Hierarchical optimal controller for predictive power allocation

Abstract

Hierarchical optimal controller for predictive power distribution. Method and system for hybrid vehicle control. A high level controller and a low level controller are provided. The high level controller uses preview information and a model of the low level controller to calculate optimized tuning parameters for the low level controller. The low level controller uses driver inputs and current operating conditions to calculate optimized torque distribution for the hybrid engine.

Description

Background Technology

[0001] Various algorithms for controlling power distribution and / or battery use in hybrid vehicles have been described and designed. A hybrid vehicle is any vehicle comprising a first power system (first power source) and an electric power system (second power source), the first power system consuming fuel from a refillable or replaceable onboard fuel source, and the electric power system using a motor driven by electricity stored in a rechargeable storage device. The first power system can use, for example and not limited to, fuel cells, gasoline, diesel, propane, natural gas, hydrogen, or any other suitable chemical and / or technology, any of which can be an engine as used herein. Several layouts are known, including layouts where the electric motor cannot be mechanically disconnected from the second power system, and layouts using a clutch to allow the second power system to be disconnected. Power distribution is the determination of how much power to obtain from each power system at any given time.

[0002] US 9,114,806 discloses methods for obtaining predictive data, including route-related data, to calculate "system-efficient engine power" (which appears to be the maximum engine efficiency achieved by using a given engine speed) and an engine power curve (the power required to meet driving demands), and to identify the discrepancy between the two curves using a forward-looking approach. The battery is used to bridge the gap between the system-efficient engine power and the power required to meet anticipated driver demands.

[0003] U.S. Patents 9,193,351 and 10,124,678 perform route analysis and segment the upcoming route. In U.S. Patent 9,193,351, the patent teaches calculating an optimized state of charge (SOC) for each segment and managing torque distribution to achieve the optimized SOC (as a setpoint) at the end of each segment. U.S. Patent 10,124,678 describes each segment of the upcoming route and optimizes power distribution control based on the segment type. Segment usage instructions will not be planned in real-time to reflect the actual operation of the system because the entire route has already been segmented and planned. Interactions between the driver, the vehicle ahead (or other vehicles), and the system will make consistency with such pre-planned route segments inefficient. For example, driver actions inconsistent with the segmented route plan may cause the battery SOC to deviate from the segment SOC setpoint; suboptimal actions will be required to return to the "planned" state.

[0004] Other algorithms handle specific situations. For example, U.S. Patent 10,894,482 provides a specific solution for predicting upcoming downhill route segments, where the battery SOC is reduced to a lower limit during anticipated recharging during the downhill segment. U.S. Patent Publication 20050228553 identifies the distance and route to the destination based on previous usage and navigation data, and promptly shuts off non-battery power sources (internal combustion engines) to allow the use of electricity upon arrival, thus minimizing battery SOC.

[0005] What is needed is a more comprehensive solution across the predicted route, which allows for optimization to achieve the best fuel economy on the navigation preview, without being locked into a rigid solution that cannot react to actual operation. Summary of the Invention

[0006] The inventors have recognized that, among other things, the problem to be solved requires new and / or alternative configurations and methods for hybrid vehicle control. In some examples, a hierarchical control architecture is used with both a high-level controller and a low-level controller. In some examples, the high-level controller uses preview information and a model of the low-level controller to calculate optimized tuning parameters for the low-level controller and transmits these optimized tuning parameters to the low-level controller. Additionally, in some examples, the low-level controller receives the optimized tuning parameters and uses these parameters, driver input, and the current operating state to calculate an optimized torque distribution for the hybrid engine.

[0007] A first illustrative and non-limiting example takes the form of a control system for a hybrid vehicle having a first power source in the form of a battery having a state of charge (SOC) and a motor configured to provide a first torque, and a second power source configured to provide a second torque. The control system includes: a high-level controller having an associated first memory storing one or more readable instruction sets for high-level controller tasks; and a low-level controller having an associated second memory storing one or more readable instruction sets for low-level controller tasks; wherein the high-level controller tasks include: receiving preview information related to the hybrid vehicle's driving path, and one or more current states of the vehicle. The system uses a model of a low-level controller, one or more system dynamic models of the hybrid vehicle, and one or more current state variables to calculate optimal fuel consumption within a predicted range, including a portion of the driving path, and calculates one or more low-level control variables used by the low-level controller; and transmits one or more low-level control variables to the low-level controller; wherein the low-level controller tasks include: receiving one or more low-level control variables from the high-level controller and one or more current state variables of the hybrid vehicle; calculating a first torque and a second torque to provide optimal fuel consumption within minimum and maximum limits for each of the first torque and the second torque; and issuing control signals to the first and second power sources based on the first torque and the second torque.

[0008] Alternatively or additionally, the higher-level controller is configured to calculate one or more control values ​​at a first sampling rate; the lower-level controller is configured to calculate a first torque, a second torque, and an on / off signal at a second sampling rate; and the second sampling rate is a sampling rate faster than the first sampling rate. Alternatively or additionally, the second power source uses fuel, and the lower-level controller task includes instructions for determining the on / off command of the first power source.

[0009] Additionally or alternatively, the hybrid vehicle further includes a transmission configured to receive first and second torques and deliver power to one or more drive wheels of the hybrid vehicle using at least two selectable gear ratios; wherein: the low-level controller task includes instructions for calculating a gear ratio request delivered to the transmission; and the model of the low-level controller used by the high-level controller includes a model of the low-level controller calculating the gear ratio request. Additionally or alternatively, the instructions for calculating the gear ratio request include: identifying feasible gear ratios; calculating the ratio of wheel torque request to transmission torque request; finding an optimal torque allocation for each feasible gear ratio; calculating the cost of the optimal torque allocation for each feasible gear ratio; and selecting the gear ratio with the lowest cost from the feasible gear ratios. Additionally or alternatively, the instructions for calculating the gear ratio request are configured to calculate the cost of the optimal torque allocation for each feasible gear ratio as a sum of fuel power cost, electricity cost, and conversion cost.

