Keywords

1 Introduction

The industry trend toward off-road and military vehicle electrification offers benefits like performance and efficiency, but poses challenges, including higher power demands and limited charging access in off-road conditions. Hybridization, with its hybrid powerpack management, emerges as a key solution to these issues, with Sivakumar's research highlighting its benefits like improved fuel economy and quieter operation [1].

Several power management strategies for series hybrid tracked vehicles have been explored across various industries. Wang's research investigates the power allocation for a hybrid electric bulldozer, implementing a rule-based method with four key guidelines that relate to battery charge and the required power, subsequently demanding power from generator unit [2]. Furthermore, Shabbir has enhanced control strategies of thermostat and power following, aiming to reduce fuel usage by moderating the shifts between the rules [3]. Zhai has also contributed with a heuristic energy management system known as the Optimal Primary Source Strategy (OPSS), which is designed to improve fuel efficiency and battery management in series hybrid electric tracked vehicles [4]. Zhang has furthered the development of rule-based and optimization-based approaches by integrating intelligent control elements like fuzzy logic filters [5]. This integration aims to address the challenges of fuel consumption that arise from frequent variations in engine loads, the non-ideal operation of generator sets, and disturbances from hydraulic pump torque.

In addition to the researches on power management, environmental impact of military conditions on electric powertrains is also investigated from the existing literature. Studies show that for a military vehicle application both the high-power demands and the extreme temperatures that electric drive systems of military vehicles must endure, as highlighted in the report by Stott and colleagues from the US Army Engineer Research and Development Center (ERDC) [6], should be considered. Thus, priority is given to performance and adaptiveness of the power management algorithm in this work from military field.

In this paper, the focus is placed on an adaptive power management system specifically designed for a series hybrid electric tracked vehicle, incorporating the literature investigation combined with an emphasis on adaptiveness to operating conditions of powerpack components. Figure 1 presents main components of a series hybrid tracked vehicle. The architecture includes two principal units: the electric traction unit and hybrid powerpack. While the traction unit is combination of electric motors, gearboxes and brake systems, the powerpack consists of battery, generator set and power distribution unit that integrates the power generated from these sources. The objective of the power management system is to dynamically regulate power distribution between the battery and the generator set, considering system variables such as driver demands, and the battery's state of charge. The findings obtained are aimed to be tested on Kaplan Hybrid [7], which is a hybrid tracked vehicle developed by FNSS.

Fig. 1.
figure 1

Main Components of a Series Hybrid Tracked Vehicle

Full size image

2 Power Management System (PMS) Development

The development of the PMS consists of several key stages: Modeling, Algorithm Design, and Verification.

2.1 Modeling the Hybrid Powerpack Plant

To test and evaluate PMS, a hybrid powerpack model with inputs including engine speed and generator torque requests, and actual outputs of these requests and the electric current at the powerpack’s output bus is created.

The accelerator pedal position is first translated into a torque request for the traction unit by referencing the electric motor's full load curves at the given speed. This torque, when multiplied by traction motor speed and unit efficiency, produces the power demand, as illustrated in Eq. (1). PD, App, TM,max, ωM and ηTr represents driver power demand, accelerator pedal position, maximum motor torque, traction motor speed and traction unit efficiency respectively.

$$ P_D = App\,T_{M,max} (\omega_M )\omega_M \eta_{Tr} $$
(1)

Next, the genset's power output is computed by multiplying the generator torque with engine speed which is controlled by output torque of the PI controller and informed by experiment-based lookup tables that factor in turbo lag and the engine's full load performance. Similar to the traction unit, the genset torque is limited by speed dependent full load curves. Computation is presented by Eq. (2) in which PGS, TG,req, TG,max, ωe and ηGS represents genset’s power output, requested generator torque, maximum generator torque, engine speed and GenSet efficiency respectively.

$$ P_{GS} = min\left[ {T_{G,req} ,T_{G,max} \left( {\omega_e } \right)} \right]\omega_e \eta_{GS} $$
(2)

The genset's excess power over the driver's demand becomes the battery's power input illustrated by PBattIn in Eq. (3) which may indicate discharge or charge. This power determines the battery's state of charge (SoC) through a simplified coulomb approach based on Movassagh’s research [8]. Equation (4) shows the transient SoC calculation, including bus voltage (VBus), battery capacity (QBatt) in Ah, and time step (Δt).

$$ P_{BattIn} = P_{GS} - P_D $$
(3)
$$ SoC\left( t \right) = SoC\left( {t - 1} \right) + \left[ {P_{BattIn} \left( t \right)} \right]\left[ {V_{Bus} Q_{Batt} } \right]^{ - 1} \Delta t $$
(4)

