Deposits form on the air-fuel ratio sensor used for air-fuel ratio control in direct-injection gasoline spark-ignition (DISI) engines. They affect the signal response of the sensor, which provides the problems of the control of air-fuel ratio. In this study, the formation mechanism of deposits on the air-fuel ratio sensors was investigated using a gasoline DISI engine and the factors that affect the response of the air-fuel ratio sensor were clarified. As a result, it is found that the influence of PN concentration, gas flow rate, and gas flow speed have a large influence on the response time of the signals of the air-fuel sensors.

View full abstract

A demonstration test was conducted using a heavy-duty vehicle equipped with a hydrogen engine to transport cargo from summer to winter. A sampling survey of the lubricating oil revealed that the amount of water in the oil increased during the winter season. The water condensing in oil showed a strong relationship with both exhaust gas temperature and humidity. It was shown that water condenses in the oil when the temperature of the oil was lower than the dew point of blow-by gas. This result described that heating oil is effective for preventing oil from emulsifying in winter.

View full abstract

A technology that captures CO₂ in engine exhaust gas has been studied. From the viewpoint of efficient development, it is useful to consider a CO₂ capture system using models. Therefore, based on the experimental results of CO₂ adsorption characteristics, the model accounting for external film diffusion and intraparticle diffusion was considered. As a result, the adsorption rate model that can reproduce the experimental CO₂ breakthrough curves was established.

View full abstract

A system that can continuously capture CO₂ in engine exhaust gas under the on-board constraints has been investigated. The feasibility study was carried out with the 1D-CFD and real-scale experiments. As a result, it was shown that the calculation model could simulate the behavior of CO2 adsorption very well, that continuous CO₂ adsorption/desorption was possible, and that about 50% of the CO₂ could be captured on WLTC mode.

View full abstract

For the purpose of predicting the transmission efficiency during operation of a rubber belt-type CVT for motorcycles, including gear shifting, we developed theoretical equations. Conventional studies have often dealt with the theorization of constant speed states, and it has been difficult to predict the transmission efficiency because a theoretical analysis method for the behavior during gear shifting has not been established. We compared the results of calculations based on the developed theoretical equations with experimental results, demonstrating their effectiveness.

View full abstract

This paper proposes a method for detecting abnormalities in the electric drive system of a large EV that combines two motors and planetary gears. This drive system is controlled and driven to balance the rotational speed difference and torque of motors 1 and 2. The proposed method detects abnormal decreases in torque in one motor in comparison to the normal torque output of the other motor by checking changes in the control amount for balancing the torque of the two motors. This method was verified under operating conditions in which the actual torque of motor 2 decreased in accordance with an increase in rotor temperature. As a result, it was confirmed that a decrease in actual torque due to an increase in rotor temperature could be detected with high accuracy.

View full abstract

With the advancement of electric vehicle (EV) performance, the increasing number of electrical components required to connect various power devices has led to issues such as higher part count, weight, and cost. This research introduces a novel power unit that integrates all power conversion functions within the drive motor and inverter. This unit not only combines AC regular charging and rapid DC charging functions but also enhances drive output and efficiency compared to conventional systems by changing motor winding connections. By developing such an integrated power unit, we aim to improve EV performance while reducing costs. This paper reports on the operating systems of the integrated power unit and the results of actual machine verification, demonstrating the successful integration of AC and DC charging into the motor drive system and paving the way for more compact and efficient EV designs.

View full abstract

For the popularization of electric vehicles, it is important to improve battery performance and safety. Therefore, we have investigated the use of laminated steel sheets for cell cases to improve the impact resistance of pouch-type Lithium-ion batteries. In this study, a pouch-type Lithium-ion battery using laminated steel sheets for the cell case was made, and its charge-discharge behavior and safety were evaluated by a crushing test using a round bar, which showed good results.

View full abstract

In order to achieve front-loading for power unit development, it is necessary to predict pollutant emissions performance of the full vehicle in the early stages of development. In this study, a methodology for building an EiL environment by coupling a concept model of a hybrid power unit and an engine testbed is described. Additionally, this paper discusses a prediction method for pollutant emission performance using EiL.

