284 excitation Effects 0.000 claims description 36
- 238000002474 experimental method Methods 0.000 claims description 22
- 238000006073 displacement reaction Methods 0.000 claims description 15
- 238000013016 damping Methods 0.000 claims description 8
- 230000000694 effects Effects 0.000 claims description 7
- 238000005312 nonlinear dynamic Methods 0.000 claims description 7
- 230000008859 change Effects 0.000 claims description 3
- 230000002596 correlated effect Effects 0.000 claims description 3
- 230000001419 dependent effect Effects 0.000 claims description 3
- 230000010354 integration Effects 0.000 claims description 3
- 101710163391 ADP-ribosyl cyclase/cyclic ADP-ribose hydrolase Proteins 0.000 claims 3
- 238000011160 research Methods 0.000 abstract description 5
- 238000010586 diagram Methods 0.000 description 10
- 238000012360 testing method Methods 0.000 description 10
- 230000008901 benefit Effects 0.000 description 6
- 230000008569 process Effects 0.000 description 5
- 238000005259 measurement Methods 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 238000012938 design process Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/15—Vehicle, aircraft or watercraft design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/17—Mechanical parametric or variational design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Geometry (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Aviation & Aerospace Engineering (AREA)
- Automation & Control Theory (AREA)
- Vehicle Body Suspensions (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses a design method of an active suspension system based on a cascade disturbance observer, which comprises the following steps: the method comprises the following steps: establishing an active suspension system model, and obtaining a sprung mass dynamic model of the system as follows:
step two: designing an active suspension active disturbance rejection control scheme based on a cascade disturbance observer; aiming at an 1/4 vehicle active suspension system, the finite time stability of the system is satisfied by an active disturbance rejection control scheme based on a cascade disturbance observer and a supercoiling algorithm; a cascaded disturbance observer is used for replacing a traditional linear expansion state observer, so that accurate observation precision of limited time can be obtained; the active suspension controller designed by the method can remarkably improve the vibration reduction performance of an active suspension system, and the cascade disturbance observer structure does not need excessive model information and observation bandwidth, has better robustness and convergence precision, and can provide theoretical and experimental references for active disturbance rejection control research of the active suspension.
Description
Technical Field
The invention relates to a control method of an automobile active suspension, in particular to a design method of an active suspension system based on a cascade disturbance observer.
Background
An active suspension system is an important vibration isolation element of an automobile and has gained a lot of attention, the design of a control algorithm is crucial in the active suspension control research, and the research of theory and experiment is an essential link in order to ensure the stability and comfort of the system;
in the design process of the existing active suspension system, the parameter activeness, unknown nonlinear dynamics and uncertainty of a model generally exist; the model parameters of the system need to be measured before design, more time and cost are spent, the measurement precision cannot reach the standard, part of nonlinear dynamics is difficult to measure, and the control precision, the vibration reduction effect, the comfort and the like of the designed active suspension system cannot well meet the design requirements; in addition, in the conventional active disturbance rejection control scheme, only a linear extended state observer is adopted, the limited time convergence performance cannot be obtained, and the observation bandwidth is limited by hardware equipment and cannot be increased randomly, so that the interference rejection performance is poor;
therefore, in the design process of the active suspension system, an effective robust observer design scheme is urgently needed, and the centralized uncertainty of the system can be accurately compensated and eliminated under the condition of not increasing the observation bandwidth, so that the model-free limited time control effect is achieved.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention aims to provide a design method of an active suspension system based on a cascade disturbance observer, aiming at 1/4 main active suspension system,
the finite time stability of the system is met by an active disturbance rejection control scheme based on a cascade disturbance observer and combining a supercoiling algorithm, a traditional linear expansion state observer is replaced by the cascade disturbance observer, the finite time zero error observation precision can be obtained, the effect of improving the vibration damping performance of the active suspension system is obviously achieved, excessive model information is not needed for a control structure, better robustness is achieved, and reference is provided for the research of the active suspension system.
In order to achieve the purpose, the technical scheme of the invention is realized as follows:
a design method of an active suspension system based on a cascade disturbance observer, the design method comprising the following steps:
the method comprises the following steps: model for building active suspension system
Based on 1/4 vehicle active suspension system, an active suspension system model is established, and the dynamic relation of the sprung mass of the system model is obtained as follows:
the built system model comprises a sprung mass layer, a motor drive active suspension layer, an unsprung mass layer and a tire layer;
step two: a cascade disturbance observer is added into an active suspension system model, an active suspension system capable of realizing finite time convergence and active disturbance rejection performance is designed, the cascade disturbance observer is composed of an inner ring observer and an outer ring observer, and the method comprises the following steps:
s1, selecting a state variable x of a system on the basis of an active suspension system model 1 ＝z s ，
x 3 P, dynamic model of sprung mass
Rewritten as the following state space equation:
wherein: x is the number of 1 Representing the vertical displacement of the sprung mass, x 2 Denotes x 1 First derivative of (a), x 3 Which represents the total uncertainty of the system,
representing the rate of change of the uncertainty, b representing the inverse of the nominal sprung mass, u representing the motor output force,
a derivative representing the uncertainty in the concentration;
s2, designing an inner loop observer of the cascading disturbance observer on the basis of an active suspension system model, wherein the inner loop observer actively estimates and inhibits external disturbance, unknown nonlinear dynamics and uncertainty of the active suspension system;
s3, after the inner loop observer is set, an outer loop observer of the cascading disturbance observer is further designed, and the outer loop observer can further reduce the residual observation error under the condition that the bandwidth of the inner loop observer is not increased;
s4, after the outer ring observer is arranged, a controller of the cascade disturbance observer is further designed through a supercoiling algorithm, and the supercoiling algorithm is designed at a control end to ensure the limited time stability of the active suspension system;
and S5, verifying the effectiveness of the active suspension system.
