Krisada Wannatong

Simulation algorithm for piston ring dynamics

Abstract Piston ring dynamics play important roles on the lubricant characteristic of reciprocating engines which lead to the consequences of engine wear and vast amount of lubricating oil consumption. Due to the complexity of motion, a study of motions and effects of the piston ring is mostly conducted in a simulation program. This paper shows a theoretical work and a new simulation algorithm of the 3D piston ring motions. The simulation concept is to determine the positions of the piston ring, which are the solutions of the Newton and Euler equations. Well known models like mixed lubrication model, asperity contact model, and blow-by/blow-back flow model were used in this study. The new simulation algorithm consists of four processes: construction of calculation nodes, use of finite different method, determination of the non-linear equation system, use of parallel computational technique. Two sets of the experimental studies were conducted for simulation verification. First, the gas flows through the static rectangular piston ring pack. Second, the blow-by/blow-back gas flow through the piston ring pack of a single cylinder diesel engine. The good agreement between the experimental and simulation results indicates that the developed simulation program can illustrate the piston ring motions and blow-by/blow-back flow. Since there is no algebraic equation in the ring profile, the presented simulation technique is suitable for the complicated surface of the piston and piston rings.

Robust Common-Rail Pressure Control for a Diesel-Dual-Fuel Engine Using QFT-Based Controller

Despite promising future, the diesel-dual-fuel engine, with diesel as pilot and natural gas as main, abounds with challenges from high NOx emission and knock especially at high speed and low load. To cope with these challenges, variation of common-rail pressure provides another desirable degree of freedom. Nevertheless, crippling with complicated dynamics, pressure wave inside the transporting rail, disturbance from varying of injections, engine speed variation, and actuator limitation, common-rail pressure control has relied on the simple PID to deliver only marginally satisfactory result. Some attempts to achieve better control have resulted in either too complicated or not too robust control system. We devise a controller from the quantitative feedback theory. Besides being able to quantitatively enforce specifications such as tracking, plant input and output disturbance rejections, and stability margin, the controller is designed from a simple model, whose parameters are allowed to be uncertain, hence robustness. The resulting controller has low order and is readily implementable. Experiment with a common-rail system in a Ricardo Hydra engine, modified to run dual fuel, shows the controller’s effectiveness over the PID.

Air/Fuel Ratio Control in Diesel-Dual-Fuel Engine by Varying Throttle, EGR Valve, and Total Fuel

From our experiences in converting diesel engine into diesel-dual-fuel engine with natural gas as primary fuel, accurate air/fuel ratio control is vital to the high engine performance, good vehicle drivability, and low emissions. Two components enter in calculating the air/fuel ratio, namely, the amount of fresh air and the amount of diesel and natural gas. Throttle and EGR valve are two actuators directly affect the amount of air, and the desired total fuel determines how much fuel should be injected at an instance. As opposed to inactive, fully opened throttle in typical diesel engine, the throttle in diesel-dual-fuel engine is regulated to cover wider range of desired air/fuel ratio. As a result, the problem of controlling the amount of air in diesel-dual-fuel engine becomes that of multi variables in which both throttle and EGR valve are involved. We present a novel algorithm that breaks the multi-variable control problem into two single-variable problems. The throttle and EGR valve are regulated one at a time as determined by a switch-and-hold logic that optimizes the throttle opening to reduce pumping loss. An algorithm is also proposed to prevent the throttle from fully closed. Because the fuel path is much faster than the air path, including in the algorithm is the adjustment of the desired total fuel according to the air/fuel ratio tracking error. We found that adjusting the total fuel within a limit helps improve the transient response. Experiments were performed on an engine test bed and a pickup truck; both engines were modified from four cylindrical diesel engines to run diesel-dual-fuel, where the natural gas is injected at each intake port. The test bed experiment showed that the air/fuel ratio was accurately regulated with widest range possible. The pickup truck was commanded to follow the new European driving cycle. The results showed that the throttle opening was optimized at all time. INTRODUCTION Due to the steeper price of diesel and gasoline, alternate and lower price fuel such as compressed natural gas (CNG) has become more attractive to the