Method Of Controlling A Vehicle Drivetrain And Controller In The Drivetrain

Abstract

A method of controlling a vehicle drivetrain by an electric motor, in order to synchronize the speed of an internal combustion engine and the speed of gears in the drivetrain, wherein if a speed synchronization error esync(t) is controlled to remain within a prespecified region for a specific period of time, the synchronization is finished, and a gear may be engaged.

Description

BACKGROUND OF THE INVENTION

Automatic transmissions generally include at least one gear set and a plurality of shift elements. Selective actuation of the shift elements adjusts the gear ratio of the automatic transmission and shifts the automatic transmission between its various gears.

In dedicated hybrid transmission (DHT) at least one secondary power source such as an electric motor is integrated with a combustion engine to enable electric driving, improve fuel consumption and driving performance.

Automated engagements of dry or wet clutches in cars and trucks are typically used for the actuation of energy flow management. Several control algorithms for synchronization have been proposed in literature, based on techniques such as fuzzy logic, optimal control, feedback linearization and PI control.

Certain automatic transmissions include dog clutch shifting elements. During various gear shifts, the dog clutch is engaged. Dog clutches are generally smaller, less expensive and more efficient than friction clutches (dry or wet) or synchronizers. Yet, engaging the dog shows certain challenges. For example, the dog clutch has to generally be synchronized prior to engaging the dog clutch. When the dog clutch is synchronized, components of the dog clutch rotate at a common speed. Conversely, such components may grind against each other and make a loud noise or even be damaged if the dog clutch is engaged while the dog clutch is not synchronized. However, synchronizing the dog clutch can be difficult and time consuming. A separate synchronizing mechanism, such as a synchronizer, can consume valuable space within a transmission and add to an overall cost of the automatic transmission.

In a hybrid powertrain system without friction clutches and without synchronizers, the engine speed has to be controlled precisely in order to synchronize the gears, favorably by an electric machine. One of the most important performance indices that affect the quality of a gear change is the shifting duration, which is in turn closely related to the speed synchronization performance. Controller's priority should be fast and accurate tracking control of a speed reference trajectory as dictated by the vehicle speed and the selected gear. However, many exogenous parameters such as unmodelled dynamics, unknown load disturbances and delays might significantly deteriorate the desired performance or even induce instabilities. Therefore, speed synchronization process needs to be robust with respect to unmodelled dynamics, model uncertainties, unknown load torque disturbances and system delays.

In order to improve transient response of the system, early solutions, as well as most industrial applications, have used feedforward control or combined feedforward with feedback action to improve accuracy and ensure closed loop stability. Other common speed synchronization techniques use conventional P/PI/PID control sometimes combined with pole placement to boost transients. To enhance the speed regulation control capability, many scientists used sliding-mode control (SMC) for speed synchronization. SMC has been widely adopted in the controller design since it is very robust to parameter variations and external disturbances. However, SMC tends to cause chattering and a boundary layer is needed, which deteriorates the disturbance rejection capability of the controller. In order to increase robustness, many authors have employed optimization methods like H∞-based LQR (Linear Quadratic Regulator) control with preview actions or open loop optimal control methods (Model Predictive Control). However, Ho controllers are being over conservative and sacrifice transient responses. Moreover, optimization control methods either require the solution of a linear matrix inequality (LMI), which is a rather tedious task, or are computationally demanding, which is an important performance index in embedded solutions, rendering their implementation on embedded control systems hard. Other robust methods include disturbance observers to ensure disturbance rejection.

There are different problems associated with the existing solutions:

Model based controller solutions in feedforward and feedback combination heavily rely on the exact knowledge of the systems dynamics and can deteriorate its performance when not modelled accurately. They are vulnerable in the case of erroneous system model, unknown dynamics and disturbances.

The performance of conventional P/PI/PID control might be insufficient when dealing with unknown disturbances and high speed and accuracy is demanded. Model uncertainties might lead eventually to poor performance, arising closed loop stability issues.

Sliding mode control tends to cause chattering and a boundary layer is needed, which deteriorates the disturbance rejection capability of the controller.

H∞ controllers are being over conservative and sacrifice transient responses.

Optimal control solutions (LQR/MPC) require solution of linear matrix inequality (LMI), which is a cumbersome task and increases the computational effort if done numerically.

Observer based solutions increase complexity of the solution because of complicated control algorithms.

An example of optimization control method is disclosed by Xiaoyuan Zhu, Hui Zhang, Zongde Fang “Speed synchronization control for integrated automotive motor—transmission powertrain system with random delays”, Mechanical Systems and Signal Processing, Volumes 64-65, 2015, Pages 46-57, ISSN 0888-3270.

SUMMARY OF THE INVENTION

It is an object of the present invention to provide good quality of engine speed synchronization in gear shifts with dog clutches, by addressing the following requirements:

• • The demand of fast transient performance and good accuracy despite of system uncertainties and unknown disturbances;
  • Minimizing the computational efforts and the complexity of the solutions already used to cope with system robustness.

This object is achieved by a method described and a Speed Synchronization Controller (SSC) which is also described.

According to the invention, a method of controlling a vehicle drivetrain by an electric motor in order to synchronize a speed of an internal combustion engine and a speed of gears in the drivetrain is provided. The method employs a Speed Synchronization Controller (SSC) based on Prescribed Performance Control (PPC) technique, which imposes certain pre-defined bounds on the speed synchronization tracking performance by employing a transformation of the speed synchronization error. Specifically, the speed synchronization error is forced to evolve within a predefined region, which is bounded by certain Prescribed Performance Limits. Those limits are defined using a decaying function of time called Performance Function and a Maximum Overshoot Parameter. In order to prohibit hazardous reaction of the system due to controller failure when the speed synchronization error reaches or surpasses the Prescribed Performance Limits, an additional set of safety bounds that lie within the Prescribed Performance Limits is employed, called the Safety Margins.

