Power Take Off in Regenerative Auxiliary Power System

Modelling and Configuration

      Many different factors and criteria should be considered in order to design an acceptable RAPS. To power the service vehicle's auxiliary devices from braking recovered energy, a regenerative braking system should be integrated into the vehicle powertrain. The connection of the RAPS to the vehicle powertrain, total system configuration, safety, weight, and size of components are the most important factors. As the first step, integration of the RAPS to the vehicle powertrain should be investigated in detail. It is vital that connection of the RAPS does not cause any major modification to the existing vehicle; otherwise, the initial cost of the total system and the safety concerns will reduce industry interest. 

     The other important point is that the designed RAPS must be modular and easy-to-install in order to reduce installation time and costs. Additionally, this research is intended for use in any service vehicle with auxiliary devices, so RAPS components and especially their models should be generic, modular, and flexible for the creation of scalable powertrains and RAPS components models.

     Given these considerations, in this study, the RAPS components are designed with the different configurations of different powertrains in mind. By utilizing this method, RAPS components can be designed separately and added to an existing vehicle. In the first part of this chapter, the design configuration of the RAPS is discussed, and the two most possible categories for integration configurations are described. The modelling concepts of different RAPS components are explained in the last part.

      

System Configuration and Potential Energy Recovery

As illustrated in Figure 3.1, a RAPS consists of different electrical and mechanical parts. Input mechanical energy will be extracted from the vehicle's powertrain through the "RAPS Connection Point". There are different options for this connection point, and they will be explained in later sections. This extracted mechanical energy is in the form of torque and angular velocity. Based on the specifications and limitations of the generator, it is possible to change the range of the input angular velocity to the acceptable range of the generator.

Figure 3.1 Regenerative auxiliary power system (RAPS) configuration

The "Generator" will convert the mechanical energy into the electrical energy and produce electrical energy, in the form of current and voltage, to charge the "Electrical Energy Storages". The output energy of the generator will be sent to the "Electrical Energy Storages" through the "Power Electronics" module. The "Power Electronics" module controls the input energy flow to the "Electrical Energy Storages". The "Power Electronics" module function is to monitor the electrical energy flow and maintain the current and voltage in a form and range that is suitable for charging the "Electrical Energy Storages". As previously mentioned, the option for plug-in charging is considered for the proposed RAPS. The "Plug-in Charging" part will send the plug electrical energy to the "Electrical Energy Storages" through the "Power Electronics" module. The "Power Electronics" module has the ability to convert the plug-in AC electricity to the acceptable DC power by the "Electrical Energy Storage".

The "Electrical Energy Storages" stores the electrical energy and provides power to the vehicle's auxilary devices. The "Power Management Controller" will monitor the power demand, battery state of charge (SOC), vehicle status, and brake pedal signal, and based on these conditions, it will control the system's power flow. It should be noted that the one-way arrows in Figure 3.1 show the power flow in the RAPS, but the two-way arrow between the whole RAPS and the "Power Management Controller" is the flow of data and state conditions information.

System Integration to Powertrain

There are different parts in a vehicle powertrain that can be used to extract mechanical power for the generator. Possible configurations can be catagories in two main groups based on the amount of the demanded power; however, as explained in detail in the following sections, the limitation of the powertrain and connecting components are also very important in this classification. Two main groups of possible configurations are defined as:

Low Power Demand System (Serpentine Belt) Configuration

For service vehicles with power demands, it can be considered that the generator, which is mostly an alternator in this configuration, is connected to the engine's serpentine belt. The serpentine belt is a continuous belt used to power different devices, which are extracting their power directly from the engine. In this system, the serpentine belt connects the engine crankshaft to multiple devices including the alternator, water pump, air pump, air conditioning compressor, power steering pump, etc. In service vehicles, there is usually a space allocated to the attachment of another device, usually a compressor, to be powered by the serpentine belt. In the design of RAPS with low power demands, this space, as shown in Figure 3.2, can be used to install the extra generator. In addition, for service vehicles without this extra space, there is the option to utilize a higher output alternator instead of the Original Equipment Manufacturer (OEM) alternator.