[0010] Additionally or alternatively, the high-level controller tasks include: using preview information to estimate vehicle speed in the prediction window; combining vehicle speed and preview information to estimate the required combined torque in the prediction window; calculating the optimal fuel consumption to satisfy the estimated combined torque required in the prediction window to generate an optimized fuel equivalence factor; and transmitting the optimized fuel equivalence factor to the low-level controller.

[0011] Alternatively or additionally, the low-level controller task includes calculating the optimal torque distribution between the first and second power sources using the current state and fuel equivalence factor. Alternatively or additionally, the high-level controller task relies on a driver model to calculate one or more low-level control variables, and the low-level controller task relies on driver control actions to calculate the first and second torques. Alternatively or additionally, the high-level controller task uses input from the navigation system to determine preview information instead of driver control actions, and the low-level controller task does not use input from the navigation system.

[0012] Other examples include hybrid vehicles having a first power source in the form of a battery having a state of charge (SOC) and a motor configured to provide a first torque, a second power source configured to provide a second torque, and a control system as in any of the aforementioned control system examples.

[0013] Another illustrative and non-limiting example takes the form of a method for controlling a hybrid vehicle powertrain having a first power source and a second power source, the first power source being in the form of a battery having a state of charge (SOC) and a motor configured to provide a first torque, the second power source being configured to provide a second torque, the hybrid vehicle powertrain being controlled by each of the following: a high-level controller having an associated first memory storing one or more readable instruction sets for high-level controller tasks; a low-level controller having an associated second memory storing one or more readable instruction sets for low-level controller tasks; wherein the method includes the high-level controller: receiving preview information related to the driving path of the hybrid vehicle. The method includes: receiving one or more current state variables of the vehicle; using a model of a low-level controller, one or more system dynamics models of the hybrid vehicle, and one or more current state variables, calculating optimal fuel consumption within a predicted range including a portion of the driving path, and calculating one or more low-level control variables used by the low-level controller; and transmitting one or more low-level control variables to the low-level controller; and the method further includes the low-level controller: receiving one or more low-level control variables from a high-level controller and one or more current state variables of the hybrid vehicle; calculating a first torque and a second torque to provide optimal fuel consumption within minimum and maximum limits for each of the first torque and the second torque; and issuing control signals to the first and second power sources based on the first torque and the second torque.

[0014] Alternatively or additionally, the higher-level controller calculates one or more control values ​​at a first sampling rate; the lower-level controller calculates a first torque, a second torque, and an on / off signal at a second sampling rate; and the second sampling rate is a sampling rate faster than the first sampling rate. Alternatively or additionally, the second power source uses fuel, and the method includes the lower-level controller determining and issuing an on / off command for the first power source.

[0015] Additionally or alternatively, the hybrid vehicle includes a transmission configured to receive first and second torques and deliver power to one or more drive wheels of the hybrid vehicle using one of at least two selectable gear ratios; wherein the method further includes: a low-level controller calculating a gear ratio request to be delivered to the transmission and issuing a gear ratio request to the transmission; and a model of the low-level controller used by the high-level controller including a model of the low-level controller calculating the gear ratio request.

[0016] Additionally or alternatively, the lower-level controller calculates the gear ratio request by: identifying feasible gear ratios; calculating the ratio of wheel torque request to transmission torque request; finding the optimal torque allocation for each feasible gear ratio; calculating the cost of the optimal torque allocation for each feasible gear ratio; and selecting the gear ratio with the lowest cost from the feasible gear ratios. Additionally or alternatively, the lower-level controller calculates the cost of the optimal torque allocation for each feasible gear ratio as the sum of fuel power cost, electricity cost, and changeover cost.

[0017] Additionally or alternatively, the method further includes a high-level controller: using preview information to estimate vehicle speed in a prediction window; combining vehicle speed and preview information to estimate the required combined torque in the prediction window; calculating optimal fuel consumption to satisfy the estimated combined torque required in the prediction window to generate an optimized fuel equivalence factor; and transmitting the optimized fuel equivalence factor to a low-level controller. Additionally or alternatively, the method further includes a low-level controller using the current state and the fuel equivalence factor to calculate the optimal torque distribution between the first and second power sources. Additionally or alternatively, the method further includes a high-level controller using a driver model to calculate one or more low-level control variables, and the low-level controller using actual driver control actions to calculate the first torque and the second torque. Additionally or alternatively, the method further includes a high-level controller using input from a navigation system to determine preview information, rather than driver control actions.

[0018] Further examples include hybrid vehicles having a first power source in the form of a battery having a state of charge (SOC) and a motor configured to provide a first torque, a second power source configured to provide a second torque, and a control system having an advanced controller and a low-level controller, the control system being configured to perform a method as described in any of the foregoing method examples.

[0019] This invention description is intended to provide an overview of the subject matter of this patent application. It is not intended to provide an exclusive or exhaustive explanation. The detailed description is included to provide further information regarding this patent application. Attached Figure Description

[0020] In accompanying drawings that are not necessarily drawn to scale, similar reference numerals may describe similar components in different views. Similar reference numerals with different letter suffixes may indicate different instances of similar components. The accompanying drawings generally illustrate the various embodiments discussed in this document by way of example and not limitation.

[0021] Figure 1 It is a block diagram of the overall control architecture;

[0022] Figure 2 This is a flowchart of the gear selection method;

[0023] Figure 3 This is a flowchart of the overall approach;

[0024] Figure 4 An illustrative hybrid vehicle control architecture is shown; and

[0025] Figure 5-9 Several illustrative hybrid vehicle powertrain architectures are shown. Detailed Implementation

[0026] Figure 1 This is a block diagram of the overall control architecture. The advanced controller 100 obtains current state data 102, which may include, for example, but is not limited to, current wheel torque request, current vehicle / wheel speed, current battery SOC, and current auxiliary power demand. The advanced controller 100 also obtains preview information 110, which is used to determine, for example, but not limited to, wheel torque request, future vehicle / wheel speed, future battery SOC estimate, and / or future auxiliary power demand. Preview information can be obtained using predictive models (e.g., driver prediction model, forward vehicle prediction model, auxiliary power prediction model, etc.).