2.2 Algorithm Design of Power Management System

The developed power management algorithm combines a high-level controller for power allocation between the generator set (genset) and battery with a low-level controller that commands the genset to produce the required power. As outlined in Fig. 2, the power allocation algorithm employs an adaptive state-based power-following approach to determine the share of power provided by the battery and genset, increasing the generator's contribution as demand rises or the state of charge (SoC) of the battery reduces to charge the battery up to a desired charge level. Based on actual state calculated using driver demand and battery SoC, PMS also decides whether requested traction power should be limited in order not to exceed the desired battery current limits at states 6 and 8. Meanwhile, the genset drive system controls the engine and generator through speed and torque commands respectively to fulfill the power generation demands.

Fig. 2.
figure 2

Power Allocation Algorithm

Full size image

This work's primary contribution lies in its adaptiveness to operating conditions of powerpack components, which is accomplished through dynamic normalization of boundary conditions. Initially, the instantaneous available hybrid power, sum of the maximum battery discharge and available genset power, is calculated based on operating factors such as auxiliary power, bus voltage, and the temperature of the battery cells as well as the engine coolant at that time step. Then, all power limits are normalized relative to the maximum hybrid power. As a result, the state boundaries change adaptively and they are illustrated in Fig. 2 in normalized format. In the final step, the driver's demand is also normalized, and the state is determined. This method allows vehicle to operate the generator at varying power levels, influenced by the operating conditions for same traction demand.

In addition to the state determination and transition algorithm, a power allocation function is designated and refined based on the vehicle concept. The powerpack is battery-driven in states 1 to 3, as depicted in Fig. 2, while state 4 ensures a smooth transition between states 3 and 5 through hysteresis by adhering to previous state rules. States 5 to 8 activate hybrid mode, with GenSet power outputs adjusted by a hybrid function, optimized via a real-time brute-force method and parameter adjustment during DIL simulations. For instance, when prioritizing silence, the function permits greater battery discharge in hybrid states 4 and 5. Conversely, if the vehicle's objective is high performance, PMS is operated at 5 and 6 states and battery's SoC is maintained at higher levels maximizing generator usage by fine-tuning the hybrid function accordingly. This hybrid function at state 5 operates similar to power following methods and is modeled as a first-order polynomial in relation to the battery's state of charge. Its character is designated through the optimization of polynomial coefficients. Overall, PMS is designed to modulate the powerpack, promoting a higher discharge tendency at states with greater SoC and generator-driven charging tendency at states with lower SoC.

2.3 Simulations and Verifications

The power management system undergoes testing through Driver-in-the-Loop (DIL) and Model-in-the-Loop (MIL) methods. DIL testing utilizes a real-time driving simulator tailored for hybrid tracked vehicles, where the driver navigates the vehicle in a virtual environment transmitting the power demands to the ground through corresponding control blocks, powerpack blocks and the traction unit. The MIL approach, moreover, evaluates the system's performance with offline simulations, calibrating control parameters based on postprocessed results.

The battery power output distribution chart, observed during a model-in-the-loop (MIL) simulation, is presented in Fig. 3. In this simulation, the hybrid power allocation function has been optimized to ensure high performance while maintaining the battery's SoC at around 60%, charging below this level and discharging above it. Figure 3 also demonstrates that high SoC coupled with high power demand leads to maximum battery discharge, while lower SoC and power demand results in maximum battery charging, as expected in a vehicle with the main aim of performance. The hybrid function also ensures smooth transitions, effectively preventing undesired engine power fluctuations.

Fig. 3.
figure 3

Battery Power Output Distribution from MIL Simulation

Full size image

Moreover, real-time driver-in-the-loop (DIL) simulations are conducted to evaluate the system under transient loading conditions, such as high-speed steering maneuvers, using the DIL simulator specifically developed for series-hybrid electric tracked vehicles. Real-time power requirements are monitored and recorded throughout the vehicle's operation as illustrated in Fig. 4. Using the observations from these simulations, fine-tuning studies of power management system is supported.

Fig. 4.
figure 4

Steering Maneuver from DIL Simulation

Full size image

3 Conclusion

An examination of the simulation outcomes shows that thanks to the algorithm's adaptive capacity, the power management system coordinates the power sources to dynamically meet demands without exceeding the battery's and generator set's limitations. Additionally, the genset fulfills the required power generation within acceptable delay thresholds. In summary, the abilities of the power management algorithm and developed MIL/DIL environments underscore their importance for the future of hybrid tracked vehicles.