Figure 3.2 Available space to install generator in the low power demand targeted vehicle "Ford Transit Connect Cargo XL-2010"

Figure 3.3 shows schematically the low power configuration and attachment of the generator to the engine. In this figure, the small arrows show the mechanical or electrical power flow and the big arrows emphasize the direction of the energy flow. The drawback of the low power demand configuration is that the maximum captured kinetic energy during braking is limited by the size of the generator-alternator. Thus, the regenerative energy is limited by the available space for the generator and the characteristics of the serpentine belt, especially its maximum tension capacity. This configuration is the main solution for vehicles, such as front wheel vehicles, with little space for adding the generator to the powertrain. 

Figure 3.3 Low power demand systems (the serpentine belt) configuration

High Power Demand System (Power Take Off) Configurations

           In service vehicles with high power demands, the generator can be driven by a power take off (PTO) module. In a vehicle, the engine produces power and transfers it to the wheels through the bell housing, transmission, drive shaft, differential and final axles. This condition can be better illustrated in a rear drive vehicle schematic as shown  in Figure 3.4. In theory, it is possible to extract the mechanical power for the generator from 6 locations. However, PTO should allow RAPS to run directly from the engine when the vehicle is stopped. This option is possible if a PTO can be installed at a point between the engine and the transmission.

Figure 3.4 Possible power extraction (PTO Installation) points in the high power demand system configuration 

In order to extract power at each of these points, a different PTO is needed. Generally, the required PTOs can be categorized into three main types;

• "Split Shaft PTO" : As shown in Figure 3.5, there are different designs and sizes for this kind of PTO. In Split Shaft PTOs, the input shaft runs two output shafts where one is a through shaft and the other is for power take off. Considering Figure 3.4, this type of PTO can be used in the following point:

Point 1: Shaft connecting engine to the bell housing

Point 2: Connecting the shaft of the bell housing of the transmission

Point 5: Driveline shaft (drive shaft)

Point 6: Any of the axles

 

Figure 3.5 "Split Shaft PTO" 

It should be considered that in order to install the Split Shaft PTO, one of the power transferring shafts of the existing vehicle requires modification; also, for any of the considered points there should be enough space to install the PTO. Hence, from all the possible installation points for Split Shaft PTO, point 5 is the most acceptable one. This location, however, does not allow the generator run directly from the engine when the vehicle is stationary. Some Split Shaft PTO could be an option; however, it requires driver action which may make this solution generally undesirable. 

• "Side Countershaft PTO": These types of PTOs are most common, and normally, they are just referred to as PTO- without any prefix. As shown in Figure 3.6, these PTOs will be installed to the output of the transmission. In most service vehicles or heavy-duty vehicles there is a place to install these PTOs. Considering Figure 3.4, this type of PTO can be used in:

Point 4: Transmission output to the drive shaft

However, installing a PTO for generator power extraction at Point 4 has the same setback as Point 5 and may not be desirable.

Figure 3.6 "Side Countershaft PTO"

• "Transmission Aperture PTO": In most heavy-duty vehicles, especially service vehicles, the option for mechanical powering of extra devices has been allocated. In these cases, as shown in Figure 3.7, the vehicle transmission has a special place for installing a "Transmission Aperture PTO" to power up other devices. Considering Figure 3.4, this type of PTO can be used in: 

Point 3: Transmission Aperture

Figure 3.7 "Transmission Aperture PTO"

There are different scenarios that should be considered for all configurations:

I) Braking: It is expected that RAPS employs the waste energy during braking and maximizes the use of regenerative braking energy. The system will be considered to be in the regenerative braking phase if all the following conditions are met: 

1. The vehicle is braking (braking signal is sent to the controller)

2. The vehicle speed is higher than a threshold (higher than 16 km/h based on suggestion)

3. The EES is not full (battery SOC level is lower than 100%)

For a low power demand system (serpentine belt) configuration during braking, in an automatic transmission, the torque converter still gets braking powers from the drive wheels to pass on the serpentine belt. However, in manual transmissions (or automatic transmissions at very low vehicle speeds) during braking, the clutch will disconnect the engine-belt from the transmission (or the torque converter does not pass the remaining braking energy.); thus, regenerative braking energy is just part of the kinematic energy in the crankshaft and moving inertia of the engine.