Further, the inner loop observer uses a topological state observer, and the outer loop observer uses a high-order sliding mode observer.
Further, the step one of the concrete steps of the active suspension system model building includes:
s1, arranging in an active suspension system model: mass of sprung mass layer is m s (ii) a The active control force u of the motor-driven active suspension layer is generated by a servo motor, and the spring force of the suspension is F s Damping force of suspension F d Mass of unsprung mass layer is m u (ii) a Simplifying the tyre components of the tyre layer into a parallel distributed spring and damper with the stress of F respectively t And F b (ii) a Vertical displacement of sprung mass z s Vertical displacement of unsprung mass z u The vertical excitation displacement of the ground is z r The vertical displacement of each part is measured by an encoder, the vertical acceleration of the sprung mass is measured by an accelerometer, and the vertical excitation of the ground is generated by a servo motor;
s2, the dynamic differential equation of the active suspension system model can be expressed as follows:
wherein:
is a time-varying unknown parameter that is,
a nominal portion of the sprung mass is shown,
is sprung mass acceleration;
s3, when the active suspension system is disturbed by the quality parameters, rewriting the formula (2) as follows:
the centralized uncertainty of the system is defined as:
wherein:
representing the inverse of the nominal sprung mass, p being the central uncertainty of the system, F Δ In order for the external interference to be unknown,
is the sprung mass vertical acceleration;
the sprung mass dynamics of the resulting system are:
wherein
For sprung mass acceleration, ρ is compensated with a cascaded observer and the controlled system equation (5) is guaranteed to satisfy the finite time stability, i.e. within a finite time
Further, the controlled system formula (5) has the following properties and theorem:
(1) attribute 1: the active suspension system is a bounded input and bounded state system, and the first derivative of the input is bounded;
(2) attribute 2: the central uncertainty ρ of the system is unknown, but itContinuously conductible with respect to Lipschitz, i.e.
Wherein the parameter L can be determined experimentally;
(3) theorem 1: the following second order systems exist:
if c is 1 > 0 and c 2 > 0, the trajectory x of the system 1 、x 2 、
Convergence to zero point within a finite time, the convergence time t < 2V 12 (x 0 ) Y, where x 0 Representing the initial state of the system, gamma being dependent on a parameter c 1 And c 2 V (x) is a strict Lyapunov function and satisfies
In the formula c 1 And c 2 Are two normal numbers;
(4) theorem 2: the following high-order systems exist:
wherein: x is the number of i Representing the state of the system, wherein n is the order of the system, | omega | < L, and L is a bounded normal number;
designing a high-order sliding mode observer by using an equation (7) as follows:
wherein:
is x i Is estimated by，
sign represents a sign function; if the gain k i Satisfy the requirement of
k 3 1.1L, the above-described high-order sliding-mode observer is time-limited accurate.
Further, the specific steps of designing the inner loop observer in step two S2 include:
(1) firstly, designing a centralized uncertainty rho of an inner loop observer estimation system, and defining a state observer as follows:
wherein:
is the bandwidth of the state observer,
is x i I is 1,2,3,
represents the inverse of the nominal sprung mass, u being the control input;
(2) subtracting equation (1) from equation (9) yields the following observed error dynamics:
wherein:
i＝1,2,3，e i which is indicative of an error in the observation,
a derivative representing the uncertainty in the concentration;
(3) order to
When i is 1,2,3, we can get:
wherein:
further, in the formula (11), a constant σ exists i > 0 and finite time T 1 > 0, for arbitrary bounding
If T > T 1 And
when, there is |. epsilon i (t)|≤σ i Wherein i is 1,2,3,
k is an integer greater than or equal to 3, the error will be within the finite time T 1 Internally converge to
Wherein: o (-) represents a direct scale factor,
is the bandwidth of the state observer.
Further, the designing step of the outer loop observer in step two S3 includes:
(1) order to
u n ＝u c +u s Wherein
Measured by an inner loop observer, formula (1) was substituted to obtain:
wherein: u. of s For the finite time compensation control law to be designed,
is the observation error of the inner loop observer and satisfies the continuously-derivable Lipschitz condition, i.e.
M is an
A parameter that is positively correlated;
(2) order to
Can obtain the product
(3) Applying a high order sliding mode observer as in equation (14) to the residual
And (3) estimating:
wherein: z is a radical of 1 、z 2 、z 3 Is an estimator of the observer, k 1 、k 2 、k 3 To observer gain, x 1 Is estimated as
(4) Definition of
Then the observed error dynamics can be obtained:
wherein:
is an error variable, k, of a higher order sliding mode observer 1 、k 2 、k 3 Is the observer gain;
(5) according to theorem 2, if the observed gain is chosen to be
k 3 1.1M, then
Will converge to zero within a limited time.
Further, the specific steps of designing the controller in step two S4 include:
(1) let u s ＝u t -z 3 Alternatively, formula (12) may be:
wherein: u. of t Is a super-spiral control law,
estimating an error for a high-order sliding mode observer;
(2) according to theorem 2, there is a finite time
Equation (16) will become a second order integration chain as follows:
(3) applying theorem 1 to design control law u of integral chain t The finite time stability of the system can be satisfied, and the following control law u is designed t ：
Wherein: c. C 1 And c 2 Is any positive parameter, and upsilon is an integral part of a control law;
(4) substituting the control law (18) into the system (17) yields:
(5) according to theorem 1, x 2 、υ、
Will converge to zero in a limited time, i.e. satisfy
The limited time stability of the system can be guaranteed.