The method includes the following steps:

• • I. During initialization of the synchronization cycle at time zero, a synchronization Performance Function, a synchronization Maximum Overshoot Parameter and a corresponding set of Safety Margins are defined, such that the speed synchronization error is initiated within the area defined by the Safety Margins and correspondingly within the Prescribed Performance Limits. A Safety Margin violation flag, which indicates whether the speed synchronization error lies within the area defined by the Safety Margins or not, is initialized to OFF, that signifies that the speed synchronization error lies within this area.
  • II. At every step of the synchronization cycle, a transformed synchronization error is calculated, by employing a transformation of the speed synchronization error, which imposes the synchronization Prescribed Performance Limits on the speed synchronization error performance. In order to avoid a hazardous controller reaction, the Safety Margins are employed via the Safety Margin violation flag. In case the speed synchronization error reaches the Safety Margins through an abrupt speed change, the Safety Margin violation flag is set to ON and the Prescribed Performance Limits are immediately expanded with the use of a safe Performance Function and a safe Maximum Overshoot Parameter. If, during synchronization, a new target speed is requested (“change-of-mind”), then the synchronization cycle is interrupted and re-initialized with the use of a new synchronization Performance Function and a new synchronization Maximum Overshoot Parameter. At the end of every synchronization cycle step, a requested motor torque is calculated as the output of a Speed Synchronization Controller (SSC), which includes a PPC part and a switching-PI part. The calculated motor torque is set to the Starter-Generator (SG).
  • III. The synchronization cycle is finished once the speed synchronization error has settled in a prespecified small region around zero for a specific period of time.

In an exemplary embodiment, when the synchronization is finished, a gear in the drivetrain is engaged, e.g. by a command sent by a DHT manager.

In an exemplary embodiment, the synchronization Performance Function, the synchronization Maximum Overshoot Parameter and the Safety Margins are chosen such that the initial speed synchronization error lies within the Safety Margins and correspondingly within the Prescribed Performance Limits.

In an exemplary embodiment, the Safety Margins lie within the Prescribed Performance Limits at all times.

In an exemplary embodiment, the Safety Margins violation flag is set to be OFF or to remain OFF when the speed synchronization error lies within the Safety Margins, which are within and at a safety distance of the Prescribed Performance Limits, and set to be ON or remain ON when the speed synchronization error lies on the Safety Margins or surpasses them.

In an exemplary embodiment, upon an abrupt speed change, if the speed synchronization error is forced to reach the Safety Margins at a time instance, the Prescribed Performance Limits are expanded with the choice of a safe Performance Function and safe Maximum Overshoot Parameter, such that the speed synchronization error continues to evolve within the expanded Prescribed Performance Limits.

In an exemplary embodiment, upon a “change-of-mind” situation, the synchronization cycle is interrupted and re-initialized, using a new Performance Function, which is chosen such that the speed synchronization error continues to evolve within new Prescribed Performance Limits.

In an exemplary embodiment, in order to modulate the speed synchronization error with respect to the required Prescribed Performance Limits, a transformed error is determined using a synchronization Transformation Function or a safe Transformation Function.

In an embodiment, the synchronization Transformation Function and the safe Transformation Function are smooth and strictly increasing, defining bijective mappings such that when the speed synchronization error reaches the Prescribed Performance Limits, the transformed error tends to infinity.

According to another aspect of the present invention, a Speed Synchronization Controller SSC is provided for synchronizing the speed of an internal combustion engine and the speed of gears in a vehicle drivetrain by the method as described above.

In an exemplary embodiment, the Speed Synchronization Controller SSC further includes a Prescribed Performance Control PPC part calibrated to force the speed synchronization error to evolve within the Prescribed Performance Limits.

In an exemplary embodiment, the Speed Synchronization Controller SSC further includes a switching-PI part, which switches between a proportional-integral PI part and proportional P part according to whether the speed synchronization error is within a specific threshold or not.

According to another aspect of the present invention, an electronic control unit is provided for controlling a vehicle drivetrain by the method described above, wherein the electronic control unit includes at least a Speed Synchronization Controller SSC as described above and a DHT Manager configured to supervise the Speed Synchronization Controller.

According to another aspect of the present invention, a vehicle drivetrain is provided, comprising an internal combustion engine coupled to a gearbox configured to transfer torque from the internal combustion engine to an output shaft, wherein the gearbox includes dog clutches, further comprising an electric machine integrated on an input side of the dog clutches in a P0, P1 or P2 configuration and an electronic control unit as described above.

In an exemplary embodiment, the electric machine is a Starter-Generator SG coupled to a crankshaft of the internal combustion engine.