For the high power demand system (PTO) configuration, segments of the drive wheel braking powers, kinetic energy in the crankshaft, and moving inertia of engine during braking can be converted into the regenerative braking power. For this configuration, in manual transmissions during braking, the clutch will disconnect the engine from the transmission and other parts of the driveline. Therefore, if point 1 is considered as the RAPS connection point to vehicle drivetrain, regenerative braking energy will be just part of the kinetic energy in the crankshaft and moving inertia of the engine. Conversely, if any of points 3,4,5 and 6 is considered as the connection point, regenerative braking energy will be limited to part of the braking power from the drive wheels. It should be noted that in most vehicles targeted for high power demand configuration, the bell housing (clutch/torque-converter) is a part of transmission package since connection point 2 will have the same conditions as connection point 3 in most cases.

For the high power demand system (PTO) configuration in automatic transmissions, the bell housing will not disconnect any part of the vehicle powertrain during braking; thus, regenerative braking energy can be obtained from the drive wheel braking powers, kinetic energy in the crankshaft, and moving inertia of engine altogether. Additionally during very low vehicle speeds, the second condition of the regenerative braking phase is not satisfied; therefore, there will be no regenerative braking when that bell housing disconnects the engine from the other parts of vehicle powertrain.

II) Vehicle Movement: If regenerative braking energy is not sufficient to charge the batteries, the power management controller charges the batteries directly from the engine during peak engine performance. In this scenario, since the moving power is transferred from the engine all the way to the drive wheels, RAPS can extract the power in all of the aforementioned possible configurations.

III) Vehicle Stops: The RAPS target is to prevent the vehicle from idling; hence, while the vehicle is stopped, the engine ought not be active. During long stops, the auxiliary devices power consumption decreases the battery SOC to a critical level. In this condition, the system should allow a generator to be run directly from the engine when the vehicle is stopped. A serpentine belt configuration will work without any limitations in this scenario. For the PTO configurations, the power extraction point should be anywhere between the engine and the vehicle transmission, or the option of a PTO with a disconnecting clutch and driver action should be considered.

Given all of these scenarios, the "Transmission Aperture PTO" is the ideal and the most cost effective solution for the realization of the RAPS in high power demand service vehicles.

Components Modelling

As the system model will be used for the optimization process, the components models should be generic, modular and flexible. The components models need to be scalable so that the optimization method can determine their optimal sizes.

There are two modelling approaches that can be used:

a) Forward-looking: As shown in Figure 3.8, modelling and simulation starts from the driver's points of view. The driver's demanded power is sent to the powertrain components, and the resulting power that is available from the powertrain is fed to the final drive and wheels. This type of modelling is more realistic compared to the backward looking models.

Figure 3.8 General forward-looking vehicle model

b) Backward Looking: In this approach, the required power is determined based on the known drive cycle data. Illustrated in Figure 3.9, this power demand is then calculated and transferred through the powertrain components to the engine or another main power source. Through this process and using the components' efficiencies, the power needed in each component is calculated. In this approach, the detailed dynamics of the components and the vehicle system is not considered; however, this will crate less complicated models compared to the forward-looking models, which require solving differential equations.

Figure 3-9: General backward looking vehicle model

Based on the fact that in this study, only the overall power consumption of the vehicle is of interest, the effects of vehicle dynamics due to the suspension can be safely ignored even though the model is not as realistic as the forward-looking model. The backward-looking approach modeling is chosen for this work to fulfill the purpose of optimization. Simultaneously, the developed vehicle model formulation is suitable for the power consumption objectives. The total system model consists of powertrain components (engine, bell housing, transmission, differential) and RAPS components (batteries, power electronic, generator, auxiliary load). The goal of this research is to design and optimize a suitable regenerative braking system, which can be added to a service vehicle’s powertrain. As previously mentioned, the modifications to the vehicle powertrain should be kept to a minimum to make the changes affordable for the industrial purposes. Thus, there should not be any change to the configuration of the mechanical components of the vehicle. The powertrain components are modeled using the scalable backward-looking approach proposed by the Guzzella and Rizzoni [43]. In this approach, the actual consumed power of a component is calculated by multiplying the required component torque and component current velocity given the component’s efficiencies.