Further, the design method further includes a step S5 of verifying the effectiveness of the active suspension system, which includes the following specific steps:
s1, square wave signal excitation experiments prove that the control method obtains the minimum acceleration amplitude value, so that the vibration reduction performance of the system is improved;
s2, sine signal excitation, and experiments prove that the control method obtains the minimum acceleration amplitude under the same observer bandwidth, the vibration reduction effect is superior to the traditional ADRC and LQR control, and the dynamic stroke of a suspension and the dynamic stroke of a tire are smaller than the passive control, so the overall performance is improved;
s3, analyzing results, wherein the results show that the acceleration of the control method is reduced by 69% under square wave excitation relative to ADRC control; under sine excitation, the acceleration of the control method is reduced by 82% relative to ADRC control; the comfort performance and the vibration damping performance of the suspension are improved, and the experimental result is in accordance with the theoretical analysis, so that the sprung mass acceleration of the active suspension system is reduced to zero in a limited time theoretically under the active suspension active disturbance rejection control scheme based on the cascade disturbance observer.
The invention has the beneficial effects that: the invention provides a design method of an active suspension system based on a cascade disturbance observer, and compared with the prior art, the design method has the following improvement:
(1) aiming at an 1/4 vehicle active suspension system, an active suspension Active Disturbance Rejection (ADRC) control scheme based on a Cascade Disturbance Observer (CDO) is designed, and compared with the traditional ADRC control, the CDO active suspension ADRC scheme can accurately compensate the unknown dynamics of the active suspension system without increasing the observation bandwidth, so that higher control precision and limited time stability are obtained;
(2) experiments show that under different interference excitations, the acceleration root mean square value of the control scheme designed by the invention under square wave excitation is respectively reduced by 69% and 87%, and the acceleration root mean square value under sine excitation is respectively reduced by 82% and 89%, so that the control scheme designed by the invention is obviously superior to the traditional ADRC and LQR control;
(3) the 1/4 vehicle main driven suspension system designed by the control method provided by the invention has the advantages that the vehicle body vertical acceleration is remarkably reduced (as shown in figures 6 and 10), and meanwhile, the suspension dynamic stroke and the tire dynamic stroke are also smaller than the maximum limit value of 10mm (as shown in figures 7, 8, 11 and 12), and the 1/4 vehicle main driven suspension system designed by the invention is further proved to have the advantages of good comfort, strong robustness and limited time stability.
Drawings
Fig. 1 is a diagram of an active suspension system model according to the present invention.
Fig. 2 is a control flow chart of a conventional ADRC control scheme of embodiment 1 of the present invention.
Fig. 3 is a control flow chart of the ADRC control scheme based on CDO in embodiment 1 of the present invention.
Fig. 4 is a diagram of a square wave signal for road surface excitation in the square wave signal excitation experiment in embodiment 2 of the present invention.
FIG. 5 is a diagram of the sprung mass acceleration signal from the square wave signal excitation experiment of embodiment 2 of the present invention.
Fig. 6 is a diagram of a test signal of a suspension dynamic stroke in a square wave signal excitation experiment in embodiment 2 of the invention.
Fig. 7 is a signal diagram of a tire dynamic stroke test in a square wave signal excitation experiment in embodiment 2 of the invention.
Fig. 8 is a motor control force diagram of a square wave signal excitation experiment in embodiment 2 of the present invention.
Fig. 9 is a diagram of sprung mass acceleration signals from a sinusoidal signal excitation experiment in embodiment 2 of the present invention.
Fig. 10 is a signal diagram of a suspension dynamic stroke test in a sine signal excitation experiment in embodiment 2 of the invention.
FIG. 11 is a signal diagram of a tire dynamic stroke test in a sinusoidal signal excitation experiment of example 2 of the present invention.
Fig. 12 is a motor control force diagram of a sine signal excitation experiment in embodiment 2 of the present invention.
Detailed Description
In order to make those skilled in the art better understand the technical solution of the present invention, the following further describes the technical solution of the present invention with reference to the drawings and the embodiments;
referring to fig. 1-12, a method for designing an active suspension system based on a cascaded disturbance observer, the method comprising the steps of:
the method comprises the following steps: building active suspension system model
Based on 1/4 vehicle active suspension system, active suspension system model is established to obtain the dynamic spring load mass of the system modelThe method comprises the following steps:
the built system model comprises a spring mass layer, a motor drive active suspension layer, an unsprung mass layer and a tire layer;
the specific process for building the active suspension system model comprises the following steps:
s1, setting in an active suspension system model (shown in figure 1): mass of sprung mass layer is m s (ii) a The active control force u of the motor-driven active suspension layer is generated by a servo motor, and the spring force of the suspension is F s Damping force of suspension F d The spring mass of the unsprung mass layer is m u (ii) a Simplifying the tyre components of the tyre layer into a parallel distributed spring and damper with the stress of F respectively t And F b (ii) a Vertical displacement of sprung mass z s Vertical displacement of unsprung mass z u The vertical excitation displacement of the ground is z r The vertical displacement of each part is measured by an encoder, the vertical acceleration of the sprung mass is measured by an accelerometer, and the vertical excitation of the ground is simulated by a servo motor;
s2, the dynamic differential equation of the active suspension system model can be expressed as:
in the present active suspension system model, the sprung mass generally varies with the load and the number of passengers, so it can be decomposed into
Wherein
Is a time-varying unknown parameter
Representing a nominal portion of sprung mass;
s3, when the active suspension system is disturbed by the mass parameter, the equation (2) is rewritten as follows:
the centralized uncertainty of the system is defined as:
wherein:
representing the inverse of the nominal sprung mass, p being the central uncertainty of the system, F Δ In order for the external interference to be unknown,
is the sprung mass vertical acceleration;
the sprung mass dynamics of the resulting system are:
wherein:
for sprung mass acceleration, ρ is compensated with a cascaded observer, ensuring that the controlled system equation (5) meets the finite time stability, i.e. within a finite time
The partial attributes and theorem existing in the controlled system formula (5) are as follows:
(1) attribute 1: the active suspension system is a bounded input and a bounded state, and the derivative of the input is bounded;
(2) attribute 2: the collective uncertainty ρ of the system is unknown, but it is continuously derivable with respect to Lipschitz, i.e.
