Engine ECU Research Articles

Deep long short-term memory neural networks as virtual sensors for marine diesel engine NOx prediction at transient conditions

Virtual Sensors based on deep-learning models for predicting the NOx emissions of a Diesel Engine under transient conditions were developed and verified. Raw data from laboratory experimental measurements, under marine transient loading cycles, were used for training and evaluation of the developed models. NOx prediction under transient conditions is often inaccurate by implementing conventional methods since they fail to capture the dynamic behavior of internal combustion engines. The proposed model is based on Long Short-Term Memory (LSTM) Networks. A Deep Feed-forward Neural Network (DFNN) was also developed to validate the LSTM. The LSTM input is a time sequence of past measurements of the inputs while the DFNN only uses the most recent measurements. The Bayesian Hyberband Optimization (BOHB) algorithm determined the structure and parameters of each network. Each model uses the same inputs and is directly derived from the engine ECU. The LSTM validation showed that the model can generalize and accurately predict the NOx emissions under transient loading compared to the DFNN.

Dynamic Analysis of Common Rail Injection System for Diesel Engine based on AMESim/Simulink Co-Simulation

In common rail diesel engine, the characteristics of fuel injection is concerned both the structure parameters of injector and control strategy of control system (ECU). Thus, fuel injection system, engine performance system and ECU constitute an interactive feedback system. In this paper, first of all, using the AMESim/Simulink co-model for the common rail diesel engine prepared in the previous work, simulation and verification are performed in the AMESim/Simulink co-model based on the initial values of the diesel engine used in the case. Then it is done Simulation and Validation on the AMESim/Simulink co-mode. The engine dynamometer and injection rate test rig makes it possible to take measurements of the engine performance and injection rate in order to compare with the modelling results. This allows the obtained co-model to be validated. Validated model is very effective in describing and analyzing the transient behavior connections among injectors, high-pressure pump, and controller and engine parameters. This co-model was then used to analyze the effects of various structural parameters of injector on the response characteristics of the fuel injection system and engine operation.

Automation of Engine ECU Calibration through CAN with Python Machine Learning Algorithms

Engine ECU subsume the numerous control functions of electrical systems in the vehicle based on various sensors inputs and the control parameters present inside ECU such as maps, multiplication factors, constants and so on. These control parameters need to be calibrated effectively for better performance as well as to meet stringent emission norms. Since, the most efficient way of calibration is through CAN by means of XCP/CCP protocol, this process involves logging and processing of the ECU data to estimate the appropriate values followed by downloading the modified values to the ECU manually. Even though from theoretical calculations it is possible to estimate the approximate parameter values, these values need to be validated in engine test beds or on road and are fine tuned to attain the optimum results by repeating the same trial under same test conditions numerous times. After each trial, the data is analyzed and new set of data is determined which is downloaded to ECU before the next trial. This process is carried out until optimum results are achieved which is time consuming. In this paper, a new approach has been explained which will eliminate the human interference during the trials and speeds up the process of establishing the master slave communication between PC and ECU through any CAN transceiver hardware with the help of python, and its machine learning algorithms to carry out the analysis tasks between successive trials which develops regression models for predicting the parameter values based on the previous trials with in a shorter period of time increasing the human potential of calibration.

ESTIMATION OF THE ACCURACY OF FUEL CONSUMPTION MEASUREMENT BASED ON DATA FROM THE ON-BOARD CAN-BUS

Determining the dependence of actual fuel consumption on operating conditions is the most important task in vehicle transport, since the cost of hydrocarbons is a significant part of the cost of transportation. Currently, there are many ways to experimentally measure fuel efficiency: using external flow meters installed in the fuel line, equipping the tank with additional level sensors, using information from the electronic engine control unit. The latter approach is of greatest interest, since it requires minimal financial investment, is easy to install, and there are no significant changes to the design. The question arises about the measurement accuracy according to this method, since the ECU calculates the fuel consumption according to a mathematical model, calibrated and programmed at the factory. This article offers a solution to this problem. At the first stage, NNSTU specialists have developed and installed on a vehicle equipment for measuring fuel consumption simultaneously by two methods – using an external high-precision Kistler Dfl3x-5bar flow meter, installed in the fuel line, and on the basis of data from the engine ECU, obtained using information CAN-bus. At the second stage, tests were carried out (moving in a steady state and the driving cycle) after this, using results of tests, we have concluded a numerical difference in the methods for determining fuel consumption. At the end of the article, tabular results of comparison of measurements and graphs of the determined parameters are demonstrated.

