Jonas Sjöblom

AN EVALUATION OF ORTHOGONAL SIGNAL CORRECTION APPLIED TO CALIBRATION TRANSFER OF NEAR INFRARED SPECTRA

Orthogonal signal correction (OSC) is a technique for pre-processing of, for example, NIR-spectra before they are subjected to a multivariate calibration. With OSC the X-matrix is corrected by a subtraction of variation that is orthogonal to the calibration Y-matrix. This correction can then be applied to new spectra that are going to be used in predictions. The aim of this study is to investigate if the OSC transform makes the spectra less dependent of instrument variation. This may result in easier calibration model transfer between different instruments without creating or re-analysing the whole calibration sample set. OSC was applied to NIR-spectra that were used in a calibration for the water content in a pharmaceutical product. Partial Least Squares calibrations were then compared to other calibration models with uncorrected spectra, models with spectra subjected to multiplicative signal correction, and a number of other transfer methods. The performance of OSC was on the same level as for piece-wise direct standardisation and spectral offset correction for each individual instrument and PLS-models with both instruments included.

Quantitative FT-Raman analysis of two crystal forms of a pharmaceutical compound

A pharmaceutical active compound H appears in two polymorphs, A and B, that are stable below and above room temperature, respectively. The A- and B-forms were found to have distinct FT-Raman spectra, in particular for a band at 1716 cm-1 (A-form) or 1724 cm-1 (B-form). Mixtures of A- in B-form were prepared, and the relative intensity of the characteristic bands at 1716 and 1724 cm-1 was found to be proportional to the relative amounts of A- and B-form in the mixtures. A calibration was made which was linear in the range from 1.8 to 15.4% (w/w) of A- in B-form. The results were compared with other methods for analysis of polymorphs: FT-IR spectrometry, differential scanning calorimetry, and powder X-ray diffractometry. A novel FT-Raman sample presentation method for inhomogeneous samples is presented.

Bridging the Gap Between Lab Scale and Full Scale Catalysis Experimentation

A new experimental set-up applied to engine emission experiments is presented which enables experiments where reactor conditions can be varied independently. These conditions include residence time, which can be varied via flow reduction and the possibility to use very different sizes of catalyst, temperature (using cooler or heaters) and concentrations (by dilution or adding gases). The system is designed to enhance knowledge transfer between lab scale and engine bench scale catalysis experimentation and to enable efficient design of experiments and catalytic reactor modeling. Two examples are presented dealing with capture of particulates and selective NOX reduction over a silver–alumina catalyst. Different experimental issues and their resolutions are also presented.

New approach for microkinetic mean-field modelling using latent variables

A detailed microkinetic model for the storage and reduction steps in a lab-scale NOX storage reactor was taken from the literature and a large umber of parameters were subject to fitting to a new set of experimental transient data. The parameters chosen for fitting were selected by ensitivity analysis of the Jacobian matrix of all adjustable parameters using latent variable (LV) models (partial least squares, PLS). The analysis of the LV structure of the Jacobian matrix is important because it indicates the experimental rank, i.e. how many parameters that re relevant to fit. Two slightly different methods both using LV models are presented in order to select good candidate parameters for fitting as ell as a method to fit linear combinations of many parameters that only span the experimental space. By using this methodology the risk of overfitting is reduced and the chances for successful fitting are increased by supplying an appropriate umber of parameters that also are uncorrelated (independent) for the given experiment. This methodology contributes to a better understanding of the catalytic process as well as a potential for more efficient parameter fitting for complex ransient heterogeneous catalytic systems. The application of this kind of sensitivity analysis offers a new approach to microkinetic modelling. 2006 Elsevier Ltd. All rights reserved.

Combined Effects of Late IVC and EGR on Low-load Diesel Combustion

The increasing demand for improved efficiency of diesel engines requires more advanced combustion solutions. These solutions include the use of variable valve timings in combination with more traditional methods such as EGR, turbocharging and advanced injection systems. By modifying the characteristics of the charge air, further hardware optimization becomes possible. In the current investigation, the effect of late intake valve closing (LIVC) was investigated together with the effect of (external) exhaust gas recirculation (EGR) in a single cylinder heavy duty diesel engine. Different injection timings and injection pressures were investigated. The mass flow of oxygen was kept constant in order to show how the density and temperature of the reactant mixture affect the combustion and emission characteristics. The combustion results showed that if the oxygen mass flow is kept constant, an EGR approach is more efficient than LIVC in lowering fuel consumption due to the effects of increased cylinder gas flow which improves fuel conversion efficiency. It was shown that the ignition delay for a fixed combustion phasing was independent of EGR but could be increased by LIVC. The peak pressure was more strongly affected by EGR due to the larger gas flow but this response can be reduced by means of LIVC. After compensating for combustion timing effects, the reduced peak pressure was mainly attributed to reduced effective compression ratio resulting from the LIVC. The results show how variable valve timing can be used as one important tool to obtain better combustion characteristics and thus enable more efficient powertrains.