View full abstract

As an EV community bus field test was conducted in Iizakaonsen area, Fukushima, EV’s energy consumption and solar irradiance under various driving and weather conditions were measured. Then, simulation models were developed to analyze the impact of PV generation on EV energy consumption, which are under different ambient temperatures, solar heights and driving conditions, such as various driving distance and route slope. Finally, the potential for electric energy reduction with installation of PV system was evaluated, with the result showing that 36% electric energy consumed by EV can be covered by PV generation with 365 days per year and total 12,000 km operation.

View full abstract

Tread profile of tire is closely related to cornering characteristics, and a tread profile that increases cornering stiffness is important when considering handling stability. In order to increase cornering stiffness, it is important to increase the lateral shear stiffness of the tread. However, detailed investigations have revealed that this is not always the case. In particular, cases where predictions and reality are reversed can cause difficulties in performance design. The purpose of this research is to elucidate the mechanism behind this reversal phenomenon. The causal relationship between tread profile and cornering stiffness was quantitatively clarified through calculations using a detailed FEM model and experiments using prototype tires.

View full abstract

Currently, to meet tension stiffness target of the hood, CAE is used repeatedly to determine suitable mastic sealer arrangement. Several complex constraints and combination of the constraints needed for manufacturing require numerous cases of simulation. In this research, we developed a design technology that can rapidly generate multiple arrangement patterns using machine learning. Using this technology, it is expected to reduce simulation costs which will improve the efficiency of the hood development process.

View full abstract

Bicycles are classified as non-motorized vehicles in road traffic law. This law stipulates that, in principle, bicycles must ride on the roadway on roads where sidewalks and roadways are separated. However, on main line roads with heavy traffic volume, the regulated velocity for vehicles is high, ranging from 40km/h to 50km/h. Furthermore, on main line roads, not only passenger vehicles but also large size vehicles such as trucks pass through, and the traffic volume is high. Therefore, based on the above principles, it is difficult to ensure the safety of cyclists. For this reason, there are many cases where bicycles are permitted to ride on sidewalks as an exception to the principles. In recent years, collisions between bicycles and pedestrians resulting in the death or serious injury of pedestrians have become a social problem. Therefore, it is also important to consider the safe coexistence of pedestrians and cyclists on sidewalks. From the viewpoint of the vehicle, the velocity of bicycles is faster than that of pedestrians, and their behaviors are difficult for drivers to grasp. Furthermore, if a bicycle collides with a vehicle, there is a high possibility that the cyclist will be killed or seriously injured. Therefore, it is best to avoid accidents involving bicycles as much as possible. In this paper, the author analyzes accident statistics data for collisions between passenger vehicles and bicycles, and between large size trucks and bicycles. As a result, the relationship between the collision position on the vehicle side, the collision velocity, and the degree and type of injury to the cyclist was clarified. Based on these findings, the author would like to discuss the desirable coexistence of automobiles and bicycles in road traffic.

View full abstract

To achieve global shape optimization using isogeometric analysis, we have developed a scheme that automatically repeats the optimization process while adjusting the number of control points in the structure. By incrementally updating both the number of control points and the search area, the design space can be explored more extensively, minimizing the risk of shape breakdown. To validate the effectiveness of the proposed shape optimization method, we evaluate the performance of the optimized structure through a computational example.

View full abstract

In this study, a head and leg injury prediction method in car-to-pedestrian collisions was constructed by using FE analyses and deep learning. Thirty-seven collisions between small adult female and sedan-type vehicles were supposed. A prediction model for HIC and a leg bending moment was built using time series data of pressure obtained from tube-type sensor inside bumper absorber during a collision as input data. As a result of deep learning, the coefficient of determination was 0.86 for HIC and 0.74 for bending moment, and the degree of injury to pedestrians could be predicted.

View full abstract

Recently, research and development of autonomous driving technology has been actively conducted. It is necessary to realize safe and comfortable autonomous driving functions. This study aims to elucidate the relationship between human sensibility and posture-holding behavior for braking during autonomous driving. In this paper, we classify subjects by performing cluster analysis on the results of sensory evaluation. Therefore, subjects are classified into three groups. Then the characteristics of posture-holding behavior of each group were shown and the relationship between psychology and posture-holding behavior was investigated. The tendency of posture-holding behavior and psychological evaluation for each group is shown.