According to the invention, a method is provided to improve gear shifting quality through enhancing speed synchronization performance, and a controller to perform the method based on a recently introduced technique, called Prescribed Performance Control (PPC) is proposed. This technique imposes certain pre-defined bounds on the tracking performance by employing a transformation of the speed synchronization error. The error transformation acts like a barrier function, since it tends to infinity as the actual error approaches the pre-defined bounds of the Prescribed Performance region. The Prescribed Performance region is bounded by certain Prescribed Performance Limits, defined with the use of a decaying function of time called Performance Function and a Maximum Overshoot Parameter. With an appropriate choice of the Performance Function and the Maximum Overshoot Parameter, it is guaranteed that the speed synchronization error converges to a predefined arbitrarily small residual set, with rate no less than a prespecified value, exhibiting maximum overshoot less than a sufficiently small preassigned constant. Moreover, by appropriately parameterizing the controller, it is guaranteed that the speed synchronization error evolves at the center of the Prescribed Performance region, as it is equally repelled by the Prescribed Performance Limits. In this way, the controller performance is smoother, and the speed converges to its desired value with minimum overshoot. The proposed controller is strictly error driven and model free, which in turn makes it robust to system uncertainties and unknown load disturbances. Moreover, its simple form makes the implementation easy, avoiding any cumbersome calculations or computationally demanding executions.

The PPC method is a generic approach to control unknown nonlinear systems. Due to its general form, it attracts a vast area of applications. Some of the most distinguished ones are control of robotic systems, control of autonomous underwater vehicles, control of quadrotor, control of small-scale unmanned helicopters.

The SSC controller, which is disclosed in this invention, includes a PPC part and an extra switching-PI part with conditional integration. Moreover, it was enhanced with safety functionalities to ensure its risk-free application into the actual vehicle system. As the majority of known PPC-related works use simulations to verify the validity of their approach, a major contribution according to this invention is the commissioning and testing of the method into a real system with actual experiments in a moving vehicle.

In principle, the method is not only applicable in P0 or P1 hybrid powertrain but could also be applied in a P2. However, there is usually a friction clutch between the combustion engine and the electric machine P2, which could be used for speed synchronization. Therefore, there is no need or benefit of synchronizing the combustion engine with the electric machine.

In arrangement P3 or P4, this method is not applicable, except if it is a combination with a P0 or P1 machine.

Employing the proposed Speed Synchronization Controller SSC, which includes a combination of a PPC part and a switching-PI part, for engine speed synchronization during gear shifts provides the following:

• • Prescribed tracking performance of engine speed: engine speed tracking performance is a priori dictated by the appropriate selection of certain designer-specified Performance Functions that encapsulate the desired transient and steady-state performance specifications,
  • Robust, model-free solution: the Speed Synchronization Controller does not incorporate any knowledge of the model dynamics and the external disturbances and doesn't need an estimation or approximators to do so. It is error driven rather than model-based, which makes it versatile and easily reusable to similarly structured systems,
  • Low complexity of implementation: its structurally simple, linear-like form makes the implementation on embedded control systems straightforward. Moreover, it is computationally efficient in comparison to optimal control or observer-based solutions.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will become fully understood from the detailed description given hereinbelow and the accompanying drawings, which are given by way of illustration only, and thus, are not restrictive of the present invention, and wherein:

FIG. 1 shows components and an arrangement of an exemplary P1-P3 hybrid drivetrain;

FIG. 2 shows a control software architecture block diagram;

FIG. 3A shows a tracking error and Prescribed Performance Limits in case of e(0)≥0 and 0<M<1;

FIG. 3B shows a tracking error and Prescribed Performance Limits in case of e(0)<0 and 0<M<1;

FIG. 4A shows a further tracking error and Prescribed Performance Limits in case of e(0)≥0 and M=0;

FIG. 4B shows a further tracking error and Prescribed Performance Limits in case of e(0)<0 and M=0;

FIG. 5A shows a further tracking error and Prescribed Performance Limits in case of e(0)≥0 and M=1;

FIG. 5B shows a further tracking error and Prescribed Performance Limits in case of e(0)<0 and M=1;

FIG. 6 shows the proposed control scheme;

FIG. 7 shows a PPC protection measure;

FIG. 8 shows a further PPC protection measure;

FIG. 9 shows a synchronization cycle;

FIG. 10 shows a synchronization algorithm flowchart diagram; and

FIG. 11 shows a speed synchronization error and an integrator disabling threshold iDisableThres.

Corresponding parts are marked with the same reference symbols in all figures.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

The following description of the preferred embodiment(s) is merely exemplary in nature and is in no way intended to limit the invention, its application, or uses.

FIG. 1 shows the components and the arrangement of an exemplary P1-P3 hybrid powertrain. An internal combustion engine 3, ICE is coupled to gearbox 4, which transfers the torque over a given amount of gears to the output shaft, which propels the wheels of the vehicle. In a dedicated hybrid transmission (DHT), the gears are changed using dog clutches instead of synchronizers. This makes the gearbox smaller, more efficient and less costly. The function of the synchronizers, matching the speed of the input side, especially the 3, ICE, to the speed of the output side of the dog clutch, which has a speed proportional to the wheels, is done by actively controlling the speed of an electric machine. This electric machine has to be integrated on the input side of the dog clutches, which is the case in P0, P1 and P2 configurations. In the example, shown in FIG. 1, the High-Voltage Starter-Generator 1, SG is connected to the crankshaft of the ICE, which makes this a P1 machine. Whenever the dog clutches are in a neutral position, the ICE and the SG may rotate freely and independent of the speed of the output shaft and thus of the vehicle speed. In order to engage a gear by closing a dog clutch, the speed of the input side is synchronized to the speed of the output side by the SG. A speed control algorithm is used to do this quickly and precisely. The algorithm is implemented on an electronic control unit 2, ECU which has an interface 2.2 to the SG, e.g. a transmission control unit or a powertrain control unit. In principal, the ICE could also be used to control the speed of the input side, but a combustion engine ICE is less dynamic and has strong torque fluctuations, which makes the electric machine SG one solution for the speed synchronization.

The control software architecture diagram is seen in FIG. 2.