[0027] In an illustrative example, preview information 110 can be obtained using navigation system data, which can be obtained through communication with a remote server, such as via cellular or other communication. The navigation system data may include information about the upcoming road. For example, given the destination, the drawn route, or the currently traveled road, the system can determine data including speed limits, expected driver behavior, and / or traffic data to construct a speed profile that the vehicle can be expected to follow. In some examples, a driver prediction model can be used to predict driver behavior. For example, in some embodiments, it may be assumed that the vehicle will accelerate from its current speed to a legal speed limit, or another speed limit, such as a comfort-based speed limit (e.g., taking into account road curvature and vehicle sway limits), or until it reaches the vehicle ahead, and then remain at the speed limit or a safe following distance from the vehicle ahead. The preview information may also include traffic signal information, where the speed profile is further adjusted to take into account, for example, but not limited to, the effect of an upcoming stop light, which, if expected, will require the vehicle to decelerate or stop and then accelerate.

[0028] Then, by taking into account resistance, road curvature, road gradient, etc., the speed curve can be used to estimate the resultant force or total torque required for the vehicle to cross the upcoming road. The torque demand can include a series of samples sufficient to cross at least a portion of the upcoming road within a desired prediction range (such as several seconds, tens of seconds, or longer). Furthermore, the current and / or anticipated future demand for auxiliary power can be added to the torque demand. Additional preview information 110 is available, and it is not necessary to calculate the preview information based on this specific example.

[0029] The advanced controller 100 will also utilize the model of the low-level controller 104 and the system model 106, which captures system dynamics and constraints. The low-level controller model may include analyses that the low-level controller will use to determine output settings, including equations shown in more detail below. The system model 106 may, for example, incorporate a model of a motor-generator unit, such as using an efficiency plot (MGU) that correlates motor efficiency with, for example, motor speed, motor torque, and / or motor power. The system model 106 may also include a model of, for example, another power source (such as a fuel cell or internal combustion engine), including an engine plot providing engine efficiency data for engine speed and torque requirements. The system model 106 may also incorporate fuel and electricity consumption for each engine start-up event, which can be used to calculate engine start / stop penalties. The advanced controller may utilize a dynamic model of the battery, which correlates battery SOC evolution with applied power (both charging and discharging).

[0030] Using preview information and predictive models, the higher-level controller optimizes the tuning parameters and / or other control inputs of the lower-level controller to minimize the cost function of the chosen option evaluated within the predictive range. The cost function may include terms related to fuel consumption, SOC limit violations, etc., within the predictive range. While the following discussion focuses on fuel efficiency-driven approaches, other or additional factors may be included in other examples. For instance, carbon or other emissions (NOx) may be modeled and used in the calculations. For example, the generation of carbon dioxide as a combustion byproduct may be considered alone or together with other emissions in the optimization calculations used in the lower-level controller, where the model is reflected in the higher-level controller.

[0031] The higher-level controller can assume that the tuning of the lower-level controller can be adjusted at any time within the prediction range (producing N optimal tuning parameter sets, where N is the length of the prediction range) or only adjusted at specific time instances (producing fewer than N optimal tuning parameter sets).

[0032] A specific example includes a high-level controller 100 calculating the equivalent factor λ of a low-level controller that implements an equivalent energy minimization strategy (ECMS). The ECMS method optimizes the power distribution between power sources (e.g., the engine and the MGU) to minimize energy consumption. The equivalent factor λ can be understood as the conversion coefficient between fuel energy and electrical energy. Because the analysis operates across a prediction range, which can be several seconds to tens of seconds or longer, a series of optimized equivalent factors will be generated as a time vector λ* = {λ*(1), λ*(2), λ*(3)…λ*(n)}. At least a first sample of this series of optimized equivalent factors is passed from the high-level controller to the low-level controller, which uses the received optimized equivalent factors as tuning parameters for performing control operations. This analysis occurs in an environment that changes rapidly as the vehicle moves, and therefore only one or a few samples of the calculated series of optimized equivalent factors may be used before restarting the calculation.

[0033] In the illustrative example, the advanced controller 100 is configured or programmed (by storing machine-executable instructions / code on a non-transitory medium) to minimize the cost function:

[0034] λ * =arg minJ HL (λ) {Equation 1}

[0035] Where λ is the equivalent factor vector optimized across the entire prediction range, and λ* is the optimal equivalent factor vector. Cost function J HL This is the high-level controller cost function, which is further described here:

[0036]

[0037] Where φ is the terminal term of the high-level cost function and represents the penalty for violating the state-of-charge (SOC) limit of the battery at the end of the prediction range. For example, a penalty can be applied if the target SOC is not met, or if the SOC exceeds the high or low boundary, where, for example, different penalty weights can be applied based on one or more of which violations occur, how long the violations last, and how far the violations exceed the boundary. N is the length of the prediction range in the sample, and i is a discrete-time instance. As described, a low-level controller is modeled, and therefore... The term represents the optimal low-level control action (which may include, but is not limited to, engine torque requests, engine start / stop requests, and / or transmission gear requests) given by ECMS predictions modeled by the high-level controller. Similarly, in Equation 2, for the i-th element of the time series, v i τ whl，i These are the vehicle speed and wheel torque requests, respectively. iThis is the integral term of the high-level controller cost function, which includes fuel consumed within the forecast range and violations of the SOC limit during the forecast range, and is further described in Equation 3:

[0038]

[0039] In equation 3, This is the fuel consumption rate, taking into account the optimal low-level control action (such as the modeled one) relative to the previous control action at the vehicle speed. The following terms... and K SOC It is the weight of violations of the upper or lower limit of SOC within the prediction range.