The parameter L can be set through experiments;
(3) theorem 1: the controlled system formula (5) has the following second-order system:
if c is 1 > 0 and c 2 > 0, the trajectory x of the system 1 、x 2 、
Convergence to zero point within a finite time, the convergence time t < 2V 1/2 (x 0 ) Y, where x 0 Representing the initial state of the system, gamma being dependent on the parameter c 1 And c 2 V (x) is a strict Lyapunov function and satisfies
In the formula c 1 And c 2 Are two normal numbers;
(4) theorem 2: the controlled system formula (5) has the following high-order system:
wherein: x is the number of i And (3) representing the state of the system, wherein n is the order of the system, | omega | is less than or equal to L, and L is a bounded normal number.
The following high-order sliding-mode observer (HOSMO) is designed using equation (7):
wherein:
is x i The estimated amount of (a) is,
sign represents a sign function; if the gain k i Satisfy the requirement of
k 3 1.1L, the above-described high-order sliding-mode observer is time-limited accurate.
Step two: adding a cascade disturbance observer into an active suspension system model to design an active suspension system capable of realizing finite time convergence and active disturbance rejection control performance, wherein the cascade disturbance observer is composed of an inner ring observer and an outer ring observer, the inner ring observer uses a topological state observer, and the outer ring observer uses a high-order sliding mode observer, and the method comprises the following steps:
s1, selecting the state variable of the system as x 1 ＝z s ，
x 3 P, dynamic model of sprung mass
Rewritten as the following state space equation:
wherein: x is the number of 1 Representing the vertical displacement of the sprung mass, x 2 Denotes x 1 First derivative of (a), x 3 Which represents the total uncertainty of the system,
representing the rate of change of the uncertainty, b representing the derivative of the nominal sprung mass, u representing the motor output force,
the derivative of the uncertainty in the concentration is represented.
S2, designing related parameters of an inner ring observer of the cascade observer on the basis of an active suspension system model, wherein the specific process is as follows:
(1) first, an inner loop observer is designed to estimate the central uncertainty ρ of the system, defining the following Linear Extended State Observer (LESO):
wherein:
is the bandwidth of the state observer,
is x i I is 1,2,3,
represents the inverse of the nominal sprung mass, u being the control input;
(2) subtracting equation (1) from equation (9) yields the following observed error dynamics:
wherein:
i＝1,2,3，e i which is indicative of an error in the observation,
a derivative representing the uncertainty in the concentration;
(3) order to
When i is 1,2,3, we can get:
wherein:
the following theorem exists in equation (11): exists with a constant σ i > 0 and finite time T 1 > 0, for arbitrarily bounded
If T > T 1 And
when, there is |. epsilon i (t)|≤σ i Wherein i is 1,2,3,
k is an integer larger than or equal to 3, the error will be in a finite time T 1 Then converge to
O (-) represents a direct scale factor,
is the bandwidth of the state observer.
In the step, the centralized uncertain dynamics can be roughly estimated by adopting the LESO, so that the observation error is converged to a smaller boundary, and the unknown dynamics of the control system is compensated; in LESO, by increasing the observation bandwidth
The observation error can be reduced, but the observation bandwidth is limited by the noise and sampling frequency of the sensor, so the observation bandwidth is not suitable to be overlarge, and in order to further improve the observation precision and obtain better vibration reduction performance, a high-order sliding mode observer (HOSMO) is adopted to perform accurate nonlinear compensation.
S3, after designing the inner loop observer, further designing relevant parameters of the outer loop observer of the cascade observer, wherein the specific process is as follows:
(1) order to
u n ＝u c +u s Wherein
Measured by an inner loop observer, formula (1) was substituted to obtain:
wherein: u. of s For a finite time compensation control law to be designed, e 3 Is the observation error of the inner loop observer and satisfies the continuously-derivable Lipschitz condition, i.e.
M is an
A parameter that is positively correlated;
(2) order to
Can obtain the product
(3) Applying a high order sliding mode observer as in equation (14) to the residualAnd (3) estimating:
wherein: z is a radical of 1 、z 2 、z 3 Is the state of the observer, k 1 、k 2 、k 3 To observer gain, x 1 Has an estimation error of
(4) Order to
Then the observed error dynamics can be obtained:
wherein:
is an error variable, k, of a higher order sliding mode observer 1 、k 2 、k 3 Is the observer gain;
(5) according to theorem 2, if the observed gain is chosen to be
k 3 1.1M, then e i (i ═ 1,2,3) will converge to zero in a finite time;
the step of applying the observer of the outer ring can further compensate the estimation error caused by the observer of the inner ring, thereby accurately compensating the unknown nonlinear dynamics of the system, and because the observation error generated by the observer of the inner ring is smaller, the adoption of a high-order sliding mode observer (HOSMO) as the observer of the outer ring has the advantages that no overlarge observation gain is needed, and the generation of obvious sliding mode buffeting is avoided; further, the controller of the cascaded disturbance observer is designed by utilizing a supercoiled algorithm (STA) in the next step to meet the limited time stability of the system.