 

General Vehicle Model

 In order to fulfill the need for a simple model for the backward-looking modeling approach, a simple vehicle body model is considered. This model only utilizes the drive torque, rolling resistance, and the resistive aerodynamic forces. The vehicle body model receives the demanded longitudinal velocity of vehicle (𝑉_𝐷𝑒𝑠) and demanded longitudinal acceleration of vehicle (𝐴_𝐷𝑒𝑠) as the input data. Based on these, the model calculates the torque (𝑇_{𝑊ℎ𝑒𝑒𝑙}) and angular velocity (𝜔_{𝑊ℎ𝑒𝑒𝑙}) at the wheels (or axles) as follows:

Wwheel = (Vdes)/(Reff)

Twheel = (Ftotal)x(Reff)

in which: 

𝐹𝑇𝑜𝑡𝑎𝑙 = 𝐹𝐷𝑟𝑖𝑣𝑒 + 𝐹𝐷𝑟𝑎𝑔 + 𝐹𝑅𝑅

and

𝐹𝐷𝑟𝑖𝑣𝑒 = 𝑀𝑇𝑜𝑡𝑎𝑙 𝐴𝐷𝑒𝑠

Fdrag = (1/2)(Cd)(𝜌)(Vdes)^2(A)

𝐹𝑅𝑅 = 𝐶𝑅𝑅 𝑀𝑇𝑜𝑡𝑎𝑙 𝑔 cos 𝛼

 

where 𝐹_{𝑇𝑜𝑡𝑎𝑙} , 𝐹_{𝐷𝑟𝑖𝑣𝑒} , 𝐹_{𝐷𝑟𝑎𝑔} , and 𝐹_𝑅𝑅 represents total vehicle longitudinal force, drive force, total aerodynamics drag force, and total tire rolling resistance force, respectively. 𝐶_𝐷 is the drag coefficient, 𝜌 is the air density, 𝐴 is the frontal area of vehicle, 𝐶_𝑅𝑅 is the rolling resistance of the tire, 𝑔 is the gravitational acceleration, 𝛼 is the road grade angle in radian, 𝑅_𝐸𝑓𝑓 is the effective tire radius, and 𝑀_{𝑇𝑜𝑡𝑎𝑙} is total vehicle mass, which is considered to be: 

 

𝑀𝑇𝑜𝑡𝑎𝑙 = 𝑀𝑉𝑒ℎ + 𝑀𝐸𝐸𝑆 + 𝑀𝐶𝑎𝑟𝑔𝑜

 

where 𝑀_𝑉𝑒ℎ is the vehicle mass before installing the EES packs or loading the cargo, 𝑀_{𝐶𝑎𝑟𝑔𝑜}  is the cargo weight of the vehicle and 𝑀_𝐸𝐸𝑆 is the mass of the electrical energy storage (EES) and other electrical parts. The total drive power demand (𝑃_{𝐷𝑟𝑖𝑣𝑒}), the power needed to move the vehicle, is equal to:

 

𝑃𝐷𝑟𝑖𝑣𝑒 = 𝜔𝑊ℎ𝑒𝑒𝑙 𝑇𝑊ℎ𝑒𝑒𝑙 + 𝑃𝐿𝑜𝑠𝑡

 

in which 𝑃_{𝐿𝑜𝑠𝑡} is the lost power in the powertrain system.

 

Final Drive (Differential)

The demand torque and angular velocity at the wheels (or axles) are the inputs for the final drive. After applying the final drive ratio (𝑁_𝐹), the demand torque (𝑇_𝐷𝐿) and angular velocity (𝜔_𝐷𝐿) at the driveline are calculated as:

 

𝜔𝐷𝐿 = 𝑁𝐹 𝜔 𝑊ℎ𝑒𝑒𝑙
Tdl = (Twheel)/(Nnr*nf)

 where 𝜂_𝐹 is the final drive efficiency.

