Search Results

Recognizing road conditions using onboard sensors is significant for the performance of intelligent vehicles, and the road profile is a widely accepted representation both in the temporal and frequency domains, greatly influencing driving quality. In this paper, a recurrent neural network embedded with attention mechanisms is proposed to reconstruct the road profile sequence. Firstly, the road and vehicle sensor signals are obtained in a simulated environment by modeling the road, tire, and vehicle dynamic system. After that, the models under different working conditions are trained and tested using the collected data, and the attention weights of the trained model are then visualized to optimize the input channels. Finally, field experiments on the real vehicle are conducted to collect real road profile data, combined with vehicle system simulation, to verify the performance of the proposed method.

Plug-In Hybrid Vehicles (PHEV) have been of significant importance recently to comply with future CO2 and pollutant emissions limit. However, performance of these vehicles is closely related to the energy management strategy (EMS) used to ensure minimum fuel consumption and maximize electric driving range. While conventional EMS concepts are developed to operate in wide range of scenarios, this approach could potentially compromise the fuel consumption benefit due to the omission of route and traffic information. With the advancements in the availability of real-time traffic, navigation and driving route information, the EMS can be further optimized to extract the complete potential of a PHEV. In this context, this paper presents application of predictive energy management (PEM) functionalities combined with information such as live traffic data to reduce the fuel consumption for a P1/P3 configuration PHEV vehicle.

Unmanned Aerial Vehicles (UAVs) encounter various uncertainties, including unfamiliar environments, signal delays, limited control precision, and other disturbances during task execution. Such factors can significantly compromise flight safety in complex scenarios. In this paper, to enhance the safety of UAVs amidst these uncertainties, a control accuracy prediction model based on ensemble learning abnormal state detection is designed. By analyzing the historical state data, the trained model can be used to judge the current state and obtain the command tracking control accuracy of the UAV at that instant. Ensemble learning offers superior classification capabilities compared to weak learners, particularly for anomaly detection in flight data. The learning efficacy of support vector machine, random forest classifier is compared and achieving a peak accuracy of 95% for the prediction results using random forest combined with adaboost model .

Heavy-duty transportation is one of the sectors that contributes to greenhouse gas emissions. One way to reduce CO2 emissions is to use drop-in fuels. However, when drop-in fuels are used, i.e., higher blends of alternative fuels are added to conventional fuels, solubility problems and precipitation in the fuel can occur. As a result, insolubles in the fuel can clog the fuel filters and interfere with the proper functioning of the injectors. This adversely affects engine performance and increases fuel consumption. These problems are expected to increase with the development of more advanced fuel systems to meet upcoming environmental regulations. This work investigates the composition of the deposits formed inside the injectors of the heavy-duty diesel engine and discusses their formation mechanism. Injectors with internal deposits were collected from field trucks throughout Europe. Similar content, location and structure were found for all the deposits in the studied injectors.

A viable option to reduce global warming related to internal combustion engines is to use renewable fuels, for example methanol. However, the risk of knocking combustion limits the achievable efficiency of SI engines. Hence, most high load operation is run at sub-optimal conditions to suppress knock. Normally the fuel is a limiting factor, however when running on high octane fuels such as methanol, other factors also become important. For example, oil droplets entering the combustion chamber have the possibility to locally impact both temperature and chemical composition. This may create spots with reduced octane number, hence making the engine more prone to knock. Previous research has confirmed a connection between oil droplets in the combustion chamber and knock. Furthermore, previous research has confirmed a connection between oil droplets in the combustion chamber and exhaust particle emissions.

The formation of deposits in the fuel systems of heavy-duty engines, using drop-in fuels, has been reported in recent years. Drop-in fuels are of interest because they allow higher levels of alternative fuels to be blended with conventional fuels that are compatible with today’s engines. The precipitation of insolubles in the drop-in fuel can lead to clogging of fuel filters and internal injector deposits, resulting in increased fuel consumption and engine drivability problems. The possible mechanisms for the formation of the deposits in the fuel system are not yet fully understood. Several explanations such as operating conditions, fuel quality and contamination have been reported. To investigate injector deposit formation, several screening laboratory test methods have been developed to avoid the use of more costly and complex engine testing.

The development of intelligent and networked vehicles has enhanced the demand for precise road information perception. Detailed and accurate road surface information is essential to intelligent driving decisions and annotation of road surface semantics in high-precision maps. As one of the key parameters of road information, road roughness significantly impacts vehicle driving safety and comfort for passengers. To reach a rapid and accurate estimation of road roughness, in this study, we develop a neural network model based on vehicle response data by optimizing a long-short term memory (LSTM) network through the particle swarm algorithm (PSO), which fits non-linear systems and predicts the output of time series data such as road roughness precisely. We establish a feature dataset based on the vehicle response time domain data that can be easily obtained, such as the vehicle wheel center acceleration and pitch rate.