An auxiliary software part, called DHT Manager, shall be implemented in the ECU, to supervise the SSC functionality. A Speed Synchronization Error Debouncer detects and indicates if the speed synchronization error remains within a prespecified region, defined by a predetermined minimum threshold, for a specific time period. Reading the drivetrain signals such as vehicle speed, requested gear and actual gear and monitoring the DHT and SSC state, the DHT Manager shall decide when speed synchronization should take place and when it is safe to proceed to gear engagement.

When the speed of the input and output side are synchronized, the clutch is closed using a gear actuator GA, 5. While SG and ICE are synchronizing, e.g. during a gear change, they cannot deliver torque to propel the vehicle. Therefore, a second electric machine 6 may be used to fill the so-called torque gap during a shift. This electric machine 6 has to be connected to the output side of the dog clutches, which is called the P3 position. In FIG. 1 the second electric drive 6 is integrated in a P3 position.

Problem Statement

The equation of motion of the 3-phase induction motor used as Starter-Generator SG, when no gear is engaged, is

J _res{dot over (ω)}_SG=T_SG+T_ICE/i_SG  (1)

where J_res is the moment of inertia of the rotating parts, e.g. of the electric machines and the ICE, T_SG and T_ICE are torques provided by the SG and the ICE, i_SG is the gear ratio between SG and ICE and ω_SG is the angular velocity of the rotor of SG.

Speed Synchronization Problem:

SG angular speed ω_Sg(t) is requested to track a desired target trajectory ω_tgt(t) as dictated by the vehicle velocity v_veh and the desired gear ratio i_tgt

${\omega }_{tgt}=\frac{v_{veh}}{3.6·r_{\mathrm{wheel}}}·i_{tgt}$

The goal is to design a state feedback controller, without using any information of the system dynamics, to solve the Speed Synchronization Problem. The speed synchronization should be fast and accurate despite of system uncertainties and unknown disturbances. Moreover, it should be guaranteed that the speed synchronization error converges to a predefined arbitrarily small residual set, with rate no less than a prespecified value, exhibiting maximum overshoot less than a sufficiently small preassigned constant.

Prescribed Performance Preliminaries

This section summarizes preliminary knowledge on Prescribed Performance, originally published by C. P. Bechlioulis and G. A. Rovithakis in “Robust Adaptive Control of Feedback Linearizable MIMO Nonlinear Systems with Prescribed Performance,” in IEEE Transactions on Automatic Control, vol. 53, no. 9, pp. 2090-2099, October 2008.

Considering in FIG. 3A and FIG. 3B a generic tracking error e(t)∈R, Prescribed Performance is achieved if e(t) evolves strictly within a predefined region that is bounded by a decaying function of time. The mathematical expression of Prescribed Performance is given ∀t≥0, by the following inequality:

−Mρ(t)<e(t)<ρ(t), ∀t if e(0)≥0

−ρ(t)<e(t)<Mρ(t), ∀t if e(0)<0  (2)

where 0≤M≤1 is a Maximum Overshoot Parameter and ρ(t) is a bounded, smooth, strictly positive and decreasing function, satisfying

$\underset{t\to \infty }{\lim }\rho \left(t\right)={\rho }_{\infty }>0,$

called Performance Function. The aforementioned statements are illustrated in FIG. 3A and FIG. 3B for an exponential Performance Function ρ(t)=(ρ₀−ρ_∞)e^−lt+ρ_∞, with ρ₀, ρ_∞, l strictly positive constants and Maximum Overshoot Parameter M, where in FIG. 3A is depicted the case of e(0)≥0 and 0<M<1, whereas in FIG. 3B is depicted the case of e(0)<0 and 0<M<1.

The special cases of Maximum Overshoot Parameter M=0 are depicted in FIG. 4A in case of e(0)≥0 and FIG. 4B in case of e(0)<0, while the special cases of Maximum Overshoot Parameter M=1 are depicted in FIG. 5A in case of e(0)≥0 and FIG. 5B in case of e(0)<0.

The constant ρ₀ is selected such that (2) is satisfied at t=0, e.g.

−Mρ(0)<e(0)<ρ(0), if e(0)≥0

−ρ(0)<e(0)<Mρ(0), if e(0)<0

The constant ρ_∞ represents the maximum allowable size of e(t) at the steady state that may be set arbitrarily small to a value reflecting the resolution of the measurement device, thus achieving practical convergence of e(t) to zero. Furthermore, the decreasing rate of ρ(t), which is related to the constant 1, introduces a lower bound on the required speed of convergence of e(t).

To introduce Prescribed Performance, an error transformation is incorporated modulating the tracking error e(t) with respect to the required performance bounds imposed by ρ(t) and M. More specifically

$\begin{array}{cc}\epsilon \left(t\right)=E\left(\frac{e\left(t\right)}{\rho \left(t\right)}\right)& \left(3\right)\end{array}$

where ε is the transformed error and E(⋅) is a smooth, strictly increasing Transformation Function defining a bijective mapping:

E:(−M,1)→(−∞,∞), if e(0)≥0

E:(−1,M)→(−∞,∞), if e(0)<0  (4)

A candidate Transformation Function, could be

$\begin{array}{cc}E\left(\frac{e\left(t\right)}{\rho \left(t\right)}\right)=\{\begin{array}{c}a·\ln \left(\frac{M+\frac{e}{\rho }}{1-\frac{e}{\rho }}\right),\mathrm{if}e\left(0\right)\ge 0\\ a·\ln \left(\frac{1+\frac{e}{\rho }}{M-\frac{e}{\rho }}\right),\mathrm{if}e\left(0\right)<0\end{array}& \left(5\right)\end{array}$

where α is a positive design constant. Notice that since ρ(0) is selected such that (2) is satisfied at t=0, ε is finite owing to (4).