S4, designing parameters of the cascaded observer controller, wherein the specific process is as follows:
(1) let u be s ＝u t -z 3 Alternatively, formula (12) may be:
wherein: u. of t In order to be a superspiral control law,
estimating an error for a high-order sliding mode observer;
(2) according to theorem 2, there is a finite time
The system (16) therefore becomes a second order integration chain as follows:
(3) applying theorem 1 to design control law u of integral chain t The finite time stability of the system can be satisfied, and the following control law u is selected t ：
Wherein: c. C 1 And c 2 For any positive parameter, v is the integral part of the control law.
(4) Substituting the controller (18) into the system (17) can result in:
(5) according to theorem 1, x 2 And
will converge to zero in a limited time, i.e. satisfy
Therefore, the system isThe limited time stability can be ensured. Thus, under the CDO-based ADRC scheme, the acceleration of the active suspension system will theoretically decrease to zero in a finite time.
Through the establishment and derivation processes of the control method, a cascaded interference observer is constructed in the proposed active disturbance rejection design scheme; the cascaded disturbance observer consists of an inner loop observer and an outer loop observer; the inner ring observer uses a linear expansion state observer, and the outer ring observer uses a high-order sliding mode observer; the inner loop observer actively estimates and suppresses concentrated external disturbances, unknown nonlinear dynamics and uncertainties of the active suspension system; the outer loop observer can further reduce the remaining observation error without increasing the bandwidth of the inner loop observer; and designing a supercoiling algorithm at a control end to ensure the limited time stability of the active suspension system.
S5, comparison of technical schemes (example 1):
the control scheme of the conventional ADRC and the control scheme of the CDO-based active suspension ADRC of the present invention:
(1) the conventional ADRC technical solution is shown in fig. 2, only one Linear Extended State Observer (LESO) is used to observe the unknown dynamics of the system, and the controller adopts a conventional Proportional Derivative (PD) controller, so the control accuracy and convergence accuracy of the system are not high;
(2) the control scheme of the active suspension ADRC based on the CDO is shown in figure 3, and the scheme is based on the traditional LESO, an observer of an outer ring can be applied to further compensate estimation errors caused by an inner ring observer LESO, so that unknown nonlinear dynamics of a system can be accurately compensated, and the errors are converged to zero; secondly, different from the traditional PD controller, the supercoiled controller can ensure the finite time convergence of the system state, and further improves the robustness and the steady-state performance; meanwhile, because the observation error generated by the inner loop observer is smaller, the adoption of the HOSMO as the outer loop observer has the advantages that excessive observation gain is not needed, and obvious sliding mode buffeting is avoided;
through comparison of the technical schemes, compared with the traditional technical scheme of the ADRC based on the LESO, the technical scheme provided by the invention has the advantage that the limited time convergence can be met under the condition that the observation bandwidth is not increased, and further illustrates the feasibility of the technical scheme provided by the invention.
Step three, experimental verification and result analysis (example 2):
in order to verify and research the effectiveness of the control scheme of the invention, a hardware loop experiment is carried out by applying a two-degree-of-freedom flexible active suspension prototype, wherein the sampling frequency is 1000Hz, and the bandwidth of an observer is taken
The test time was 15s, the state of each part was measured by the encoder, and the sprung mass velocity was obtained using a low pass filter. To illustrate the uniqueness of the control scheme of the present invention, it is compared to a conventional ADRC control scheme, taking the proportional and differential gain as k P ＝401、k D 40, limited by measurement noise and sampling frequency, the observation bandwidth is taken as
Since the observation bandwidth is sensitive to measurement noise, increasing the observation bandwidth does not improve the control performance.
In the experiment, the following four control schemes were compared, respectively: 1, passive control; 2 conventional LQR control; 3 conventional ADRC control; CDO-based ADRC control (CDO-ADRC), where the control gain of LQR is K ═ 24.6648.87-0.473.68.
S1. Square wave signal excitation experiment
After test setting according to experiments, taking L as 30 and M as 10;
(1) firstly, testing by using a square wave signal with the amplitude of 0.2cm, wherein the actually generated waveform is shown in figure 4;
(2) the impact resistance of the system can be verified by using the square wave signal, and the test results are shown in fig. 5-8:
in the evaluation of the control performance of the active suspension, the magnitude and root mean square value of the acceleration of the sprung mass (vehicle body) are closely related to the comfort of the vehicle, and as can be seen from fig. 5, the CDO-ADRC control scheme provided by the invention achieves the minimum acceleration amplitude value under the condition of not needing detailed information of a model and having the same observer bandwidth, so that the vibration damping performance of the system is greatly improved; furthermore, to ensure the safety performance of the system, the suspension dynamic stroke and tire dynamic stroke waveforms are shown in fig. 6 and 7, respectively, both of which are less than the maximum limit ± 1 cm; FIG. 8 is a graph of the active control forces generated by the various control schemes, and it can be seen that the controller proposed herein has a weak buffeting that meets the actuator requirements.
S2. sine signal excitation experiment
The sine wave z is used for simulating the road unevenness for testing r (t) ═ hsin (2 pi ft), where h is 0.2cm, the excitation frequency is taken to be the natural frequency of the sprung mass, i.e. f is 3 Hz;
the experimental test results are shown in fig. 9-12, and it can be seen from the graphs that CDO-ADRC obtains the minimum acceleration amplitude under the same observer bandwidth, so the vibration reduction effect is obviously better than that of the traditional ADRC and LQR control; the suspension and tire stroke is also smaller than in passive control, so the overall performance is improved, and fig. 12 shows the control force of the motor required by each control scheme.