[0040] Next:

[0041]

[0042] As mentioned above, φ can be the penalty for violating the SOC limit at the end of the prediction range (the Nth sample), where Equation 4 illustrates how the weights and maximum / minimum calculations are performed. The high-level controller models a series of control actions from the low-level controller, generating the three control signals shown in Equation 5:

[0043]

[0044] in The term represents the optimal low-level control action modeled by the high-level controller, and and G * These are the optimal engine torque, engine on / off, and gear ratio requests calculated by the low-level controller, modeled by the high-level controller. As described, the low-level controller can use an ECMS control scheme, and therefore the high-level controller can model the low-level controller for ECMS control. If the low-level controller uses a different control scheme, the high-level controller can adjust accordingly. Some examples can further extend the ECMS control scheme to account for additional variables such as battery wear, other component health / aging, emissions, etc.

[0045] Equation 6 describes the optimal MGU torque for modeling:

[0046]

[0047] Where τ* mau,iThe optimized, modeled MGU torque request for the i-th sample is a function of the optimal low-level controller action. As shown, the MGU torque request is obtained by dividing the wheel torque request by the gear ratio request and subtracting the torque supplied by the second power source (the engine or internal combustion engine (ICE) in these equations), which is the product of the second power source torque request and the on / off control signal of the second power source, both of which are modeled requests as a function of a vector of equivalent factors.

[0048] The advanced controller will also determine the SOC of the MGU battery:

[0049]

[0050] The SOC of i+1 is a function of the previous SOC sample, the optimal MGU, the torque of the second power source, the optimal second power source on / off request (shown as engine on / off request in the equation), the optimal gear request, and the vehicle speed.

[0051] In some examples, engine on / off can be handled by the low-level controller (note the discussion below regarding the limitations of functionality for specific hybrid powertrain architectures). In non-restrictive examples, the ECMS algorithm can be extended to handle engine on / off by using instantaneous optimization in engine torque optimization. For example, ECMS uses the current time sample to provide fuel-optimal engine torque and engine on / off signals for a specific time sample. In one example, the low-level controller may determine that engine torque is not needed and that the engine can be turned off, and thus the low-level controller makes this determination and signals the engine to turn off; when torque is needed from the engine again, the low-level controller makes that determination and signals the engine to turn on or start. The high-level controller models the low-level controller by obtaining the engine torque and engine on / off for each sample in the preview window, and then optimizes the low-level controller parameters so that the control is optimal for the current time sample and also relative to the prediction range. For example, the high-level controller can enhance the low-level controller model to use engine start or engine stop penalties when modeling the low-level controller actions, so that the low-level controller is provided with an energy-optimal set of parameters, rather than having to keep track of both the optimization problem and the set of rules for engine on / off. Other options can be used within the graded control strategy for engine on / off and / or transmission gear settings.

[0052] When a high-level controller models low-level control signals, cost function minimization can be achieved as follows:

[0053] Expressed as shown in Equation 8:

[0054]

[0055] The cost function can be expressed as:

[0056]

[0057] Where P {f} P represents the power consumed by the fuel path, which is generated by the use of the secondary fuel source as specified by the current and previous control signals and vehicle speed vi. {e} This represents the power consumed by the electrical path (including battery power consumption and on / off events), and the last factor R Δu It is the penalty for the change in torque request Δτ from the second power source (usually an internal combustion engine, although other power sources can be used). ice，i The weight.

[0058] The MGU torque can be determined from the second power source torque, wheel torque, and gear request as follows:

[0059]

[0060] The power consumed by the fuel path is characterized by Equation 11:

[0061]

[0062] in This is the fuel consumption rate, and LHV is the lower calorific value of the fuel, so the power consumed is the product of fuel usage and LHV. The power consumed via the electrical path is characterized by Equation 12:

[0063] P {e}，i =η(u LL，i (λ), τ whl，i v i )·P {MGU} (τ mgu，i u LL，i (λ), u LL，i-1 (λ), v i Equation 12

[0064] Where η is the MGU efficiency, which can vary with received control signals, wheel torque requests, and vehicle speed indications. MGU power consumption P {MGU} This is related to the torque request received by the MGU, vehicle speed, and current and previous control signals.

[0065] Higher-level controllers can impose the following constraints on the modeling and optimization of the behavior of lower-level controllers:

[0066]

[0067]

[0068] eng on,i ∈{0,1}

[0069]

[0070] here, τ ice，i , It is the lower / upper limit of the torque of the second power source (such as an engine or ICE) within the prediction range. τ mgu，i , It represents the lower / upper limit of the MGU torque within the predicted range, and It is the set of transmission ratios that are allowed at time i, calculated relative to the transmission input speed / torque limit.

[0071] As illustrated in the detailed example above, the high-level controller calculates the optimized time vector of the equivalent factor λ by selecting the optimal equivalent factor at each time point. This optimal equivalent factor minimizes fuel consumption over the entire window, is adjusted to account for any violations of SOC general limits and / or terminal limits (which may include end-of-period SOC targets), and is constrained by overall system requirements. Furthermore, in this example, the equivalent factor is then passed to the low-level controller and used as a tuning parameter for the ECMS implemented in the low-level controller.