Estimation of the accuracy of fuel consumption measurement based on data from the on-board CAN-bus

The legal requirements for vehicle exhaust emissions are becoming increasingly stringent. Emissions are directly related to fuel consumption. Therefore, reducing fuel consumption is a very urgent task. To solve this problem, it is necessary to assess the influence of various factors on the actual fuel consumption. To determine the fuel consumption, there are a large number of different methods using external flow meters installed in the fuel line, installing additional level sensors in the fuel tank. A very simple and convenient method is based on the use of information from the electronic engine control unit. This method requires minimal financial investment and does not require significant structural changes. The question arises about the measurement accuracy according to this method, since the ECU calculates the fuel consumption according to a mathematical model, calibrated and programmed at the factory. To assess the accuracy of this method, a set of tests was carried out. The tests were performed at constant vehicle speed and according to the standard driving cycle. During the tests, the fuel consumption was recorded using a high-precision calibrated flow meter installed in the fuel line and on the basis of data from the engine ECU. Based on the test results, a conclusion was made about the size of the discrepancy in the methods for determining the fuel consumption.

Systematic Methodology for Analysis and Control of Real Driving Emission for Heavy Duty Vehicles Using Virtual Test Bed

<div class="section abstract"><div class="htmlview paragraph">Development of future efficient and cleaner heavy duty engines are no longer limited to laboratory development under standard conditions. In order to address the global issues like climate change and poor air quality in its true sense, future advanced and existing heavy duty diesel engines should also be demonstrating emission conformity compliance as per legislations under real driving conditions using PEMS testing. In India starting from Apr 2023, heavy duty vehicles would be tested for in-service conformity and presently they are under monitoring phase. With the introduction of RDE (Real Driving Emission) the effort, cost and time requirements could be tremendous in order to meet conformity compliance over real driving conditions including the range of ambient conditions for the said period as per the norms. Hence, it would be of great importance if we can utilize virtual test bed (VTB) to off-load several initial field testing and calibration efforts by virtually running the vehicles over simulated cycles with various operating conditions that includes simulating payload, traffic, ambient condition change and predicting emissions and their conformity for these generated routes and operating conditions. However, this would indeed require a good correlation between the actual engine and vehicle behavior and the plant model. In this work an attempt has been made to understand how VTB can be used for this purpose. The PEMS testing data for the first heavy duty vehicle with GVW above 12 ton was used for one level of model validation were reasonable matching of various trip parameters, characteristics trends and emissions were observed for running the simulated cycle in VTB. Later we performed a parametric variations including payload, traffic and ambient conditions using a second vehicle with GVW above 12 ton category with different tonnage and application to understand their impact on trip characteristics, emissions and conformity factors and finally the results for one parameter was validated by conducting a PEMS test. With this approach, in contrast to conventional method we utilized virtual test bed with integration of engine ECU, plant model (engine +after treatment) and PUMA test system to investigate the possibility of generating drive cycle, performing parametric variations, simulating extreme conditions and understanding their impact on trip characteristics, emissions, and conformity factor moreover such findings will definitely help in understanding the scope of calibration work required to comply with real driving emission norms.</div></div>

Vehicle Health Monitoring System Using CAN

Modern vehicles are integrated with more number of ECUs and sensors which make them smarter in terms of engine decisions, performance, fuel efficiency, security and stability. ECUs also provide useful data in the form of diagnostic codes via OBDII (on-board diagnostics) protocol, these codes determine a specific issue in the vehicular system which is used by service technicians to address the issues occurred. This paper deals with the exploring the potential of OBDII codes from engine ECU and compute the data on an Android device to monitor the vehicle health status in real-time. The whole idea is to provide portable diagnosis solutions and integration of such system into in-vehicle entertainment like Google’s Android Auto and Apple’s CarPlay which can be one of key component for V2X and IOVs (Internet of Vehicles).