Design of the Speed Synchronization Controller SSC:

Let the speed synchronization error be defined as the difference between the SG angular actual speed and the SG target speed, i.e. e_sync(t)=ω_SG(t)−ω_tgt(t). The goal is to impose certain Prescribed Performance Limits, depicted with dashed line in FIG. 9, as formed by a synchronization Performance Function ρ_sync(t) and a synchronization Maximum Overshoot Parameter M_sync. To this end, an SSC controller, as shown in FIG. 6 is employed, whose output torque T_SG(t) is designed as the sum of a PPC part T_PPC(ε_sync(t) and a switching-PI part T_P-PI(e_sync(t))

T _SG(t)=T_PPC(ε_sync(t))+T_P-PI(e_sync(t)).  (6)

The PPC part T_PPC(ε_sync(t)) is constructed with the use of the transformed synchronization error

${\epsilon }_{sync}\left(t\right)=E_{sync}\left(\frac{e_{sync}\left(t\right)}{{\rho }_{sync}\left(t\right)}\right),$

where E_sync(⋅) is a synchronization Transformation Function defining a bijective mapping:

E _sync:(−M_sync,1)→(−∞,∞), if e_sync(0)≥0

E _sync:(−1,M_sync)→(−∞,∞), if e_sync(0)<0

The switching-PI part T_P-PI(e_sync(t)) is further defined as:

$\begin{array}{cc}{T}_{P-\mathrm{PI}}\left(e_{sync}\left(t\right)\right)=\{\begin{array}{c}{T}_{\mathrm{PI}}\left(e_{sync}\left(t\right)\right),❘e_{sync}\left(t\right)❘\le \mathrm{iDisable}\mathrm{Thres}\\ T_P\left(e_{sync}\left(t\right)\right),❘e_{sync}\left(t\right)❘>\mathrm{iDisable}\mathrm{Thres}\end{array}& \left(7\right)\end{array}$

where T_PI(e_sync(t)) is a PI part, T_P(e_sync(t)) is a P part and iDisableThres is the integrator disabling error threshold, which determines the condition for the switching between the PI part and P part. Specifically, the T_P-PI(e_sync(t)) is switching according to whether the speed synchronization error e_sync(t) is within the region defined by iDisableThres threshold (|e_sync(t)|≤iDisableThres), depicted as the area within the dotted lines of FIG. 11, or not.

In practice this would indicate that if the speed synchronization error e_sync(t) is outside the area defined by iDisableThres (|e_sync(t)|>iDisableThres), the controller integral part is disabled, whereas if the speed synchronization error is within this area (|e_sync(t)|≤iDisableThres), the controller integral part is active. In order for this feature to be effective, the iDisableThres should be chosen less than the initial speed synchronization error (e_sync(0)). FIG. 11 shows an example where iDisableThres is chosen as ⅓ of the initial speed synchronization error (e_sync(0)).

The purpose of using the proposed switching-PI part, also known as “conditional integration”, is to inhibit the integral term of the controller when the speed synchronization error (e_sync(t)) is big, in order to avoid large transients due to controller saturation (integrator windup). This may happen in case of large setpoint changes, large disturbances, delays or equipment malfunctions. Integrator windup may also occur when several controller terms are combined, as in our case, in driving the Starter-Generator SG, 1. In such situations, the controller might be easily saturated during transient by the nonlinear PPC term T_PPC(ε_sync(t)) in FIG. 6. The integral term is active only when the speed ω_SG(t) of SG is very close to the target speed ω_tgt(t), in order to eliminate any steady state errors.

A solution to the Speed Synchronization Problem is provided by demanding the SG torque T_SG(t), as calculated by the output of SSC controller of equation (6), shown in FIG. 6.

PPC Protection Against Numerical Errors

An important computational issue of the Prescribed Performance Control PPC lies on the existence of the transformed synchronization error

${\epsilon }_{sync}\left(t\right)=E_{sync}\left(\frac{e_{sync}\left(t\right)}{{\rho }_{sync}\left(t\right)}\right),$

calculated with the use of Transformation Function E_sync(⋅). Note that if the speed synchronization error e_sync(t) abruptly reaches or surpasses the Prescribed Performance Limits, shown as dashed line in FIG. 9, due to an unexpected disturbance, controller saturation, “change-of-mind” situations, e.t.c., the transformation E_sync(⋅) is not defined, e.g. the argument of the logarithm in (5) is either infinite or negative. This leads to a numerical error in the calculations and fatally causes unexpected system behavior due to controller failure. Such phenomena are met frequently in industrial applications and it is important to protect the controller against them.

Distinction is made between two numerically hazardous cases HCa in FIG. 7 and HCb in FIG. 8, defined as:

• • Case HCa: Abrupt speed change, which could lead the speed synchronization error e_sync(t) outside the Prescribed Performance Limits (dashed line in FIG. 7), due to disturbance, insufficient controller action, e.t.c. (FIG. 7).
  • Case HCb: “Change-of-mind” situations, where the target speed changes discontinuously during synchronization due to new gear request (FIG. 8).
  • Safety Margins: Let us define the strictly positive functions σ_min(t), σ_max(t), where 0<σ_min(t)≤σ_max(t), satisfying the following inequalities:

−M_syncρ_sync(t)<−σ_min(t)<σ_max(t)<ρ_sync(t), ∀t if e_sync(0)≥0

−ρ_sync(t)<−σ_max(t)<σ_min(t)<M_syncρ_sync(t), ∀t if e_sync(0)<0  (8)

The functions σ_min(t), σ_max(t) define the Safety Margins (dotted lines in FIG. 7, FIG. 8 and FIG. 9) and should be chosen in a reasonable way to be close to the Prescribed Performance Limits (dashed lines in FIG. 7, FIG. 8 and FIG. 9) but yet allow a safe distance from them.