[0072] The advanced controller can perform its model-based optimization calculations using preview data and the current system state at a relatively lower repetition rate than the advanced controller. For example, in various examples, the advanced controller may perform the analysis at intervals of approximately 0.1 seconds to approximately 10 seconds, or approximately 0.5 seconds to approximately 3 seconds, or approximately 1–2 seconds. The advanced controller will respond to the demands imposed upon it by the road (slope, curvature, etc.), the driver (whether human or autonomous), losses (friction, drag), and any other inputs (e.g., use of auxiliary electrical systems), using equivalent factors received from the advanced controller, and performing the actual optimization using an analysis similar to Equations 8–12 above. For example, the advanced controller may perform its analysis at higher frequencies in the range of approximately 10 kHz to approximately 100 Hz. In one example, the advanced controller operates at 1 Hz (1-second intervals), and the advanced controller operates at 100 Hz (0.01-second intervals). That is, the advanced controller may operate at a sampling rate at least twice, at least ten times, or at least one hundred times that of the advanced controller. The difference in the repetition rates of the two controllers is optional.

[0073] The high-level controller 100 transmits the first element of the optimal tuning parameters (at least corresponding to the first time instance within the prediction range) to the low-level controller 120. For example, if the low-level controller implements ECMS and the high-level controller optimizes N samples of λ within the prediction range, then λ*(1) corresponding to the first element of the optimal vector λ* is transmitted to the low-level controller. Similarly, if the high-level optimization is performed less frequently than the sampling time of the high-level optimization and the provided sample coverage is longer than the time interval between iterations of the high-level control scheme, several first elements can be transmitted directly. For example, if the high-level controller performs its analysis every second and uses 0.1 seconds as the time interval for its analysis, then ten samples of the optimized vector will be provided. In other examples, the high-level controller analysis uses a sampling time interval equal to the repetition rate of the high-level controller, so only one sample is provided to the low-level controller.

[0074] Constraints can be considered on the tuning parameters of the optimized low-level controller (e.g., the maximum change in parameters to prevent sudden changes in the behavior of the low-level controller). The model used by the high-level controller may be the same as or different from the model used by the low-level controller.

[0075] In one example, the lower-level controller uses a formula similar to Equations 8-12 above, as shown in the following section. If updated to reflect the lower-level controller receiving λ*(1) from the higher-level controller, the cost function minimization can be expressed as shown in Equation 13:

[0076]

[0077] The cost function can be expressed as:

[0078]

[0079] Where P {f} P represents the power consumed via the fuel path generated by the use of a second source of fuel. {e} This represents the power consumed by the electrical path (including battery power consumption and on / off events), and the last factor R Δu It is the penalty for the change in torque request Δτ from the second power source (usually an internal combustion engine, although other power sources can be used). ice The weight.

[0080] The MGU torque can be determined from the second power source torque, wheel torque, and gear request as follows:

[0081]

[0082] The power consumed by the fuel path is characterized by Equation 16:

[0083]

[0084] in This is the fuel consumption rate, and LHV is the lower calorific value of the fuel, so the power consumed is the product of fuel usage and LHV. The power consumed via the electrical path is characterized by Equation 17:

[0085] P {e}，j =η(u LL，j (λ * (1)), τ whl，j v j )·P {MGU} (τ mgu，j u LL，j (λ * (1)), u LL，j-1 v j Equation 17

[0086] Where η is the MGU efficiency obtained from the MGU diagram, which can vary with the received control signals, wheel torque requests, and vehicle speed indications, and the MGU power consumption is related to the torque requests received by the MGU, the vehicle speed, and the current and previous control signals.

[0087] Constraints for low-level controller optimization include:

[0088]

[0089]

[0090] eng on,j ∈{0,1}

[0091]

[0092] here, τ ice，j , It is the lower / upper limit of the torque of the second power source (such as an engine or ICE) within the prediction range. τ mgu，j , It represents the lower / upper limit of the MGU torque within the predicted range, and It is the set of transmission ratios that are permissible at time j, calculated relative to the transmission input speed / torque limit.

[0093] Return to Figure 1The advanced controller uses preview information 110 and current state 102, along with the low-level controller model 104 and system model 106, to calculate the optimized time vector of the equivalent factor λ*. At least a first sample λ*(1) of the optimized time vector of the equivalent factor is transmitted to the low-level controller 120. The low-level controller 120 uses data from current state 102, tuning parameters λ*(1), and system model 106 to calculate the optimized solution to calculate the control signals issued to each of the first power source 130, the second power source 140, and the transmission 150. For example, the first power source 130 may be an engine (fuel cell, gasoline, diesel, propane, natural gas, hydrogen, or any other suitable chemical and / or technology, any of which may be an engine as used herein), and receives a first power source torque request (also referred to herein as an engine torque request or ICE torque request, where ICE indicates an internal combustion engine) and a first power source on / off request from the low-level controller 120. The second power source 140 can be an MGU / motor that is coupled to a rechargeable battery and receives torque requests from the MGU or motor.

[0094] In some examples, transmission 150 receives gear requests. In some examples, transmission 150 may be a continuously variable transmission, in which case the above formula will be modified to continuously optimize transmission gear requests, rather than using discrete gear ratios.

[0095] In some examples, the low-level controller will use the first sample λ*(1) of the optimized time vector of the received equivalent factor for several iterations of the low-level control analysis. For example, the analysis in the low-level controller can occur at 10 to 1000 Hz, with the same λ*(1) for the entire second (10 to 1000 iterations), while the received data about the current state 102 will be updated with each iteration as new data samples are received. In one example, the low-level controller operates at at least ten times the frequency of the high-level controller, or alternatively, approximately ten times the frequency of the high-level controller. Continuing with this example, the high-level controller can operate to recalculate the optimized time vector at one-second intervals, such that only one sample is used from each calculation. Other methods and configurations can be used alternatively.

[0096] In some examples, the high-level controller can perform its analysis at set intervals and can generate an optimization time vector of the equivalent factor, where each element of the vector is used at different intervals. For example, the high-level controller can compute the optimization time vector of the equivalent factor λ* at one-second intervals, where samples within the vector are configured to be used at 100-millisecond intervals, such that ten samples of λ* will be transmitted, each sample used for at least one iteration of the low-level controller analysis. Continuing this example, if the low-level controller operates at 100Hz, each sample of the optimization time vector of the equivalent value will be used for 10 iterations of the low-level controller analysis.