Development on ECU for the 32-bit High Pressure Common-rail Diesel Engine Based on AUTOSAR ECU Software

In order to facilitate the development of the diesel engine ECU, based on AUTOSAR ECU software architecture and MPC554 microcontroller, the ECU basic platform which includes of diesel engine ECU hardware and ECU basic software library is established, as well as the overall common rail ECU is achieved. This ECU basic platform of the diesel engine provides reliable hardware system, drive module interface, two-stage injector configuration control waveform output and OBD system. It is convenient to ECU development which improves the developing efficiency, shortens the developing cycle and has a high practical value.

Development of two-phase interleaved synchronous rectification DC-DC for commercial vehicle engine ECU

The commercial vehicle diesel engine ECU and the electronically controlled common rail system power supply developed in this paper adopt the two-phase BUCK converter as the main loop. Considering the problems of large current fluctuation, high continuum diode loss and possible non-uniform current in parallel connection of traditional BUCK converter, a control strategy of two-phase staggered parallel synchronous rectifier BUCK converter is proposed. The relationship between current fluctuation and inductance, switching frequency and duty cycle of different operating modes is theoretically analyzed, and the current fluctuation of traditional BUCK converter is compared. The prototype test results show that the converter structure and control strategy effectively reduce the output current fluctuation and improve the system power density.

가솔린엔진제어기 공기량결정함수의 시뮬레이션 환경 구축 및 파라미터 매핑 자동최적화

가솔린엔진제어기 공기량결정함수의 시뮬레이션 환경 구축 및 파라미터 매핑 자동최적화

Research on Key Technologies of Integrated Teaching Platform for Automotive Engine ECU Control System

In order to ensure that high-skilled personnel can better grasp the internal structure of ECU control system of automobile engine and improve the maintenance skills, this paper analyses the teaching platform from two aspects of hardware and software; the former including the introduction of basic characteristics and platform building, and introduce the functions and components of the integrated teaching platform for control system of automobile engine. Through the analysis, the maintenance personnel can have a better understanding of the fault diagnosis and elimination of the ECU of the automobile engine, and provide a good reference basis.

Performance Characteristics Comparison of CNG Port and CNG Direct Injection in Spark Ignition Engine

A comparative performance analysis is being carried out on a four cylinder, four stroke cycle, spark ignition engine having displacement volume 1297cc. The cylinder head of original gasoline based engine was modified by drilling holes from upper surfaces of head to individual combustion chamber to convert the engine in a CNG direct injection engine. The CNG port injection (CNG-PI) system and CNG direct injection (CNG-DI) system were incorporated with the single engine. The engine was retrofitted to run on both CNG-PI and CNG-DI system alternately with common CNG tank and other engine loading and measurement system. The engine was equipped with electrical dynamometer having rheostat type loading. The CNG direct injection system was incorporated with various sensors and engine ECU. The operating parameters can be obtained on computer screen by loading the computer with engine through switch box. The engine was run over the speed range of 1000 rpm to 3000 rpm with incremental speed of 300 rpm. The performance parameters were calculated from observations and recorded for both CNG-PI and CNG-DI system. The experimental investigation exhibits that, the average 7-8% reduction in BSFC while the engine was running with CNG-DI system as compared to that of CNG-PI system. Also the engine produced 8-9% higher brake torque and hence higher brake power. The engine gives 6-7% higher brake thermal efficiency with CNG-DI system as compared to CNG-PI system.