 

PTO

 The PTO consists of a set of gears that transfer the extracted mechanical power from the powertrain to the generator in the high power demand configuration. In order to keep the model as simple as possible, the PTO is modeled by a gear ratio (𝑁_𝑃𝑇𝑂) and an efficiency factor (𝜂_𝑃𝑇𝑂). The modeling formula for the PTO defines its output torque (𝑇_{𝑃𝑇𝑂−𝑂𝑢𝑡}), which is equal to the generator torque in the high power demand configuration, and its output angular velocities (𝜔_{𝑃𝑇𝑂−𝑂𝑢𝑡}) as:

 

Tpto-out = (Tpto-in)/(Npto)npto*Cact-pto
𝜔𝑃𝑇𝑂−𝑂𝑢𝑡 = 𝑁𝑃𝑇𝑂 𝜔𝑃𝑇𝑂−𝐼𝑛

 

in which 𝐶_{𝐴𝑐𝑡−𝑃𝑇𝑂} is the PTO activation control signal provided by the power management controller. In a high power demand configuration, the PTO is responsible for extracting the energy from the powertrain for the generator. This power extraction occurs in two general cases: i) Battery pack SOC decreases to a critical level where the power management controller decides to charge the battery pack (during the vehicle’s movement or stop); and ii) regenerative braking. In either of the two general cases, the PTO’s activation control signal will be one (active); otherwise this value will be zero (not active).

The power management controller, in general, controls the flow of power by monitoring the auxiliary power demand, battery state of the charge, vehicle status, brake pedal signal, and amount of power produced by the generator. Power management is considered as a higher level control algorithm that monitors the power flow between the vehicle powertrain, generator, and ESS as shown in Figure 3-10. The controller makes sure that the generator produces the maximum amount of power during regenerative braking to maximize the RAPS efficiency. In addition, the controller charges the batteries directly from the engine, during the peak engine performance, when the state of the charge is lower than a critical value.

Figure 3-10: Power management controller input/output signals

In order to determine when to charge the batteries with the direct power from the engine, the controller considers the following criteria:

1. Once the SOC drops below the “High-level battery SOC threshold”, the engine will be used to charge the battery packs if it is working at high efficiency (30% or higher);

2. Once the SOC drops below the critical value or “Low-level battery SOC threshold”, the engine will be used to charge the battery packs under any condition, even low efficiency or idling, to prevent battery packs from any damage and life cycle reduction;

3.  During braking conditions, there will certainly be no direct charging from the engine. 

In order to make the system model as simple as possible, the controller utilizes a rule-based strategy that is adequate for control purposes in the design optimization of the proposed RAPS. The low and high-level battery SOC thresholds are defined in the optimization process, and will be further explained in Chapter 4.

 

Transmission (Gearbox)

The demand torque and angular velocity at the transmission output shaft (driveline) are the inputs to transmission model. After applying the transmission ratio (𝑁_𝑇), the demand torque (𝑇_𝑇 ) and angular velocity (𝜔_𝑇) at the transmission are calculated as transmission model outputs:

 

𝜔𝑇 = 𝑁𝑇 𝜔𝐷𝐿
Tt = (Tdl)/(Nt*nt)

in which 𝜂_𝑇 represents the transmission efficiency. The values of the transmission ratio are calculated using a look-up table indexed by the vehicle’s longitudinal speed. It should be mentioned that the connection between the transmission and engine is provided by the bell housing (clutch / torque-convertor). Due to the fact that a simple component modeling approach is selected for this study, the effects of the bell housing part are considered in the transmission efficiency.

 

Generator

The output electrical power of the generator (𝑃_𝐺), in the form of output current (𝐼_𝐺) and output voltage (𝑉_𝐺) is calculated as: 

𝑃𝐺 = 𝐼𝐺𝑉𝐺 = 𝜂𝐺 𝑇𝐺𝜔𝐺

 

where the 𝑇_𝐺 and 𝜔_𝐺 are the input torque and angular velocity to the generator, and 𝜂_𝐺 is the generator efficiency from the lookup tables indexed by the input torque and angular velocity to the generator. In the low power demand configuration, the generator is connected to the engine through the serpentine belt. During regenerative braking, the bell housing transfers braking powers from the transmission to the serpentine belt. In this case:

Tg = (areg*Tt)/(Ng)

𝜔𝐺 = 𝑁𝐺𝜔𝑇

 

in which 𝛼_𝑅𝑒𝑔 is the regenerative braking coefficient and 𝑁_𝐺 is the generator ratio due to serpentine belt pulleys. During braking conditions (due to safety, general limitations of adding regenerative braking, system efficiency, and size limitation of the generator), conventional mechanical brakes still work. Therefore, by adding regenerative braking to the vehicle, just a percentage of the braking torque energy can be captured by the generator. The regenerative braking coefficient (𝛼_𝑅𝑒𝑔) represents this percentage. On the other hand, during vehicle movement or vehicle stop scenarios, the generator demanded torque, which is determined by the controller as it charges the battery in a critical level SOC, will be provided by the engine. In these cases:

Tg = (Te_charge)/(Ng)

𝜔𝐺 = 𝑁𝐺𝜔𝐸

 

where 𝑇_{𝐸_𝐶ℎ𝑎𝑟𝑔𝑒} is the engine demand  torque for direct  battery charging conditions  (when the generator is run by the engine instead of regenerative breaking energy). In these scenarios, the engine demanded torque for direct charging (𝑇_{𝐸_𝐶ℎ𝑎𝑟𝑔𝑒} ) will be added to the transmission demanded torque (𝑇_𝑇) and the required torque to overcome the resistance forces at the engine (𝑇_𝑅) in order to determine the total demand torque from the engine (𝑇_𝐸):

𝑇𝐸 = 𝑇𝑅 + 𝑇𝑇 + 𝑇𝐸𝐶ℎ𝑎𝑟𝑔𝑒
𝜔𝐸  =  𝜔𝑇

 

Based on the fuel consumption rate during idling, the engine idling torque is estimated and this value is considered to be the needed torque to overcome the engine resistance forces (𝑇_𝑅).

In the high power demand configuration, a generator is connected to the PTO. Therefore, generator torque and angular velocity are equal to the output torque of the PTO (𝑇_{𝑃𝑇𝑂−𝑂𝑢𝑡}) and output angular velocities of the PTO (𝜔_{𝑃𝑇𝑂−𝑂𝑢𝑡}), respectively. When considering the “Transmission Aperture” PTO configuration in the regenerative braking phase, the generator torque and angular velocity are determined as: 

𝜔𝐺 = 𝑁𝑃𝑇𝑂𝜔𝑇

 

During vehicle movement or vehicle stop scenarios, if the battery SOC decreases to the critical level, the generator will be powered directly by the engine (direct battery charging). In these scenarios, the generator torque is calculated as:

Tg = (areg*Tt)/(Npto)npto*Cact-pto
𝜔𝐺 = 𝑁𝑃𝑇𝑂 𝜔𝐸

 

Similarly, in the low power demand configuration, for the direct battery charging condition  in the “Transmission Aperture”, the PTO configuration equations (3-20) and (3-21) are  valid.

Engine

The engine model should be scalable so that the developed model can be easily modified for different vehicles. Similar to the model used in [43], the applied engine model in this study has the scalability and composability features. Using a scalable model, the vehicle components that belong to the same class (for example ICEs) can be modeled using the same basic model. The important factor is that the model should be independent from the component size and can be scaled based on a simple scalar variable such as displacement or power rating. The composability feature is concerned that the system components’ model can easily compose the other related parts.

Storing, converting, and transferring the energy in a vehicle is concerned with three domains: chemical, mechanical, and electrical. In each energy domain, power is equal to the product of flow variable and effort variable. In the powertrain analysis, angular velocity and torque are the flow and effort variables, respectively [44]. The Willan’s line modeling approach, which is proposed by Rizzoni et al. [49], normalizes the flow and effort variables to create a model that is independent of scaling. The Willan’s line concept for a generic energy converter, ICE in this case, is shown in Figure 3-11.

Figure 3-11: Willan’s line modeling concept

Figure 3-12: Common electric circuit-based models for batteries

The final SOC of all working days will be the same since the SOC starts from the same level of 95% each day and same drive cycle is used for each working day.