The key idea behind the Safety Margins (dotted lines in FIG. 7, FIG. 8 and FIG. 9) is that whenever the speed synchronization error e_sync(t) reaches them, a Safety Margin violation flag is raised (SLimViolFlag=ON in FIG. 9), which indicates that the speed synchronization error e_sync(t) could surpass the Prescribed Performance Limits (dashed lines in FIG. 7, FIG. 8 and FIG. 9) in the next time instances. Once the violation flag SLimViolFlag=ON, the controller SSC reacts according to the respective protection measure (see Protection Measures against hazardous cases HCa, HCb). During normal operation, the speed synchronization error e_sync(t) should evolve strictly within the Safety Margins (SLimViolFlag=OFF):

−σ_min(t)<e_sync(t)<σ_max(t), ∀t if e_sync(0)≥0

−σ_max(t)<e_sync(t)<σ_min(t), ∀t if e_sync(0)<0  (9)

In order to avoid getting into the aforementioned numerically hazardous situations, the following protection strategies are defined.

Protection Measures Against Hazardous Cases HCa, HCb:

In case of HCa (FIG. 7): If an unexpected condition forces the speed synchronization error e_sync(t) to reach the Safety Margins (dotted line in FIG. 7) at the time instance t=t_viol, the Prescribed Performance Limits (dashed line in FIG. 7) are immediately expanded with the choice of a safe Performance Function ρ_sƒ(t) and safe Maximum Overshoot Parameter M_sƒ, such that the speed synchronization error e_sync(t) would continue to evolve within the safe Prescribed Performance Limits (dash-dotted line in FIG. 7). This is achieved with the use of a safe Transformation Function E_sƒ(⋅), defining a defining bijective mapping:

E _sƒ:(−M_sf,1)→(−∞,∞), if e_sync(0)≥0

E _sƒ:(−1,M_sƒ)→(−∞,∞), if e_sync(0)<0

that constitutes the transformed synchronization error

${\epsilon }_{sync}\left(t\right)=E_{sf}\left(\frac{e_{\mathrm{sync}}\left(t\right)}{{\rho }_{sf}\left(t\right)}\right).$

Note that such an unexpected situation (e.g. abrupt persistent disturbance) could prohibit an optimal synchronization performance by the available control effort (controller saturation). The characteristics of the safe Prescribed Performance Limits and their main difference with respect to the synchronization Prescribed Performance Limits is that they have to be chosen in a less constrained manner, in order to allow the speed synchronization error e_sync(t) retain higher values and lower speed of convergence, without the risk of inducing numerical errors. E.g. the convergence region of the safe Prescribed Performance Limits, defined by ρ_sƒ and M_sƒ, has to be larger than the convergence region of the synchronization Prescribed Performance Limits, defined by ρ_sync∞ and M_sync, i.e. −M_sƒ·ρ_sf∞<−M_sync·ρ_sync∞<ρ_sync<ρ_sf∞, in case of e_sync(0)≥0.

In case of HCb: If a “change-of-mind” situation takes place at the time instance t=t_step, then the synchronization cycle is interrupted and re-initialized, using a new Performance Function ρ′_sync(t), and new Maximum Overshoot Parameter M′_sync, defining the new synchronization Performance Limits (dash-dotted line in FIG. 8). The new Performance Function ρ′_sync(t) has to be chosen such that the new initial speed synchronization error e′_sync(t′₀) lies within the new Prescribed Performance Limits and the new Maximum Overshoot Parameter M′_sync has to be chosen in a way to reflect the maximum allowed overshoot according to the desired synchronization performance.

The aforementioned hazardous cases HCa, HCb may occur more than once within a single synchronization cycle. Therefore, both conditions should be checked at every single iteration of the synchronization cycle, i.e. steps G and H of the synchronization algorithm flowchart diagram (FIG. 10).

Description of the Synchronization Cycle Steps in FIG. 9:

Step I—Initialization in FIG. 9:

• • Calculate the initial speed synchronization error e_sync(0),
  • Choose the form of the synchronization Performance Function (ρ_sync(t)) and the corresponding Maximum Overshoot Parameter (M_sync),
  • Choose the form of the Safety Margin functions σ_min(t), σ_max(t),
  • Set the Safety Margin violation flag SLimViolFlag=OFF,

Step II—k-Th Step of the Synchronization Cycle, See FIG. 9:

• • Calculate the speed synchronization error e_sync(t) at the specific time instance t_k,
  • Calculate the actual value of synchronization Performance Function ρ_sync(t) at the specific time instance t_k,
  • Calculate the Safety Margins σ_min(t), σ_max(t) at the specific time instance t_k,
  • Calculate the transformed synchronization error ε_sync(t) at the specific time instance t_k:

${\epsilon }_{sync}=E_{sync}\left(\frac{e_{sync}}{{\rho }_{sync}}\right),$

• • • if the Safety Margin violation flag SLimViolFlag=OFF, using the synchronization Performance Function ρ_sync(t) and synchronization Maximum Overshoot Parameter M_sync through the synchronization Transformation Function E_sync(⋅),

${\epsilon }_{sync}=E_{sf}\left(\frac{e_{sync}}{{\rho }_{sf}}\right),$

• • • if the Safety Margin violation flag SLimViolFlag=ON (case HCa), using the safe Performance Function ρ_sƒ(t) and safe Maximum Overshoot Parameter M_sf through the safe Transformation Function E_sƒ(⋅),
    • If a new target speed is requested (case HCb), interrupt the synchronization process and re-initialize the synchronization cycle (go to Step I—Initialization in FIG. 9)
  • Calculate the requested motor torque:

If |e_sync(t)|≤iDisableThres, T_SG(t)=T_PPC(ε_sync(t))+T_PI(e_sync(t))

If |e_sync(t)|>iDisableThres, T_SG(t)=T_PPC(ε_sync(t))+T_P(e_sync(t))

• • Request the SG torque T_SG(t).