Research and Analysis of Engine Combustion Based on Independent Test System

In this paper, we use the self-developed electronic control engine ECU calibration and automatic matching system, to analyze the cylinder combustion dynamics. With the first and secondary filter, collect the combustion pressure of eight-cylinder gasoline engine and crank angle variation over the entire cycle of data; measure and analyze the main parameters of the relevant data within each cylinder; draw a series of reasonable and effective variation of fuel consumption data graphs, etc. It provides an accurate basis for the best ignition timing and fuel economy optimization of eight-cylinder engine.

Efficient combustion parameter prediction and performance optimization for a diesel engine with a low throughput combustion model

In this work, an efficient implementation of a zero-dimensional model is described for the estimation of key engine parameters for combustion control in compression-ignition engines. The direct problems of the estimation of the angle of 50% of fuel mass fraction burnt (MFB50) and of the Indicated Mean Effective Pressure (IMEP) are addressed as well as the inverse problems of determining optimal start of injection (SOI) timing for target values of MFB50 and IMEP. The main focus is on the computational cost of the algorithms proposed, designed in order to keep the number of operations as low as possible without compromising the applicability of the methods to different engine configurations and operation points. Execution time of the order of few milliseconds is achieved for parameter prediction and of the order of one tenth of a second for the optimization problems, such that an implementation in engine ECU for model-based control purposes can be envisaged.

Development of an electrically driven diesel fuel pump for commercial vehicles using HILS

Common-rail fuel injection systems on modern diesel engines have fuel pumps that are mechanically driven by the crankshaft. The pumps actually house two pumping elements: a low pressure element and a high pressure element. Part of the fuel compressed by the low pressure element returns to the tank in the process of maintaining the pressure in the common-rail at adequate levels for injection. Since the returning fuel represents pumping energy loss, fuel economy improves if the returned fuel can be reduced by using a properly controlled electrical fuel pump. As the first step in developing an electrical fuel pump the fuel supply system on a six-liter diesel engine was modeled with AMESim to analyze the workload and the fuel feed rate of the injection pump; the results served as the basis for selecting a suitable servo-motor to drive the pump. A motor controller was built using a DSP and a program which controls the common-rail pressure using a proportional control method based on the target fuel pressure information from the engine ECU. Test equipment was also fabricated to evaluate the performance of the electrically driven fuel pump. In a simulated vehicle test during the first 200 seconds of the NEDC driving schedule, the electric pump satisfied the fuel pressure and flow demand of the engine, consuming only 48.3% of the energy that an engine-driven pump would require.

Fault Manager of Diagnostic System in the Gasoline Engine ECU

A structure of fault manager is established. Under the premise of combining relevant standards, the five main modules of the fault manager effectively integrated and handled fault information. This fault manager is implemented in a home-ECU for its upcoming production.

Part-load performance and emissions of a spark ignition engine fueled with RON95 and RON97 gasoline: Technical viewpoint on Malaysia’s fuel price debate

Due to world crude oil price hike in the recent years, many countries have experienced increase in gasoline price. In Malaysia, where gasoline are sold in two grades; RON95 and RON97, and fuel price are regulated by the government, gasoline price have been gradually increased since 2009. Price rise for RON97 is more significant. By 2014, its per liter price is 38% more than that of RON95. This has resulted in escalated dissatisfaction among the mass. People argued they were denied from using a better fuel (RON97). In order to evaluate the claim, there is a need to investigate engine response to these two gasoline grades. The effect of gasoline RON95 and RON97 on performance and exhaust emissions in spark ignition engine was investigated on a representative engine: 1.6L, 4-cylinder Mitsubishi 4G92 engine with CR 11:1. The engine was run at constant speed between 1500 and 3500rpm with 500rpm increment at various part-load conditions. The original engine ECU, a hydraulic dynamometer and control, a combustion analyzer and an exhaust gas analyzer were used to determine engine performance, cylinder pressure and emissions. Results showed that RON95 produced higher engine performance for all part-load conditions within the speed range. RON95 produced on average 4.4% higher brake torque, brake power, brake mean effective pressure as compared to RON97. The difference in engine performance was more significant at higher engine speed and loads. Cylinder pressure and ROHR were evaluated and correlated with engine output. With RON95, the engine produces 2.3% higher fuel conversion efficiency on average but RON97 was advantageous with 2.3% lower brake specific fuel consumption throughout all load condition. In terms of exhaust emissions, RON95 produced 7.7% lower NOx emission but higher CO2, CO and HC emissions by 7.9%, 36.9% and 20.3% respectively. Higher octane rating of gasoline may not necessarily beneficial on engine power, fuel economy and emissions of polluting gases. Even though there is some advantage using RON97 in terms of emission reduction of CO2, CO and HC, the 38% higher price and higher NOx emission is more expensive in the long run. Therefore using RON95 is economically better and environmentally friendlier. The findings provide some techno-economic evaluation on the fuel price debate that surround the Malaysia’s population in the recent years. The increased of fuel price may have limited their ability to use higher octane gasoline but it did not negatively affecting the users as they perceive.