users. Recently, CNG has been successfully used in spark-ignition (SI) engines such as those running bi-fuels. With spark-assisted ignition, using CNG does not post much practical problems. Compared to the spark-ignition engine of the same size, compressed-ignition (CI) engine has higher fuel efficiency and performance, which makes it more preferable for heavy-duty trucks. Due to physical limitations, CI engines cannot run with CNG alone. A customary modification is to inject CNG in the intake ports and to inject a smaller amount of diesel directly in the cylinder to initiate the combustion. This so-called diesel-dualPage 2 of 16 fuel (DDF) engine has a prospect of converting existing diesel trucks to run partly on CNG. Our experiences have shown that obtaining from CNG 60-70 percents by energy is possible. However, DDF engine controls become more challenging both in air and fuel paths because the engine now uses two fuels at the same time. CNG is injected at the intake ports similar to the typical SI engine, while diesel is injected in the cylinder as is done in the CI engine. DDF engine therefore has combined characteristics of SI and CI engines. In CI engine, throttle is usually inactive and fully opened for maximum fresh-air intake, and exhausted gas recirculated (EGR) valve and variable geometry turbine (VGT) are actuated to bring portion of the exhausted gas back to the cylinder to reduce temperature and therefore reduce NOx. In SI engine, throttle is usually very active in achieving lambda one, while EGR valve and VGT are normally not required due to lower-temperature combustion chamber. In DDF engine, throttle, EGR valve, and VGT are all actively actuated. For low emissions, the engine must follow a desired air/fuel ratio closely. This desired air/fuel ratio, found during calibration, has wider range than those of SI and DI engines. Since throttle is not fully opened as in CI engine, to reduce pumping loss, the throttle must be opened as wide as possible, while still follows a desired air/fuel ratio. Existing literature on air/fuel ratio control of DDF engine is either non-existent or unavailable. Since in SI engine, throttle is normally driven by drivers pedal, and its air path is not actively controlled, we surveyed instead the air path control of the CI engine. The first group formulated the air path control problem as multivariable problem and solved using H∞ -based optimization algorithms. Ref. [1] applied the H∞ mixed synthesis and a controller based on internal model principle to control the air/fuel ratio of diesel engine by actuating only the EGR valve. Ref. [2]-[3] formulated a multivariable problem in controlling the air path of a turbocharged diesel engine, where the plant has two inputs, EGR position and variable geometry turbine (VGT) position, and two outputs, mass air flow (MAF) and manifold pressure (MAP). Their control techniques are two-degree-of-freedom H∞ loop-shaping. Both the H∞ mixed synthesis and H∞ loop-shaping techniques are covered excellently in [4]. An extension to control design based on linear parameter varying (LPV) model can be found in [5]. Additional optimization algorithms that are closely related to the H∞ algorithm are as follows. Ref. [6]-[7] formulated a gain-scheduling H∞ -based control. The scheduled variables are engine speed, VGT position, and MAP. The plant model is in quasi linear parameter varying (LPV) form, with two inputs as EGR and VGT positions and two outputs as MAF and MAP. Ref. [8] presented a linear matrix inequality (LMI) based control, where a nonlinear plant having EGR valve and VGT openings as inputs and EGR flow and air/fuel ratio as outputs was formulated. Around each operating point, the plant was linearized and put in polytopic form. Ref. [9] applied LQG to controlling air/fuel ratio and burned gas fraction by regulating EGR valve and VGT. Several decentralized controls were presented in [10] to regulate MAP and MAF by actuating the EGR valve and VGT. The second group designed air path controller from nonlinear model using several nonlinear control techniques. Ref. [11] applied the dynamic feedback linearization to control of the turbocharged diesel engine air path. EGR and VGT flows were used as plant inputs. For plant outputs, instead of using EGR fraction and air/fuel ratio or MAP and MAF, they used MAP and a modified signal, a function of MAP, exhaust manifold pressure, and compressor power, as their outputs. The reasons for this choice are signal accessibility in commercial vehicle and avoiding non-minimum-phase plant model. Ref. [12] treated fuel amount and VGT position as inputs and Page 3 of 16 air/fuel ratio as output. The control design was based on the backstepping control. Ref. [13] applied sliding mode control in controlling the air/fuel ratio and EGR rate using EGR valve and VGT positions. Ref. [14] used constructive Lyapunov control in regulating the air/fuel ratio and EGR fraction by actuating the EGR valve and VGT. A simple linear control using observer can be seen in [15], which presented an output-feedback PID control, having EGR valve and