[0097] Figure 2 This is a flowchart of the method for gear selection. At 200, the method determines the feasible gear ratio. Such analysis may include identifying the current gear 212 and determining whether a shift is available, as shown at 214. For example, after a shift, the system may need to wait at least a minimum time (e.g., 500 milliseconds to 5 seconds) before a subsequent shift to avoid overly frequent gear changes, as mechanical actions require a certain minimum amount of time, and / or overly frequent shifts may be impaired driver comfort and / or cause wear on the clutch / transmission, making such frequent shifts unlikely. The feasible gear ratio 200 may be, for example, and not limited to, the current gear, the current gear plus one (unless in the highest gear), and the current gear minus one (unless in the lowest gear); if dual shifting is permitted, it may also include plus / minus two. Depending on the desired outcome, for transmissions with multiple available ranges and gear selections, the feasible gear ratios may also consider range variations.

[0098] Next, as shown at 210, the method calculates the ratio of wheel torque request to transmission input torque request for each feasible gear ratio. As shown at 220, the optimal torque distribution solution (TSS) for each feasible gear ratio is then calculated, for example using an ECMS strategy. At 230, the solution cost for each feasible gear ratio is calculated using, for example, the following formula:

[0099] H j =P {f}，j +λ * (1)P {e}，j +R Δu Δ τice，j {Equation 18}

[0100] Each power and change cost shown in Equation 18 is defined above. Then, as indicated at 240, the minimum cost is selected. With the optimal gear ratio request determined, torque distribution and engine on / off requests can then be calculated.

[0101] Figure 3This is a block diagram of the overall approach to controlling a hybrid vehicle. Starting at block 400, the High-Level Controller (HLC) receives preview and current state data. Such data can include information from various inputs shown at 470, including preview information from the Navigation System (NAV), such as that modified or influenced by the driver model, and current state information from the Transmission Relay (TRN), the hybrid battery, and the second power source (second PS). For example, the preview information can include navigation system data related to the upcoming road (whether determined using a driving map or simply the road the vehicle is currently on), which can include road gradient, road quality, road curvature, posted speed limits, current speed, weather, traffic and / or traffic control signals (stop lights, stop or yield signs, upcoming determinations such as roundabouts, etc.). The preview information from the navigation system can further include information about the vehicle's surroundings, such as the presence of vehicles ahead and how far and how fast they are. The preview information can be combined with the navigation system information and the driver model to calculate the vehicle's speed profile. For example, a driver model can determine the extent of acceleration a driver can use to reach maximum speed, what the maximum speed might be (relative to traffic and / or speed limits), and the following distance to the vehicle ahead. The driver model can be based on assumed characteristics of a human driver ("model"), observed behavior of an actual driver in the vehicle, or, for autonomous vehicles, modeled behavior of the autonomous vehicle. Current state information may include, for example, and not limited to, the current gear selection of the transmission, the current state (on / off and / or speed) of the hybrid battery SOC and the second power source (and the first power source, here the hybrid MGU, such as its speed). Other current state information may include the vehicle's current speed and the state of any auxiliary power systems (e.g., passenger compartment or cargo load heating / cooling).

[0102] Using the method illustrated above, the HLC then calculates one or more LLC control variables 410. As previously indicated, box 410 may include the HLC using the LLC control model. In some examples, at 410, the HLC determines a time vector λ* with multiple samples, and at 420, it passes one or more samples of the calculated LLC control variables to the LLC.

[0103] At box 430, the LLC receives LLC control variables and also obtains current state information. The current state information used by the LLC at box 430 may differ from the information obtained by the HLC at box 400. For example, the LLC may obtain driver control signals along with vehicle speed, transmission ratio, hybrid battery data, and the state of the second power source; however, as described, it may not obtain one or more of the driver model and optionally navigation system inputs. In other examples, the LLC may use the driver model and / or may use navigation system data.

[0104] LLC then calculates the torque used by the MGU and the second PS, which can be an engine (gasoline, diesel, or other fuel), as indicated at 440. The torque can be calculated with reference to one or more cost functions 442, and a range of available techniques can be used to minimize the torque, such as iterative solvers (which may also occur in block 410). LLC then calculates and issues control signals, as indicated at 450, including, for example, speed and torque to the first power source or MGU 452, speed, torque, and / or on / off signals to the second power source 454, signaling for the clutch 456 (which can decouple the second power source 454 when the clutch 456 is closed and recouple it when the clutch 456 is open), and gear selection or transmission ratio (for continuously variable transmissions) for the transmission 458.

[0105] In some examples, the HLC does not receive information about the driver's current control commands (e.g., steering control, braking, and / or accelerator pedal), but instead relies on driver model and navigation system inputs to determine the optimal control variables used by the LLC; in such examples, the LLC may depend on the driver's current control commands and the vehicle's current state (e.g., vehicle speed), but not on data inputs from the navigation system and / or driver model. In other examples, the LLC may simply ignore the driver model, but may include data from the navigation system in its calculations, such as customizing control actions by taking into account upcoming road or traffic controls, if desired. Such customization may include, for example, identifying upcoming stop signs or traffic lights that will require stopping / decelerating, and configuring torque distribution to charge the hybrid battery for acceleration after an upcoming stop. Such specific tasks may be separate from other optimization activities, or alternatively, may be integrated into the driver predictive model.

[0106] Figure 4 The overall control architecture is illustrated. The overall control architecture includes a vehicle controller 300, which receives data / inputs from each of the following: driver 310, various system models or other stored information 312, preview information 314 such as from navigation, communication, and vehicle sensors, and current state data 316 (such as from engine and battery sensors) related to the current state of the system; and generates control outputs for the engine 320, MGU 322, clutch 324, and transmission 326. The controller 300 can also output data to a user interface 330 that allows the driver to receive information about ongoing operations, to a diagnostic module 332 where operating data, anomalies, faults, etc., can be stored, and / or to a communication module 334 for transmission outside the vehicle, such as to a fleet manager.