상용차용 디젤엔진의 Engine ECU_ILS 시스템 개발

The automobile industry requires technological innovations to reduce fuel consumption with the public interest in environmental conservation in recent years. Thus, the hybrid system is applied not only to passenger cars but also commercial vehicles. The purpose of this paper is to develop engine ECU_ILS to develop commercial hybrid vehicles. In order to develop the engine and vehicle, the dynamometer and exhaust gas analyzer is needed. However, a lot of time and cost are required. In contrast, the model-based development environment that can be applied to a variety of test conditions can reduce development time. Therefore, a HILS system environment that can consider the behavior of actual vehicles for evaluation of the control logic, fuel consumption and exhaust gas is required. This engine ECU_ILS system was developed in this study, can analyze parameter such as the fuel injection rate, fuel injection time, fuel consumption and exhaust gas like the actual vehicle test using map data. Also, this system is expected to be able to analyze the characteristic of vehicle behavior and the development of peripheral device in relation to engine and vehicles. This HILS system can be used to develop control strategies of commercial hybrid vehicle systems in the future.

Hardware Design of Nature Gas Engine ECU Based on Single Chip

Gas supply method plays an important role in engine performance for natural gas engines (NGE). In order to overcome the low charging efficiency of pre-mixed fuel gas supply method the multi-point injection (MPI) gas engine controller hardware system is designed in this paper. For the nature gas engine controller the main controller is selected as the microcontroller MC9S12XSl28, then the circuitry for semiconductor piezoelectric air pressure sensor, potentiometer throttle position, thermistor temperature sensor, a magnetic-electric sensor for detecting the top dead center (TDC) and simulating engine speed are designed. What have done above is the base of software design. Key Words: nature gas engine; controller; hardware design

HILS기반 상용차 디젤엔진용 연료펌프의 전기구동 시스템 적용에 관한 연구

Fuel injection pressure has steadily increased in diesel engines for the purpose of improving fuel efficiency and cleaning exhaust gas, but it has now reached a point, where the cost for higher pressure does not warrant additional gains. Common rail systems on modern diesel engines have fuel pumps that are mechanically driven by crankshaft. The pumps actually house two pumping module inside: a low pressure pump component and a high pressure pump component. Part of the fuel compressed by the low pressure component returns to the tank in the process of maintaining the pressure in the common rail. Since the returning fuel represents pumping loss, fuel economy improves if the returned fuel can be eliminated by using a properly controled electrical fuel pump. As the first step in developing an electrical fuel pump the fuel supply system on a 6 liter diesel engine was modeled with AMESim to analyze the workload and the fuel feed rate of the injection pump, and the results served as basis for selecting a suitable servo motor and a reducer to drive the pump. A motor controller was built using a DSP and a program which controls the common rail pressure using a proportional control method based on the target fuel pressure information from the engine ECU. A test rig to evaluate performance of the fuel pump is implemented and used to show that the newly developed electrically driven fuel pump can satisfy the fuel flow demand of the engine under various operating conditions when the rotational speed of the pump is adequately controlled.