VGT openings as plant inputs and MAP and EGR fraction as plant outputs. A Luenberger-type observer was used to estimate EGR flow. The third group used favorable properties of model predictive control in dealing with input and output constraints. Ref. [16] applied model predictive control to control MAF and MAP having EGR valve and VGT as controlled inputs and engine speed and load as disturbance inputs. Active set strategy was used to reduce computational time, which makes the algorithm implementable in real time. Ref. [17] used lookup tables to obtain set points for air/fuel ratio and the amount of recirculated exhaust gas from engine speed and fueling rate. Set points for MAP, exhaust manifold pressure, and compressor power were then determined from the two set points via formulas. The EGR valve and VGT were commanded to track the set points. A nonlinear model predictive control was used for tracking the set points and for constraining the EGR valve and VGT openings to physical limits. Since the amount of air inside of the cylinders directly affects the combustion. Several researchers have designed air estimators and included them in their algorithms. In real-time, Ref. [18] used recursive least square algorithm to estimate air/fuel ratio in each cylinder using a single UEGO sensor placed at the exhaust pipe. The estimated air/fuel ratio was then used in an optimal control algorithm to control individual fuel injector to obtain desired air/fuel ratio in each cylinder. Ref. [19] used Kalman filter in estimating the air/fuel ratio in each cylinder from a single air/fuel ratio measurement at turbine downstream. Not many work discussed pumping loss. In [20], the VGT was actuated together with the EGR valve to control oxygen-to-fuel ratio and EGR fraction of diesel engine. Both actuators were designed to minimize pumping loss, which was defined to be the difference between exhaust manifold and intake manifold pressures. Also not many work actuated the fuel injector, in addition to the EGR valve and VGT, to control air/fuel ratio. Ref. [21] used fuzzy logic control in adjusting diesel fuel injectors pulse width to track desired air/fuel ratio, while EGR valve and VGT were commanded to follow pre-specified maps. In this paper, the plant has three inputs, which are total fuel, throttle opening, and EGR valve opening, and one output, which is the air/fuel ratio. A novel switch-and-hold logic is designed to actuate either throttle or EGR valve, one at a time, by opening the throttle as wide as possible at all time to reduce pumping loss, while still achieving the air/fuel ratio set point. This reduces the multivariable problem to two single variable problems, which can be controlled by any advanced control algorithms suitable for

Quantitative Feedback Control of Air Path in Diesel-Dual-Fuel Engine

In this paper, we investigate a multivariable control of air path of a diesel-dual-fuel (DDF) engine. The engine is modified from a CI engine by injecting CNG in intake ports. The engine uses CNG as its primary fuel and diesel as its secondary fuel, mainly for initiation of combustion. The modification is economically attractive because CNG has lower price than diesel and the modification cost is minimal. However, for DDF engine, control of the air path becomes more difficult because the engine now has combined characteristics of the CI and the SI engines. The combined characteristics come from the fact that diesel is still directly injected into cylinders (CI engine) while CNG is injected at the intake ports (SI engine.) In pure CI engine, throttle is normally fully opened for maximum air intake, while EGR valve is actively actuated to obtain low emissions. In pure SI engine, however, throttle is an active actuator, driven by pedal. The air path control of the DDF engine, therefore, is a multivariable problem, where both throttle and EGR valve are actively actuated to obtain desired outputs. Two outputs that are of particular interest are mass air flow (MAF) and manifold air pressure (MAP). They are directly available from existing vehicle sensors and are good indicators of the air path characteristic. Their desired values are normally obtained as fixed maps during engine calibration. We formulate a control problem, having throttle and EGR valve openings as inputs and MAF and MAP as outputs. Because of high level of input-to-output interactions, a fully multivariable controller is preferred over the decentralized control. A multivariable control based on the quantitative feedback theory (QFT) is designed and implemented. The QFT-based controller is attractive because its robustness amount can be quantified. Control design plant model is allowed to have uncertainty within a boundary called plant template. For all uncertainty within a plant template, a controller and a prefilter are designed from loop shaping to enforce several frequency-domain specifications such as tracking, stability, disturbance and noise rejections, and control effort limitation. A 2KDFTV Toyota CI engine is modified as a DDF engine and is installed in a test cell