S3, result analysis
To quantify the control effect of each control scheme, the root mean square value (RMS) of the acceleration is calculated using the following formula:
where the excitation time is T-15 s, table 1 is the acceleration rms:
TABLE 1 suspension Performance index comparison
As can be seen from table 1:
(1) under square wave excitation, acceleration of ADRC control based on CDO decreased by 69% relative to ADRC control;
(2) acceleration for CDO-based ADRC control drops by 82% relative to ADRC control under sinusoidal excitation;
therefore, the comfort performance and the vibration damping performance of the suspension are greatly improved, and the experimental result is in accordance with the theoretical analysis, so that the effectiveness of the control scheme is verified.
Through the technical scheme and the embodiment, the effectiveness of the 1/4 vehicle main-driven suspension system based on the CDO-ADRC is proved. The following conclusions can be drawn by comparing theoretical analysis with experiments:
(1) according to the control scheme provided by the invention, the model parameters of the system do not need to be measured in advance, and compared with the traditional ADRC control, the ADRC adopting the CDO can obtain higher control precision and limited time stability under the condition of not increasing observation bandwidth;
(2) under different interference excitations, the control scheme designed by the invention is obviously superior to the traditional ADRC and LQR control, the mean square root value of the acceleration under square wave excitation is respectively reduced by 69 percent and 87 percent, and the mean square root value of the acceleration under sine excitation is respectively reduced by 82 percent and 89 percent, so the comfort performance is greatly improved, and meanwhile, the dynamic stroke of the suspension and the dynamic stroke of the tire also meet the requirements.
The foregoing shows and describes the general principles, essential features, and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are described in the specification and illustrated only to illustrate the principle of the present invention, but that various changes and modifications may be made therein without departing from the spirit and scope of the present invention, which fall within the scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (7)
1. A design method of an active suspension system based on a cascade disturbance observer is characterized by comprising the following steps:
the method comprises the following steps: model for building active suspension system
1/4-based main-driven suspension systemEstablishing an active suspension system model, and obtaining a dynamic relation of the sprung mass of the system model as follows:
the built system model comprises a sprung mass layer, a motor drive active suspension layer, an unsprung mass layer and a tire layer;
step two: a cascade interference observer is added into an active suspension system model to design an active suspension system capable of realizing finite time convergence and active disturbance rejection performance, wherein the cascade interference observer consists of an inner loop observer and an outer loop observer, and the method comprises the following steps:
s1, selecting a state variable x of a system on the basis of an active suspension system model 1 ＝z s ，
x 3 P, dynamic model of sprung mass
Rewritten as the following state space equation:
wherein: x is the number of 1 Representing the vertical displacement of the sprung mass, x 2 Denotes x 1 First derivative of (a), x 3 Which represents the total uncertainty of the system,
the rate of change of the uncertainty is represented,
representing the inverse of the nominal mass of the sprung mass,
indicating sprung massThe nominal portion, u, represents the motor output force,
a derivative representing the uncertainty in the concentration;
s2, designing an inner loop observer of the cascading disturbance observer on the basis of an active suspension system model, wherein the inner loop observer actively estimates and inhibits external disturbance, unknown nonlinear dynamics and uncertainty of the active suspension system; the specific steps of the inner loop observer design include:
(1) firstly, designing a centralized uncertainty rho of an inner loop observer estimation system, and defining a state observer as follows:
wherein: theta is the bandwidth of the state observer,
is x i I is 1,2,3,
represents the inverse of the nominal sprung mass, and u represents the motor output force;
(2) subtracting equation (1) from equation (9) yields the following observed error dynamics:
wherein:
e i which is indicative of an error in the observation,
a derivative representing the uncertainty in the concentration;
(3) let epsilon i ＝e i /θ i-1 And i is 1,2,3, then:
wherein:
s3, after the inner loop observer is set, an outer loop observer of the cascading disturbance observer is further designed, and the outer loop observer can further reduce the residual observation error under the condition that the bandwidth of the inner loop observer is not increased; the design step of the outer loop observer comprises the following steps:
(1) order to
u n ＝u c +u s Wherein
Measured by an inner loop observer, formula (1) was substituted to obtain:
wherein: u. of s For the finite time compensation control law to be designed,
is the observation error of the inner loop observer and satisfies the continuously-derivable Lipschitz condition, i.e.
M is an
A parameter that is positively correlated;
(2) order to
Can obtain the product
(3) Applying a high order sliding mode observer as in equation (14) to the residual
And (3) estimating:
wherein: z is a radical of 1 、z 2 、z 3 Is an estimator of the observer, k 1 、k 2 、k 3 To observer gain, x 1 Has an estimation error of
(4) Definition of
Then the observed error dynamics can be obtained:
wherein:
is an error variable, k, of a higher order sliding mode observer 1 、k 2 、k 3 Is the observer gain;
(5) according to theorem 2, if the observed gain is chosen to be
k 3 1.1M, then
Will converge to zero within a finite time;
s4, after the outer ring observer is arranged, a controller of the cascade disturbance observer is further designed through a supercoiling algorithm, and the supercoiling algorithm is designed at a control end to ensure the limited time stability of the active suspension system;
and S5, verifying the effectiveness of the active suspension system.
2. The design method of the active suspension system based on the cascade disturbance observer is characterized in that the inner loop observer uses a topological state observer, and the outer loop observer uses a high-order sliding mode observer according to claim 1.