Step III (in FIG. 9)—End of the Synchronization Cycle:

If the speed synchronization error e_sync(t) remains within a prespecified region defined by the threshold MinErrThres for a specific time period Δt, which may be detected and indicated by a Synchronization Error Debouncer (shown in FIG. 2):

|e_sync(t)≤MinErrThres for Δt=SyncErrDebounceTime

then the synchronization is over, and gear may be engaged. Whereby SyncErrDebounceTime is the synchronization error debounce time period that signifies the successful completion of the synchronization cycle.

FIG. 10 shows synchronization algorithm flowchart diagram:

• • Step A—Initialization,
  • Step B—Calculate the initial speed synchronization error e_sync(0),
    • Choose the synchronization Performance Function ρ_sync(t),
    • Choose synchronization Maximum Overshoot Parameter M_sync,
    • Choose the Safety Margins σ_min(t), σ_max(t),
    • Set the Safety Margin violation flag SLimViolFlag=OFF,
  • Step C—Start the synchronization cycle,
  • Step D—Calculate the speed synchronization error e_sync(t),
  • Step E—Check if |e_sync(t)|≤MinErrThres for Δt≥SyncErrDebounceTime,
    • If Yes: to to step F,
    • If No: go to step G,
  • Step F—End of synchronization cycle,
  • Step G—Check if there is a “change-of-mind” situation (case HCb),
    • If Yes: go to Step A,
    • If No: go to step H,
  • Step H—Check if the Safety Margin violation flag SLimViolFlag==ON (case HCa),
    • If Yes: go to step J,
    • If No: go to step I,
  • Step I—Use the synchronization Performance Function ρ_sync(t) and synchronization Maximum Overshoot Parameter M_sync,
  • Step J—Use the safe Performance Function ρ_sƒ(t) and safe Maximum Overshoot Parameter M_sƒ,
  • Step K—Check if the speed synchronization error e_sync(t) is within the Safety Margins,
    • If No: go to step L,
    • If Yes: go to step M,
  • Step L—Set the Safety Margin violation flag SLimViolFlag=ON (HCa), continue with step H,
  • Step M—Calculate the transformed synchronization error E_sync (t),
  • Step N—Check if |e_sync(t)|≥iDisableThres, (see also FIG. 11)
    • If No: go to step O,
    • If Yes: go to step P,
  • Step O—Calculate the speed controller output

T _SG(t)=T_PPC(ε_sync(t))+T_PI(e_sync(t)), (see also FIG. 6 in combination with equation (7)),

• • Step P—Calculate the speed controller output

T _SG(t)=T_PPC(ε_sync(t))+T_P(e_sync(t)), (see also FIG. 6 in combination with equation (7)),

• • Step Q—Set the requested SG torque T_SG(t).

Step Q signifies the final step of a single iteration of the synchronization cycle (e.g. 1st iteration). Once the first iteration is finished at step Q, a second iteration is started by going back to step D. This iteration is repeated until step E is true and the synchronization is finished with step F.

A method of controlling a vehicle drivetrain by an electric motor (1, SG), in order to synchronize the speed of an internal combustion engine (3, ICE) and the speed of gears in the drivetrain, wherein if a speed synchronization error e_sync(t) is controlled to remain within a prespecified region for a specific period of time, the synchronization is finished, and a gear may be engaged.

The description of the invention is merely exemplary in nature and, thus, variations that do not depart from the gist of the invention are intended to be within the scope of the invention. Such variations are not to be regarded as a departure from the spirit and scope of the invention.

LIST OF REFERENCE

• 1 Starter-Generator
• 2 electronic control unit
• 3 internal combustion engine
• 4 gearbox
• 5 gear actuator
• 6 electric machine
• A to R step
• DHT Dedicated Hybrid Transmission
• ECU electronic control unit
• ED electric drive
• SG Starter-Generator
• ICE internal combustion engine
• e(t) tracking error
• ε(t) transformed error
• e_sync(t) speed synchronization error
• ε_sync(t) transformed synchronization error
• ω_tgt target angular speed
• ω_SG angular velocity of the SG rotor
• v_veh velocity of the vehicle
• r_wheel radius of a wheel
• i_tgt target gear ratio
• T_SG torque of SG
• T_ICE torque of ICE
• PPC Prescribed Performance Control
• PI Proportional-Integral
• P Proportional
• T_PPC PPC part
• T_P-PI switching-PI part
• T_PI PI part
• T_P P part
• E(⋅) Transformation Function
• E_sync(⋅) synchronization Transformation Function
• E_sƒ(⋅) safe Transformation Function
• ρ(t) Performance Function
• ρ_sync(t) synchronization Performance Function
• β_sƒ(t) safe Performance Function
• M Maximum Overshoot Parameter
• M_sync synchronization Maximum Overshoot Parameter
• M_sƒ safe Maximum Overshoot Parameter
• HCa, HCb Hazardous Case a, Hazardous Case b
• σ_min(t), σ_max(t) Safety Margin functions
• MinErrThresh minimum error threshold
• SyncErrDebounceTime synchronization error debouncer
• SLimViolFlag Safety Margin violation flag
• iDisableThres integrator disabling threshold
• SSC speed synchronization controller