Ammonia is a promising carbon-free alternative fuel for use in combustion systems. The main associated challenges are its relatively low reactivity and high NOx emissions compared to conventional fuels. Therefore, the combustion behaviour of ammonia and ammonia blends still needs to be better understood over a wide range of conditions. To this end, a comprehensive chemical kinetic mechanism C3MechV3.4, which is an update of C3MechV3.3, has been developed for improved predictions of the combustion of ammonia and ammonia blends. C3MechV3.4 has been validated using a wide range of experimental results for pure ammonia and ammonia/hydrogen, ammonia/methanol and ammonia/n-heptane blends. These validations target different data sets including ignition delay times, species profiles measured as a function of time, and/or temperature and laminar flame speeds over a wide range of conditions.

ADAS (Advanced Driver Assistance System) functions can help the driver avoid accidents or mitigate their effect when they occur, and are pre-cursors to full autonomous driving (SAE defined as Level 4+). The main goal of this work is to develop a Model-Based system to actuate the Evasive Maneuver Assist (EMA) function. A typical scenario is the situation in which longitudinal Autonomous Emergency Braking (AEB) is too late and the driver has to adopt an evasive maneuver to avoid an object suddenly appearing on the road ahead. At this time, EMA can help improve the driver’s steering and braking operation in a coordinated way. The vehicle maneuverability and response performance will be enhanced when the driver is facing the collision. The function will additionally let the vehicle steer in a predetermined optimized trajectory based on a yaw rate set point and stabilize the vehicle. The EMA function is introduced with some analysis of benchmarking data.

Due to the lack of masking effect from the engine, noise and vibration from the drivetrain system have become a critical issue in electric vehicles. To address this problem, this study aims at proposing a comprehensive optimization method with respect to multiple internal excitation sources to reduce the noise and vibration of the electric drivetrain system (EDS). A rigid-flexible coupling dynamic (RFCD) model is first proposed by incorporating the elements of electric motors, gear pairs, bearings, shafts, and housing. Based on the proposed model, optimizations concerning multiple internal excitations are carried out by designing the notch width of the stator core and optimizing the modification parameters of the tooth surface along the flank and lead direction. Meanwhile, multiple operating conditions are considered in the optimization of the tooth surface to cover a complete working condition of the EDS.

Aiming at the fault that the speed signal of the automatic transmission output shaft Hall-type speed sensor fluctuates abnormally due to the change of the air gap, the method of fault detection and diagnosis is proposed. Firstly, a limited low-pass filter module was designed according to the characteristics of the fault, and a good filter effect is achieved. Secondly, by comparing the signals before and after filtering, a residual generator is designed, and an adaptive dynamic threshold is designed by analyzing the causes and influencing factors of the residual, which can configure a reliable and effective threshold for the generated residual in real time, which improves the fault identification robustness and effectiveness. Then, a fault debounce method is designed to avoid frequent false alarms of occasional faults. Finally, simulation verification proves the effectiveness of the method.

Modern spark-ignited (SI) engines offer excellent emission reduction when operated with a stoichiometric mixture and a three-way catalytic converter. A challenge with stoichiometric compared to diluted operation is the knock propensity due to the high reactivity of the mixture. This limits the compression ratio, thus reducing engine efficiency and increasing exhaust temperature. The current work evaluated a model of conditions at inlet valve closing (IVC) and top dead center (TDC) for steady state operation. The IVC temperature model is achieved by a cycle-to-cycle resolved residual gas fraction estimator. Due to the potential charge cooling effect from methanol, a method was proposed to determine the fraction of fuel sourced from a wall film. Determining the level of charge cooling is important as it heavily impacts the IVC and TDC temperatures.

Stoichiometric operation of a Port Fueled Injection (PFI) Spark-Ignited (SI) engine with a three-way catalytic converter offers excellent CO2 reduction when run on renewable fuel. The main drawbacks with stoichiometric operation are the increased knock propensity, high exhaust temperature and reduced efficiency. Knock is typically mitigated with a reactive knock controller, with retarded ignition timing whenever knock is detected and the timing then slowly advanced until knock is detected again. This will cause some cycles to operate with non-ideal ignition timing. The current work evaluates the possibility to predict knock using the measured and modelled temperatures at Inlet Valve Closing (IVC) and Top Dead Center (TDC). Feedback effects are studied beyond steady state operation by using induced ignition timing disturbances.