with an engine dynamometer. A system identification of the air path is performed to obtain a control design model. Both simulation and experimental results on the engine show the effectiveness of the proposed control system in tracking step changes of desired MAF and MAP. INTRODUCTION In countries that can produce its own natural gas, CNG may have lower price than diesel. CNG has long been used as an alternative fuel for the SI engine. With minor modification to the engine, CNG can be injected in the intake ports and be used fully in place of gasoline. This convenience comes from the fact that there exist spark Page 2 of 21 plugs to help ignite the CNG. In CI engine, however, the CNG cannot be used fully in place of diesel because the compression ratio is not high enough to ignite the CNG. To use CNG with CI engine, unless major engine modification is done, diesel is still needed in a smaller amount to initiate the combustion. CNG can be used as primary fuel and can be injected at the intake ports. The so-called diesel-dual-fuel (DDF) engine concept can reduce total fuel cost while still obtain good efficiency and performance of the CI engine. The challenge comes from having two fuels in use at the same time. CNG is injected at the intake ports, which is mixed with fresh air before going into the combustion chamber. This SI engine-like behavior requires precise air/fuel ratio regulation. Diesel is also injected directly into the combustion chamber, where EGR amount must be carefully controlled. While steady-state mappings can be performed during engine calibration, the engine suffers from poor performance during transients. During sudden load or speed changes, fresh air normally lags due to delays from turbocharger and compressible property of air. This emphasizes the necessity to have high-performance control system of the air path. Existing literature on air path control of DDF engine is very limited. Ref. [1] presents an algorithm in choosing to actuate either throttle or EGR valve to regulate air/fuel ratio, while pumping loss is minimized. Ref. [2] adds the effect on adjusting the total fuel amount. In CI engine, normally EGR valve and variable geometry turbine (VGT) are actuated to control MAF and MAP. Existing algorithms in use can be divided into three groups, according to type of the plant model used in control design. The first group is designed based on two-by-two matrix, containing plant transfer functions. Existing algorithms are as follows: H∞ mixed synthesis [3], two-degree-of-freedom H∞ loop-shaping [4], gain-scheduling H∞ based control [5], linear matrix inequality (LMI) control [6], linear quadratic Gaussian (LQG) [7], decentralized control [8], model predictive control [9]. The second group is designed based on linear or nonlinear state-space model. Existing algorithms are as follows: dynamic feedback linearization [10], backstepping control [11], sliding mode control [12], constructive Lyapunov control [13], output-feedback PID control [14], nonlinear model predictive control [15]. The third group, probably the least popular, is designed based on unknown plant model. Existing algorithm is fuzzy logic control [16]. In this paper, we present a control system based on quantitative feedback theory (QFT), originated around fifty years ago by I. Horowitz [17]. Although QFT was originated decades ago, its applications have just been accelerated in the late 80s and early 90s due to the advent of the QFT computer-aided-design (CAD) packages [18]-[20]. QFT design is performed on the Nichols chart. Three main steps are as follows. First, time-domain specifications are converted to frequency domain. Second, frequency-domain specifications are converted to bounds on the Nichols chart. Third, loop shaping is performed to shape the open-loop plot on the Nichols chart. After the open-loop plot is shaped to satisfy all specification bounds, time-domain specifications are met. The authors are not aware of any existing literature that applies QFT to the air-path control. Some advantages over existing techniques mentioned above are as follows: Page 3 of 21 • QFT is designed on a set of plant model, called plant template, which contains all possible plants. The designer can specify any possible plant model variation, for example, different plant model for different speed and load. At the end, the resulting controller will ensure that all specifications are met for all plant variation. • Various time-domain specifications can be set. They are tracking, disturbance and noise rejection, control effort limitation, and model matching. • Specifications are converted to graphical bounds on the Nichols chart. Open-loop plot is shaped to satisfy all bounds. The designer directly sees the amount of over designs and can reduce them, resulting in the leastconservative controller. A 2KD-FTV Toyota CI engine is modified as a DDF engine and is installed in a test cell with an engine dynamometer. A two-by-two transfer function matrix, relating throttle position sensor (TPS) readings and EGR valve opening to MAF and MAP, is found from black-box system identification. Multivariable QFT-based controller and prefilter are designed and tested in simulation and experiment. MAF and MAP can follow their set points closely. The air-path control system benefits from the advantages of the QFT technique mentioned above. The paper is organized as follows. Next section presents system identification of the air path, followed by control system design section. Results are given in simulation results and experimental results sections. The paper is closed with conclusions, references, contact information, and acknowledgments. AIR-PATH SYSTEM IDENTIFICATION Compressor Turbine Throttle EGR Valve