[0107] Figure 5-9 Several illustrative hybrid vehicle configurations are shown. Figure 5 The P0 hybrid system includes a battery that powers the MGU, which is belt-coupled to the transmission drive shaft, which is also coupled to the engine. The engine and MGU cannot be decoupled here. The engine starter is also belt-coupled to the transmission drive shaft. A clutch separates the transmission from the drive shaft. The transmission outputs torque to the differential, which is coupled to the drive wheels as shown.

[0108] Figure 6 In the P1 hybrid system, the MGU is moved to the opposite side of the engine and placed directly on the drive shaft. The engine's starter is coupled to the transmission drive shaft via a belt. The engine and MGU cannot be decoupled here. A clutch separates the transmission from the drive shaft. The transmission outputs torque to the differential, which is coupled to the drive wheels as shown. The controller scheme and correlation analysis described above can be applied to either the P0 or P1 hybrid powertrain architecture with minor modifications. One adjustment is that the engine cannot be turned off while the vehicle is moving because the MGU and ICE are coupled together. As a result, engine on / off can only be used when the vehicle is stationary.

[0109] Figure 7 In the P2 hybrid configuration, the MGU is moved to the other side of the clutch. Here, the engine and starter are coupled on one side of the clutch and can be decoupled from the MGU and transmission. This makes the decision to shut off the engine more useful. However, the MGU cannot be decoupled. The transmission outputs torque to the differential, which is coupled to the drive wheels as shown.

[0110] Figure 8 The P3 hybrid now has the same engine, starter, and clutch as the P2, but the MGU has been moved to a position between the transmission and the differential. Figure 9 The P4 hybrid system places the MGU on the opposite side of the differential. An alternative to the P4 hybrid system could include multiple MGUs on each drive wheel. The controller described above is easily applied to each of the P2, P3, and P4 hybrid powertrain configurations.

[0111] Each of these non-limiting examples may exist independently or may be combined with one or more other examples in various permutations or combinations. This description includes reference to the accompanying drawings, which form part of the detailed description. The drawings illustrate specific embodiments by way of illustration. These embodiments are also referred to herein as “examples.” Such examples may include elements other than those shown or described. However, the inventors also contemplate examples in which only those elements shown or described are provided. Furthermore, the inventors contemplate examples using any combination or permutation of those elements (or one or more aspects thereof) shown or described herein, relative to a particular example (or one or more aspects thereof) or relative to other examples (or one or more aspects thereof) shown or described herein.

[0112] In the event of any inconsistency between the usage in this document and any other document incorporated by reference, the usage in this document shall prevail. In this document, as is common in patent documents, the terms “a” or “an” are used to include one or more, independent of any other instances or usages of “at least one” or “one or more.” Furthermore, in the claims, the terms “first,” “second,” and “third,” etc., are used merely as labels and are not intended to impose numerical requirements on their objects.

[0113] The method examples described herein may be implemented, at least in part, by a machine or computer. Some examples may include computer-readable or machine-readable media encoded with instructions operable to configure an electronic device to perform the methods described in the examples above. Implementations of such methods may include code, such as microcode, assembly language code, high-level language code, etc. Such code may include computer-readable instructions for performing various methods. This code may form part of a computer program product. Additionally, in one example, the code may be tangibly stored on one or more volatile, non-transitory, or non-volatile tangible computer-readable media, such as during execution or at other times. Examples of such tangible computer-readable media may include, but are not limited to, hard disks, removable disks or optical discs, magnetic tape cassettes, memory cards or sticks, random access memory (RAM), read-only memory (ROM), and the like.

[0114] The above description is intended to be illustrative and not restrictive. For example, the above examples (or one or more aspects thereof) may be used in combination with each other. Other embodiments may be used by those skilled in the art upon review of the above description. The abstract is provided to comply with the requirements of 37C.FR §1.72(b) to allow the reader to quickly determine the nature of the technical disclosure. It is submitted on the understanding that it will not be used to interpret or limit the scope or meaning of the claims.

[0115] Furthermore, in the detailed description above, various features may be grouped together to simplify this disclosure. This should not be construed as implying that any disclosed feature not claimed is necessary for any claim. Rather, the subject matter of innovation may lie in fewer than all features of a particular disclosed embodiment. Therefore, the following claims are thus incorporated into the detailed description as examples or embodiments, wherein each claim is an independent, separate embodiment, and such embodiments are contemplated to be combined with each other in various combinations or arrangements. The scope of protection should be determined with reference to the appended claims together with the full scope of equivalents conferred by such claims.

Claims

1. A control system for a hybrid vehicle, the hybrid vehicle having a first power source and a second power source, the first power source being a battery having a state of charge and a motor configured to provide a first torque, the second power source being configured to provide a second torque, the hybrid vehicle further including a transmission configured to receive the first torque and the second torque and to deliver power to one or more drive wheels of the hybrid vehicle using one of at least two selectable gear ratios, the control system comprising: An advanced controller having an associated first memory that stores one or more readable instruction sets for advanced controller tasks; A low-level controller having an associated second memory, the second memory storing one or more readable instruction sets for low-level controller tasks; The advanced controller tasks include: Receive preview information related to the driving path of the hybrid vehicle, as well as one or more current state variables of the hybrid vehicle; Use the preview information to estimate the vehicle speed in the prediction window; The required combined torque in the prediction window is estimated by combining the vehicle speed and the preview information; Using a low-level controller to calculate the transmission ratio request model, one or more system dynamics models of the hybrid vehicle, and the one or more current state variables, calculate fuel consumption within a predicted range, including a portion of the driving path requiring the combined torque to satisfy the estimated prediction window, and calculate one or more low-level control variables for use by the low-level controller, including equivalent factors, where the equivalent factors are conversion coefficients between fuel energy and electrical energy; and The one or more low-level control variables, including the equivalent factor, are transmitted to the low-level controller. The low-level controller tasks include: Receive one or more low-level control variables, including the equivalent factor, and one or more current state variables of the hybrid vehicle from the high-level controller; The first torque and the second torque are calculated using the one or more current state variables and the equivalent factor to provide fuel consumption within minimum and maximum limits for each of the first torque and the second torque; Calculate the gear ratio request delivered to the transmission; and Based on the first torque, the second torque, and the gear ratio, control signals are requested to be sent to the first power source and the second power source to control the power distribution between the first power source and the second power source based on the calculated fuel consumption, thereby minimizing the fuel consumption of the hybrid vehicle.