3. The design method of the active suspension system based on the cascade disturbance observer as claimed in claim 2, wherein the step one of building the active suspension system model specifically comprises the steps of:
s1, arranging in an active suspension system model: mass of sprung mass layer is m s (ii) a The active control force u of the motor-driven active suspension layer is generated by a servo motor, and the spring force of the suspension is F s Damping force of suspension F d Mass of unsprung mass layer is m u (ii) a Simplifying the tyre components of the tyre layer into a parallel distributed spring and damper with the stress of F respectively t And F b (ii) a Vertical displacement of sprung mass z s Vertical displacement of unsprung mass z u The vertical excitation displacement of the ground is z r The vertical displacement of each part is measured by an encoder, the vertical acceleration of the sprung mass is measured by an accelerometer, and the vertical excitation of the ground is measured by a servomotorGenerating;
s2, the dynamic differential equation of the active suspension system model can be expressed as:
wherein:
is a time-varying unknown parameter that is,
a nominal portion of the sprung mass is shown,
representing the sprung mass vertical acceleration;
s3, when the active suspension system is disturbed by the quality parameters, rewriting the formula (2) as follows:
the centralized uncertainty of the system is defined as:
wherein:
representing the inverse of the nominal sprung mass, p being the central uncertainty of the system, F Δ In order for the external interference to be unknown,
is the sprung mass vertical acceleration;
the sprung mass dynamics of the resulting system are:
wherein
For sprung mass acceleration, ρ is compensated with a cascaded observer and the controlled system equation (5) is guaranteed to satisfy the finite time stability, i.e. within a finite time
4. The design method of the active suspension system based on the cascade disturbance observer is characterized in that the following properties and theorems exist in the controlled system formula (5):
(1) attribute 1: the active suspension system is a bounded input and bounded state system, and the first derivative of the input is bounded;
(2) attribute 2: the collective uncertainty ρ of the system is unknown, but it is continuously derivable with respect to Lipschitz, i.e.
Wherein the parameter L can be determined experimentally;
(3) theorem 1: the following second order systems exist:
if c is 1 > 0 and c 2 > 0, the trajectory x of the system 1 、x 2 、
Convergence to zero point within a finite time, the convergence time t < 2V 1/2 (x 0 ) V is a, whereinx 0 Representing the initial state of the system, gamma being dependent on the parameter c 1 And c 2 V (x) is a strict Lyapunov function and satisfies
In the formula c 1 And c 2 Are two normal numbers;
(4) theorem 2: the following high-order systems exist:
wherein: x is the number of i Representing the state of the system, wherein n is the order of the system, | omega | < L, and L is a bounded normal number;
designing a high-order sliding mode observer by using an equation (7) as follows:
wherein:
is x i The estimated amount of (a) is,
sign represents a sign function; if the gain k i Satisfy the requirement of
k 3 1.1L, the above-described high-order sliding-mode observer is time-limited accurate.
5. The design method of an active suspension system based on a cascaded disturbance observer according to claim 1, wherein in equation (11), there is a constant σ i > 0 and finite time T 1 > 0, for arbitrarily bounded
If T > T 1 And when theta > 0, has |. epsilon i (t)|≤σ i Where i is 1,2,3, σ i ＝O(1/θ k ) K is an integer greater than or equal to 3, the error will be within a finite time T 1 Inner convergence to e i ＝O(1/θ k-i+1 )；
Wherein: o (-) represents a direct scaling factor and θ is the bandwidth of the state observer.
6. The method for designing an active suspension system based on a cascaded disturbance observer according to claim 1, wherein the controller design in step two S4 comprises the following specific steps:
(1) let u s ＝u t -z 3 Alternatively, formula (12) may be:
wherein u is t In order to be a superspiral control law,
estimating an error for a high-order sliding mode observer;
(2) according to theorem 2, there is a finite time
Equation (16) will become a second order integration chain as follows:
(3) applying theorem 1 to design control law u of integral chain t The finite time stability of the system can be satisfied, and the following control law u is designed t ：
Wherein: c. C 1 And c 2 Is any positive parameter, and upsilon is an integral part of a control law;
(4) by substituting the control law (18) into the system (17):
(5) according to theorem 1, x 2 、υ、
Will converge to zero in a limited time, i.e. satisfy
The limited time stability of the system can be guaranteed.