View full abstract

Data-driven approaches using machine learning methods are gaining attention as substitutes for Computer-Aided Engineering (CAE). These methods perform well on tasks that are similar to the training data. However, during the planning phase or with new designs, tasks often differ significantly, making accurate predictions challenging. Recent studies have utilized fine-tuning to adapt to new tasks, but a key challenge remains there is no generalizable method to determine which parts of the pre-trained neural network weights should be frozen or updated. To address this issue, this study proposes Average Weights-Based Fine-Tuning (AWFT), a method that dynamically decides whether to freeze or update each edge of a neural network. We report that AWFT enables fast adaptation to new tasks while maintaining high accuracy.

View full abstract

The numerical analysis of gas jet characteristics using the open CFD software OpenFOAM was validated for further effective use of gas fuel injection in an internal combustion engine. The under-expanded jets formed when hydrogen, methane, nitrogen, and helium gases are injected into the atmosphere from a gas fuel injector were compared with visualization experiments using the Schlieren method. First, the differences in calculation results due to the turbulence model are assessed in the case of hydrogen injection. Then, the k-omega SST turbulence model is selected, and the cases of gas injection at nozzle pressure ratios (NPR) of 5 and 7.5 are compared and analyzed for jet characteristics near the nozzle and downstream mixture formation. The scaling laws of axial mass fraction and Mach number distributions, and also the tip penetration, are validated.

View full abstract

During refueling of an automobile, air is entrained by fuel flow at the inlet of the filler pipe. This multiphase flow in the filler pipe was evaluated in a refueling experiment using water as the simulated fuel. It was observed a transition from spray flow to liquid film flow at a distance from the fuel nozzle exit close to 450 mm. A model of the relationship between entrained air flow rate and tank pressure was developed. In the model, droplets inside the filler pipe were assumed to have spherical and hemispherical shapes. The prediction by the hemispherical shape model was closer to the experimental values than that by the spherical model.

View full abstract

In order to improve the accuracy of hydrogen consumption measurement using a Coriolis flowmeter, we conducted chassis dynamometer tests using a fuel cell vehicle. In this report, we used high-pressure hydrogen that is more than 1MPa and a hydrogen flow meter with a pressure adjustment mechanism to improve measurement accuracy. As a result, even in the transient driving cycles such as WLTC and JE05, where past tests showed variations, test results by Coriolis flowmeter were obtained that closely matched the measured values using the gravimetric method, which was determined from the difference in the weight of the gas cylinder before and after the test.

View full abstract

The target substances at Real Driving Emission (RDE) test are mainly CO₂ and NOx. In this study, due to its high global warming potential, N₂O emission was focused as RDE tests. An on-board FTIR analyzer was installed with a conventional portable emission measurement system (PEMS) to a diesel light duty vehicle. As a result, the CO₂ concentrations showed slight differences depending on the measurement principles. It seems that it is caused by Dry/Wet correction. The emission behavior of N₂O was compared with CO and NO emission behavior to discuss N₂O emission processes.

View full abstract

In order to estimate fuel economy of engine vehicles, we usually use chassis dynamometer. Engines on these vehicles are cooled by using the cooling fan of chassis dynamometer. The recent electric vehicles (EVs) manage the thermal energy of all components during driving. Therefore, it is necessary to replicate wind flows of real driving correctly by using the cooling fan. In this study, we modified the cooling fan outlet and verified the effect of it by using an engine vehicle and electrified vehicles. During the driving test cycle, the fuel economy and emissions of the hybrid electric vehicle (HEV) was improved by using the modified fan outlet in comparison with the conventional fan outlet.

View full abstract

In this study, it was experimentally investigated whether the fatigue limit of spot-welded tensile-shear joints made of 590 MPa-class high-tensile steel sheet was improved by applying a single overload in advance. The fatigue limit was improved by two times due to a single overload of 3.04 kN. The single overload caused a large compressive residual strain on the internal surface near the nugget edge on the plate width centerline. This is the main factor behind the extension of fatigue life and improvement of fatigue limit due to a single overload.

View full abstract

In driverless autonomous driving, there is a possibility that the vehicle gets stuck when autonomous vehicles face unpredictable situations. In this study, we propose a detection mechanism that can accurately and quickly identify the stuck state among stops to remotely assist in resolving the situation. The results of the experiment demonstrate that the proposal shortens the time required to resolve the stuck state and mitigates its impact on the traffic flow.

View full abstract