The use of turbochargers with downsized internal combustion engines improves road vehicles’ energy efficiency but introduces additional sound sources of strong acoustic annoyance on the turbocharger’s compressor side. In the present study, direct noise computations (DNC) are carried out on a passenger vehicle turbocharger compressor. The work focuses on assessing the influence of grid parameters on the acoustic predictions, to further advance the maturity of the acoustic modelling of such machines with complex three-dimensional features. The effect of the boundary layer mesh structure, and of the spatial resolution of the mesh, on the simulated acoustic signatures is investigated on detached eddy simulations (DES). Refinements in the core mesh are applied in areas of major acoustic production, to generate cells with sizes proportional to the local Taylor microscale values.

One way to reduce carbon dioxide emissions from the current heavy-duty vehicles fleet is to replace fossil fuel with renewable fuel. This can be done by blending so-called drop-in fuels into the standard diesel fuel. However, problems such as insoluble impurities may arise when the fuels are mixed. These precipitates, known as soft particles, can cause deposits in the fuel system, e.g., injectors and fuel filters, reducing the engine´s performance. The most used drop-in fuel today is biodiesel which, is blended with different concentrations. To better understand how soft particles are formed in the vehicle´s fuel system, the degradation of biodiesel blends in the engine has been investigated. This study explores biodiesel blends´ degradation process by comparing the incoming fuel with the return fuel from a modern diesel engine to investigate how the fuel is affected by this process. The engine was run using different blends of biodiesel fuel.

Heavy-duty transportation accounts for significant part of the greenhouse gas emissions. Currently the most common powertrain for long-haul trucks is compression-ignited engines. In order to reduce the greenhouse gas emissions of these engines, renewable fuels, such as biodiesel can be used. Today biodiesel is used as a drop-in fuel, however when biodiesel is mixed with conventional diesel, soft particles may form. Soft particles have been identified as a mixture of insoluble impurities and degradation products in the fuel. These soft particles can lead to deposits in the injection and fuel filtration system, leading to reduced engine performance. In this paper, zinc-neodecanoate and soft particles from the degradation of biodiesel is studied. In both cases, the emphasis is on soap type contaminants. Zinc-neodecanoate has shown to lead to nozzle fouling, while soft particles from degradation of biodiesel have been found in diesel fuel filters.

Sustainable fuels can help to decrease carbon dioxide emissions in road transportation compared to standard fossil fuels. The most common sustainable fuels used today in heavy-duty applications are biodiesel and hydrogenated vegetable oil (HVO). Biodiesel and HVO are known as drop-in fuels since they are fuels that can be blended with standard diesel. However, due to changes in the chemical properties when the fuels are mixed, solubility problems in terms of precipitates may be formed. These insolubilities can lead to deposits in the fuel system, e.g., blocked fuel filters and internal injector deposits, and thus driveability problems. This study is a part of a project where the goal is to study the processes that cause the formation of deposits inside the injectors in heavy-duty vehicles. The deposits inside the injectors are known as internal diesel injector deposits (IDID).

The ignition dynamics of an ethanol-diesel direct injection compression ignition engine is investigated based on 3D RANS simulations. Experimental results of a previous test campaign on a single-cylinder research engine equipped with two direct injectors are used to validate the CFD model. Four reference engine conditions are considered, including split and overlapped injections of ethanol and diesel at low and high load. Combustion driven by the separate direct injection of pure ethanol and diesel as pilot fuel is simulated with AVL Fire and AVL Tabkin adopting the flamelet generated manifold combustion model. The in-cylinder pressure and apparent rate of heat release traces computed in the simulations are found to be consistent with the corresponding experimental results.

Thermoacoustic engine has been proven to be a promising technology for automotive exhaust waste heat recovery to save fossil fuel and reduce emission thanks to its ability to convert heat into acoustic energy which, hence, can be harvested in useful electrical energy. In this paper, based on the practical thermodynamic parameters of the automotive exhaust gas, including mass flow rate and temperature, two traveling-wave thermoacoustic engines are designed and optimized for the typical heavy-duty and light-duty vehicles, respectively, to extract and reutilize their exhaust waste heat. Firstly, nonlinear thermoacoustic models for each component of a thermoacoustic engine are established in the frequency domain, by which any potential steady operating point of the engine is available.

Alcohols offer high resistance to autoignition which is necessary to attain the required load in heavy duty (HD) spark ignition (SI) engines. Dilution increases thermal efficiency and reduces propensity to autoignition making it an important combustion strategy. Reliable and robust prediction at increased dilution is necessary to support development of high efficiency spark ignition engines and the transition to renewable fuels. A previous experimental study demonstrated 25 bar gross IMEPg for ethanol and methanol at λ=1.4 excess air ratio and over 48% indicated efficiency at λ=1.6 on a single cylinder engine. Based on this dataset, a semi-predictive model (SITurb) was fitted for a range of excess air ratios and engine loads. With the default model, poor accuracy was observed above λ=1.4. Ignition delay was incorrectly predicted at λ=1.6 and λ=1.8.