Claims

1. A method of controlling a vehicle drivetrain by an electric motor in order to synchronize a speed of an internal combustion engine and a speed of gears in the drivetrain, comprising the steps of: Step I—Initialization: Determining an initial speed synchronization error (esync(0)),Determining a form of a synchronization Performance Function (ρsync(t)) and a synchronization Maximum Overshoot Parameter (Msync),Determining a form of Safety Margin functions (σmin(t)), (σmax(t)),Setting a Safety Margin violation flag (SLimViolFlag) to be OFF,Step II—k-th step of a synchronization cycle: Determining the speed synchronization error (esync(t)) at a specific time instance tk,Determining an actual value of the synchronization Performance Function (ρsync(t)) at the specific time instance tk,Determining the Safety Margins (σmin(t)), (σmax(t)) at the specific time instance tk,Determining a transformed synchronization error (εsync(t)) at the specific time instance tk:

2. The method of claim 1, further comprising the steps of a gear in the drivetrain is engaged when the synchronization is finished.

3. The method of claim 1, further comprising the steps of choosing the synchronization Performance Function (ρsync(t)), the synchronization Maximum Overshoot Parameter (Msync) and the Safety Margins (σmin(t)), (σmax(t)) such that −Msyncρsync(0)<−σmin(0)<esync(0)<σmax(t)<ρsync(0), if esync(0)≥0−ρsync(0<−σmax(0)<esync(0<σmin(0)<Msyncρsync(0), if esync(0)<0.

4. The method of claim 1, further comprising the steps of the safety margins (σmin(t)), (σmax(t)) satisfy the conditions 0<σmin(t)≤σmax(t) and −Msyncρsync(t)<−σmin(t)<σmax(t)<ρsync(t), ∀t if esync(0≥0−ρsync(t)<−σmax(t)<σmin(t)<Msyncρsync(t), ∀t if esync(0<0.

5. The method of claim 1, further comprising the steps of setting the Safety Margin violation flag (SLimViolFlag) to be OFF or remaining OFF when the speed synchronization error esync(t) lies within the Safety Margins (σmin(t)), (σmax(t)), which are within and at a safety distance of the Prescribed Performance Limits, and set to be ON or remaining ON when the speed synchronization error esync(t) lies on the Safety Margins (σmin(t)), (σmax(t)) or surpasses them.

6. The method claim 5, further comprising the steps of wherein upon an abrupt speed change, if the speed synchronization error esync(t) is forced to reach the Safety Margin (σmin(t)), (σmax(t)) at a time instance tviol, the Prescribed Performance Limits are expanded with the choice of a safe performance function ρsƒ(t) and a safe Maximum Overshoot Parameter Msƒ, such that the speed synchronization error esync(t) continues to evolve within expanded Prescribed Performance Limits.

7. The method of claim 6, wherein upon a “change-of-mind” situation at the time instance t=tstep, the synchronization cycle is interrupted and re-initialized, using a new Performance Function p′sync(t) and a new Maximum Overshoot Parameter M′sync which are chosen such that the speed synchronization error esync(t) continues to evolve within new Prescribed Performance Limits.

8. The method of claim 1, further comprising the steps of determining the transformed synchronization error εsync(t) by the equation

9. The method according to claim 8, further comprising the steps of the synchronization Transformation Function Esync(⋅) is defining a bijective mapping: Esync:(−Msync,1)→(−∞,∞), if esync(0)≥0Esync:(−1,Msync)→(−∞,∞), if esync(0)<0

10. The method of claim 1, further comprising the steps of: providing speed synchronization controller (SSC); andproviding a vehicle having the drivetrain;synchronizing the speed of the internal combustion engine and the speed of gears in the drivetrain.

11. The method of claim 10, further comprising the steps of: providing a Prescribed Performance Control PPC part (TPPC(esync(t)))calibrating the Prescribed Performance Control PPC part (TPPC(esync(t))) to retain the speed synchronization error (esync(t)) within the Prescribed Performance Limits at all time.

12. The method of claim 10, further comprising the steps of providing a switching-PI part TP-PI(esync(t)), the a switching-PI part TP-PI(esync(t)) switching between a PI part (TPI(esync(t)) and a P part (TP(esync(t)) according to whether the speed synchronization error (esync(t)) is within a specific threshold or not.

13. The method of claim 10, further comprising the steps of: providing an electronic control unit (ECU), the speed synchronization controller (SSC) being part of the electronic control unit (ECU);providing a DHT Manager being part of the electronic control unit (ECU);controlling the vehicle with the electronic control unit (ECU);supervising the speed synchronization controller (SSC) with the DHT Manager.

14. The method of claim 13, further comprising the steps of: providing an internal combustion engine (3) being part of the drivetrain;providing a gearbox having an output shaft and being part of the drivetrain, the internal combustion engine coupled to the gear box;providing dog clutches having an input side, the dog clutches being part of the gearbox; andan electric machine integrated on the input side of the dog clutches in a P0, P1 or P2 configuration;transferring torque from the internal combustion engine to the output shaft of the gearbox.

15. The vehicle drivetrain of claim 14, further comprising the steps of providing the electric machine to be a Starter-Generator (SG) coupled to a crankshaft of the internal combustion engine.