Sliding Mode Control of Air Path in Diesel-Dual-Fuel Engine

In diesel-dual-fuel engine, CNG is injected at the intake ports and diesel fuel is injected at the cylinders. As a result of using CNG as main fuel, smaller amount of diesel is used mainly for ignition, resulting in lower fuel cost. However, stricter air path control is required because the engine now operates partly as a port fuel injection engine and partly as a diesel engine. As is evident from engine calibration, desired MAP and MAF have more abrupt change with wider range than those of diesel engine. In typical commercial truck, MAP and MAF are controlled separately using traditional controller such as PID with marginal control performance. Recently, more researchers have combined the control of MAP and MAF together as multivariable problem because both quantities reflect the behavior of the air path. In this paper, multivariable sliding mode control (SMC) is implemented in two-approaches, a model-reference-based and an integrator-augmented based. The diesel-dual-fuel engine was converted from a diesel engine and used in the engine test bed. Throttle and EGR valve were actuated to regulate MAP and MAF at their respective set points. Experimental results at an engine speed of 2000 rpm and 20% pedal showed that the two proposed algorithms delivered good tracking performance with fast action. The MAP and MAF responses were able to track their desired values with 2.5 seconds settling time and less than 10% overshoot. The integrator-augmented SMC had more response accuracy than the model-reference SMC but with more chattering. INTRODUCTION In countries where compressed natural gas (CNG) has lower price than diesel, there are vast interest in converting existing diesel light-duty trucks into those using CNG. To use CNG with compressed ignition (CI) engine, unless major engine modification is done, diesel is still needed in a smaller amount to initiate the combustion. CNG can be used as primary fuel and can be injected at intake ports. The so-called diesel-dual-fuel (DDF) engine concept can reduce total fuel cost while still obtain good performance of the CI engine. The challenge comes from having two fuels in use at the same time. CNG is injected at the intake ports, which is mixed with fresh air before going into the combustion chamber. This spark-ignition (SI) engine-like behavior requires precise air path control. Diesel is also injected directly into the combustion chamber, where exhausted gas recirculation (EGR) amount must be carefully controlled. The more stringent performance requirement of the air path requires a high-performance control system. Majority of the existing literature in air-path control is on the SI engine. In SI engine, the fuel injection is used to regulate the air/fuel ratio at a desired value. Some of the more recent works are as follows. Ref. [1] used the Page 2 of 22 second-order sliding mode control (SMC) with appropriately sliding gain to control the fueling amount to maintain a desired air/fuel ratio. Ref. [2] developed adaptive sliding gains to compensate the effect of time delay of oxygen sensor. Ref. [3] presented the adaptation of fuel-delivery model parameters and the measurement bias of air mass flow to deal with the problem caused by engine uncertainties. Some work includes designing an observer to estimate actual air inside the cylinder. Ref. [4] added the radial basis function neural network to replace observers for uncertainties compensation. Ref. [5] used observer-based SMC in which some states were estimated to reduce the plant model mismatch so that the chattering could be reduced. Ref. [6] presented a follow-up work with a direct adaptive control method using Gaussian neural networks to compensate transient fueling dynamics and the measurement bias of mass into the manifold. Rather than controlling the fuel injection amount as in SI engine, EGR valve and variable geometry turbine (VGT) are normally used to control mass air flow (MAF) and manifold absolute pressure (MAP) in the air path of the CI engine. Several control strategies have been developed for controlling the air path of CI engine. Most control systems were designed from plant models such as transfer functions matrix as in Ref. [7]-[13] or linear or nonlinear state-space model as in Ref. [14]-[18]. Some control systems were designed without having the plant models in their algorithms. Ref. [19] used fuzzy logic control. Ref. [20] used