2. The control system according to claim 1, wherein: The advanced controller is configured to calculate one or more control values ​​at a first sampling rate; The low-level controller is configured to calculate the first torque, the second torque, and the on / off signals of the first power source and / or the second power source at a second sampling rate; and The second sampling rate is a faster sampling rate than the first sampling rate.

3. The control system according to claim 1, wherein, The second power source uses fuel, and the low-level controller task includes instructions for determining the on / off command of the second power source.

4. The control system according to claim 1, wherein, The instructions used to calculate the gear ratio request include: Identify feasible transmission ratios; Calculate the ratio of wheel torque request to transmission torque request; Find the optimal torque distribution for each feasible gear ratio; Calculate the cost of optimal torque distribution for each feasible gear ratio; and Choose the transmission ratio with the lowest cost from the feasible transmission ratios.

5. The control system according to claim 4, wherein, The instructions used to calculate the gear ratio request are configured to calculate the cost of optimal torque allocation for each feasible gear ratio as a sum of fuel power cost, electricity cost, and change cost.

6. The control system according to claim 1, wherein, The low-level controller task includes calculating the optimal torque distribution between the first power source and the second power source using one or more current state variables and the equivalent factor.

7. The control system according to claim 1, wherein, The high-level controller task relies on the driver model to calculate one or more low-level control variables, and the low-level controller task relies on the driver's control actions to calculate the first torque and the second torque.

8. The control system according to claim 1, wherein, The higher-level controller task uses input from the navigation system to determine preview information, rather than driver control actions, while the lower-level controller task does not use input from the navigation system.

9. A method for controlling a hybrid vehicle powertrain, the hybrid vehicle powertrain having a first power source and a second power source, the first power source being a battery having a state of charge and a motor configured to provide a first torque, the second power source being configured to provide a second torque, the hybrid vehicle powertrain further including a transmission configured to receive the first torque and the second torque and to deliver power to one or more drive wheels of the hybrid vehicle using one of at least two selectable gear ratios, the hybrid vehicle powertrain being controlled using each of the following: An advanced controller having an associated first memory that stores one or more readable instruction sets for advanced controller tasks; A low-level controller having an associated second memory, the second memory storing one or more readable instruction sets for low-level controller tasks; The method described therein includes the advanced controller: Receive preview information related to the driving path of the hybrid vehicle, as well as one or more current state variables of the hybrid vehicle; Use the preview information to estimate the vehicle speed in the prediction window; Estimate the required combined torque in the prediction window by combining vehicle speed and preview information; Using a low-level controller to calculate the transmission ratio request model, one or more system dynamic models of the hybrid vehicle, and the one or more current state variables, calculate fuel consumption within a predicted range, including a portion of the driving path requiring the combined torque to satisfy the estimated prediction window, and calculate one or more low-level control variables used by the low-level controller, including equivalent factors, where the equivalent factors are conversion coefficients between fuel energy and electrical energy; and The lower-level control variable, including the equivalent factor, is transmitted to the lower-level controller; The method further includes the low-level controller: Receive one or more low-level control variables, including the equivalent factor, and one or more current state variables of the hybrid vehicle from the high-level controller; The first torque and the second torque are calculated using the one or more current state variables and the equivalent factor to provide fuel consumption within minimum and maximum limits for each of the first torque and the second torque; Calculate the gear ratio request delivered to the transmission; and Based on the first torque, the second torque, and the gear ratio, control signals are requested to be sent to the first power source and the second power source to control the power distribution between the first power source and the second power source based on the calculated fuel consumption, thereby minimizing the fuel consumption of the hybrid vehicle.

10. The method according to claim 9, wherein: The advanced controller calculates one or more control values ​​at a first sampling rate; The low-level controller calculates the first torque, the second torque, and the on / off signals of the first power source and / or the second power source at a second sampling rate; and The second sampling rate is a faster sampling rate than the first sampling rate.

11. The method according to claim 9, wherein, The second power source uses fuel, and the method includes a low-level controller determining and issuing an on / off command for the second power source.

12. The method according to claim 9, wherein, The lower-level controller calculates the transmission ratio request in the following manner: Identify feasible transmission ratios; Calculate the ratio of wheel torque request to transmission torque request; Find the optimal torque distribution for each feasible gear ratio; Calculate the cost of optimal torque distribution for each feasible gear ratio; and Choose the transmission ratio with the lowest cost from the feasible transmission ratios.

13. The method according to claim 12, wherein, The low-level controller calculates the cost of optimal torque allocation for each feasible gear ratio as the sum of fuel power cost, electricity cost, and change cost.

14. The method of claim 9, further comprising a low-level controller using the one or more current state variables and the equivalent factor to calculate an optimal torque distribution between the first power source and the second power source.

15. The method of claim 9, further comprising the higher-level controller using a driver model to calculate the one or more lower-level control variables, and the lower-level controller using actual driver control actions to calculate the first torque and the second torque.

16. The method of claim 9, further comprising the advanced controller using input from the navigation system to determine preview information, rather than driver control actions.

Images