7. The design method of an active suspension system based on a cascaded disturbance observer according to claim 1, wherein the design method further comprises a step S5 of verifying the validity of the active suspension system, which comprises the following specific procedures:
s1, square wave signal excitation experiments prove that the control method obtains the minimum acceleration amplitude value, so that the vibration reduction performance of the system is improved;
s2, sine signal excitation, and experiments prove that the control method obtains the minimum acceleration amplitude under the same observer bandwidth, the vibration reduction effect is superior to the traditional ADRC and LQR control, and the dynamic stroke of a suspension and the dynamic stroke of a tire are smaller than the passive control, so the overall performance is improved;
s3, analyzing results, wherein the results show that the acceleration of the control method is reduced by 69% under square wave excitation relative to ADRC control; under sinusoidal excitation, the acceleration of the control method is reduced by 82% compared with ADRC control; the comfort performance and the vibration damping performance of the suspension are improved, and the experimental result is in accordance with the theoretical analysis, so that the sprung mass acceleration of the active suspension system is reduced to zero in a limited time theoretically under the active suspension active disturbance rejection control scheme based on the cascade disturbance observer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010340486.9A CN111553021B (en) | 2020-04-26 | 2020-04-26 | Design method of active suspension system based on cascade disturbance observer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010340486.9A CN111553021B (en) | 2020-04-26 | 2020-04-26 | Design method of active suspension system based on cascade disturbance observer |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111553021A CN111553021A (en) | 2020-08-18 |
| CN111553021B true CN111553021B (en) | 2022-09-16 |
Family
ID=72003039
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010340486.9A Expired - Fee Related CN111553021B (en) | 2020-04-26 | 2020-04-26 | Design method of active suspension system based on cascade disturbance observer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111553021B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112953287B (en) * | 2021-03-26 | 2024-06-11 | 淮阴工学院 | Inverter self-adaptive control method based on variable perturbation expansion observer |
| CN113110053B (en) * | 2021-04-15 | 2022-08-09 | 广东工业大学 | Method for realizing finite time control on nonlinear active suspension system |
| CN114167715B (en) * | 2021-11-26 | 2023-12-15 | 济南大学 | Active suspension system saturated PD-SMC tracking method based on beneficial nonlinear factors |
| CN116430730B (en) * | 2023-04-11 | 2023-11-21 | 天津大学 | Helicopter active vibration damping control method based on limited time expansion state observer |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109334380A (en) * | 2018-11-16 | 2019-02-15 | 燕山大学 | Nonlinear hydro-pneumatic suspension Active Control Method based on parameter uncertainty and external disturbance |
| CN110597063A (en) * | 2019-09-24 | 2019-12-20 | 燕山大学 | Active suspension output feedback control method based on nonlinear extended state observer |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2749729B2 (en) * | 1991-08-29 | 1998-05-13 | 三菱電機株式会社 | Magnetic recording / reproducing circuit |
| FI123925B (en) * | 2012-08-17 | 2013-12-13 | Kone Corp | Procedure for managing elevator related data |
| CN205745069U (en) * | 2016-05-01 | 2016-11-30 | 安徽工程大学 | A kind of Novel magnetorheological fluid antivibrator |
| CN107220405B (en) * | 2017-04-21 | 2020-04-24 | 南京航空航天大学 | Automobile chassis system integration multi-objective optimization method based on improved cell membrane optimization algorithm |
| CN109101759B (en) * | 2018-09-04 | 2023-06-16 | 贵州理工学院 | Parameter identification method based on forward and reverse response surface method |
| CN110096750B (en) * | 2019-04-02 | 2020-12-11 | 燕山大学 | Self-adaptive dynamic surface control method considering nonlinear active suspension actuator |
| CN110077191B (en) * | 2019-04-15 | 2021-05-18 | 四川轻化工大学 | Control method for automobile active suspension system |
| CN110321665A (en) * | 2019-07-26 | 2019-10-11 | 广东工业大学 | Control method, device, equipment, medium and the vehicle of vehicle suspension system |
| CN110443003B (en) * | 2019-08-19 | 2023-04-18 | 合肥工业大学 | Control and optimal design method of active stabilizer bar system |
-
2020
- 2020-04-26 CN CN202010340486.9A patent/CN111553021B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109334380A (en) * | 2018-11-16 | 2019-02-15 | 燕山大学 | Nonlinear hydro-pneumatic suspension Active Control Method based on parameter uncertainty and external disturbance |
| CN110597063A (en) * | 2019-09-24 | 2019-12-20 | 燕山大学 | Active suspension output feedback control method based on nonlinear extended state observer |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111553021A (en) | 2020-08-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111553021B (en) | Design method of active suspension system based on cascade disturbance observer | |
| Koch et al. | Nonlinear and filter based estimation for vehicle suspension control | |
| CN110597064B (en) | Active suspension output feedback control method based on nonlinear and uncertain models | |
| Yi et al. | Observer design for semi-active suspension control | |
| US5750890A (en) | Method and apparatus for modelling a tire for use with a vehicle spindle-coupled simulator | |
| CA2217151C (en) | Method and apparatus for modeling a tire for use with vehicle couple-spindled simulator | |
| JP2016002844A (en) | Vehicle spring-upper/spring-lower relative speed estimation device | |
| Tafner et al. | Robust online roll dynamics identification of a vehicle using sliding mode concepts | |
| CN103197558B (en) | Microgyroscope fuzzy self-adaptation control method based on T-S model | |
| JP2019018773A (en) | Control system for suspension | |
| Hodgson et al. | A parameter identifying a Kalman filter observer for vehicle handling dynamics | |
| CN115958931A (en) | Nonlinear active suspension constraint interference suppression method with preset attenuation performance | |
| Imine et al. | Road profile estimation in heavy vehicle dynamics simulation | |
| CN116186891A (en) | Dynamic transfer characteristic calculation method and device based on double-chamber air suspension system | |
| Wang et al. | Longitudinal tire force estimation with unknown input observer | |
| Meetei et al. | Fully active suspension design using super twisting sliding mode control based on disturbance observer and ensemble Kalman filter | |
| Pham et al. | Comparative study of three robust observers for automotive damper force estimation | |
| Yi et al. | Observer-based control of vehicle semi-active suspensions | |
| CN104765929B (en) | A kind of nonlinear state decoupling observation procedure based on Multiple Time Scales recurrent neural network | |
| Gáspár et al. | Parameter identification of a suspension system and road disturbance estimation | |
| El Kattar et al. | Road profile estimation based on the second generation CRONE control | |
| SAILESH et al. | HOMOTOPY OPTIMIZATION FOR THE COMPRESSION OF A MULTIBODY MODEL OF A VEHICLE | |
| Ziegenmeyer | Estimation of disturbance inputs to a tire coupled quarter-car suspension test rig | |
| CN118683257A (en) | Uncertainty flexible active suspension system optimal control method based on UDE | |
| CN117192984A (en) | Active air suspension self-adaptive sliding mode control method based on disturbance observer |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20220916 |