the SMC for VGT control only. Ref. [21] developed the algorithm based on adaptive sliding mode control. Since the DDF engine has CNG injected at intake ports and diesel fuel injected at the cylinders, its air-path control can mimic those of SI engine or CI engine. Like an SI engine, air/fuel ratio is controlled. However, the optimal desired air/fuel ratio is not fixed at one, but rather changes with operating points. Ref. [22] presented an algorithm in choosing to actuate either throttle or EGR valve to regulate air/fuel ratio of a DDF engine, while pumping loss is minimized. Ref. [23] is a follow-up work by adding the effect on adjusting the total fuel amount. Like a CI engine, MAF and MAP are controlled. Ref. [24] applied quantitative feedback theory (QFT) to control MAF and MAP from throttle and EGR valve. In this paper, two multivariable SMC techniques were implemented and compared. The techniques have already existed in the literature but have not been used in air-path control applications. The SMC technique aim is to drive the system state variables to the origin. There are two phases: reaching and sliding. During reaching phase, the state variables are driven to a sliding surface and maintained there. During sliding phase, the state variables on the sliding surface moves to the origin. For tracking, a state-space model is formulated to have the tracking errors as its state variables. With SMC, the tracking errors are driven to the origin (zeros.) Some advantages over other techniques are as follows: • Plant model is normally not required in the algorithm. However, knowing the plant model helps improving the quality of the control system in two ways. First, the plant models upper bound can be used to compute, rather than trial-and-error, a controller parameter. Second, it can be feed-forwarded to cancel known terms. Doing so reduces the effort from fast-switching control, therefore reduces chattering. • During sliding phase, the system order reduces from n to n m − , where n is the number of state variables and m is the number of control inputs. Reduced order consumes less computational time. • During sliding phase, the system is not affected by matched uncertainties. Matched uncertainties are disturbances, noise, and plant uncertainties that can be directly altered by the control inputs. Two SMC techniques used in this paper are model-reference SMC and integrator-augmented SMC. In modelreference SMC, a reference model, relating the reference to the plant output, is formulated to meet various timedomain specifications. Then, the SMC controller is designed so that the errors between the reference model state variables and the plant state variables are driven to zeros. As a result, the plant output is close to the reference model output, and the specifications can be met. In integrator-augmented SMC, a state variable, Page 3 of 22 representing integral of the tracking error, is augmented to the plant state variables. Then, the SMC controller is designed to drive all state variables to zeros. The SMC used in this paper was based on the unit-vector approach, proposed in Ref. [25]. The unit-vector approach was shown to handle unmatched uncertainties effectively. The model-reference SMC and integratoraugmented SMC methods, suitable for tracking of multi-input-multi-output (MIMO) systems, were taken from Ref. [26]. Basics of the SMC method can be found from Ref. [26]Ref. [28] and references therein. Stability analysis of nonlinear systems used in this paper can be found from Ref. [29]. A 2KD-FTV Toyota CI engine was modified as a DDF engine and was installed in a test cell with an engine dynamometer. CNG was injected at each intake port as in multi-point injection. Diesel was directly injected into each cylinder as liquid spark plug. Throttle and EGR valve were actuated to regulate MAF and MAP at various set points. Simulation and experimental results have shown that the MAF and MAP can follow their set points closely with benefits from the advantages of the SMC techniques mentioned above. The paper presents topics in the following order: • Air-path system identification • Model-reference SMC design • Integrator-augmented SMC design • Experimental results • Conclusions • References, contact information, and acknowledgments AIR-PATH SYSTEM IDENTIFICATION A Toyota 2KD-FTV diesel engine was converted to a DDF engine and was installed in a test bed with an engine dynamometer. The test bed management system is AVL PUMA, and data acquisition and control hardware is National Instruments. Table 1 lists some important specifications of the engine. Figure 1 depicts the air-path system of the engine. Parameters in Figure 1 describe measurable quantities from the engine test bed. , , , T p W m are for temperature, pressure, mass flow rate, and mass, respectively. Subscripts , , , , , , a c p i x t e are for ambient, compressor, plenum, intake manifold, exhausted manifold, turbine, and exhausted pipe, respectively. Throttle and EGR valves are two actuators to control MAF and MAP, which are denoted by ac W and i p in Figure 1.