Parameter Estimation of a DOC from Engine Rig Experiments with a Discretized Catalyst Washcoat Model

In this thesis methods of parameter estimation of a Diesel Oxidation Catalyst (DOC) from engine rig experiments were investigated. The investigation did not only include methods of parameter fitting to experimental data but a large effort was also put into catalyst modeling and experimental design. Several different catalyst configurations were used with varying Pt loading, washcoat thickness and volume. To further expand the experimental space, engine operating points were chosen with a wide variation in variables (inlet conditions) and both transient and stationary operation was used. A catalyst model was developed where the catalyst washcoat was discretized as tanks in series both radially and axially and for parameter estimation a traditional gradient search method was used. Four different modeling approaches were used for parameter tuning where the most successful one tuned kinetic parameters as well as internal mass transfer parameters. It was also shown that it is of high importance that the kinetic model used has an intrinsic structure when the catalyst model separates mass transport and kinetics and when several catalyst configurations are used. A new method was evaluated where sensitivity analysis and data selection was used as a part of the parameter estimation. This methodology renders better statistical properties and should improve the parameter tuning when using gradient search methods. Furthermore, a reduced computational cost could be achieved by using only the most relevant data points during parameter tuning. The evaluated method did neither result in an improved fit to measurement data nor reduce the time for parameter tuning compared to a reference case. Adjustment of an unbalanced weighting of the residuals for the different components was identified as the most important factor for a future improvement of the method but more transient experimental data was also suggested as a possible improvement. A method of creating fast transients in concentration for a full scale engine rig system was presented and evaluated. The method included an engine rig where an SCR with urea injection and a DOC with bypass possibility were situated between the test object, which also was a DOC, and the engine. By controlling urea injection and DOC bypass a wide range of exhaust compositions, not possible by only controlling the engine, could be achieved which will improve the possibilities for parameter estimation for the modeling of the DOC in future studies.

Characterization of Particulate matter and the capture efficiency in open metal substrates

The capture efficiency (CE) of particulate matter in a novel metal substrate was evaluated using an exhaust gas after treatment system rig. The CE was measured for different temperatures, flows and channel lengths. The trends of CE showed the expected behavior as the CE increased for higher temperatures, lower velocities or longer channel lengths. The experimental results were compared against theoretical calculations of different types in order to visualize and interpret the observed CE in the novel metal substrate. Computational fluid dynamics simulations investigations indicated that inertial mechanisms on particle deposition were active in the metal substrate. It was also demonstrated that the channel length was the most significant factor for increased CE.

Use of Late IVC and EGR to Enhance Diesel Engine Optimization

The increasing demand for improved efficiency of diesel engines requires more advanced combustion solutions. In addition to traditional methods such as EGR, turbocharging and advanced injection systems, variable valve timing is now available at a reasonable production cost. The use of variable inlet valve timing provides an efficient way for Low-Temperature Combustion (LTC) which provides high thermal efficiency in combination of low emission levels. Furthermore, by modifying the characteristics of the charge air (i.e. by means of EGR, boost pressure and late inlet valve closing, LIVC), further hardware optimization becomes possible, e.g. by increasing compression ratio without reaching critical peak pressures. In the present study, the effect of LIVC was investigated together with the effect of EGR in a single cylinder heavy duty diesel engine. The engine was equipped with pneumatically controlled inlet valves and a high pressure common rail injector. Different injection timings and injection pressures were investigated at two different load points. The mass flow of oxygen was kept constant in order to show how the physical properties (density and temperature) affect the combustion and emission characteristics. The combustion results showed that if the oxygen mass flow is kept constant, EGR is a more efficient way (compared to LIVC) to lower the fuel consumption since it is accompanied with the largest gas flow and thus increased fuel conversion efficiency. The LIVC decreased the fuel consumption at low loads and reduced the emissions at both loads. Transportation of people and goods tend to increase and since internal combustion engines will remain a major power supply for many years to come, reduced fuel consumption is an utmost important way to decrease the CO2 emissions and to move towards a sustainable society. The results in this study show that variable inlet valve timing can be used as one important complementary tool to obtain better combustion characteristics and thus enabling more efficient powertrains.

Model-based experimental screening for DOC parameter estimation

In the current study a parameter estimation method based on data screening by sensitivity analysis is presented. The method applied Multivariate Data Analysis (MVDA) on a large transient data set to select different subsets on which parameters estimation was performed. The subset was continuously updated as the parameter values developed using Principal Component Analysis (PCA) and D-optimal onion design. The measurement data was taken from a Diesel Oxidation Catalyst (DOC) connected to a full scale engine rig and both kinetic and mass transport parameters were estimated. The methodology was compared to a conventional parameter estimation method and it was concluded that the proposed method achieved a 32% lower residual sum of squares but also that it displayed less tendencies to converge to a local minima. The computational time was however significantly longer for the evaluated method.

New Methodology for Transient Engine Rig Experiments for Efficient Parameter Tuning

Introduction The diesel oxidation catalyst (DOC) is a well established technology to reduce CO and hydrocarbon (HC) emissions from diesel engines. Strengthened emission standards have made the importance of the DOC even greater in recent years since it plays an indispensible role in enhancing the performance of diesel particulate filters (DPF) and selective catalytic reduction (SCR) by utilization of NO oxidation to NO2. Therefore correct prediction of the DOC performance is very important for simulations of the entire aftertreatment system. When performing kinetic parameter estimation, laboratory scale experimental data is generally used. In laboratory scale it is possible to use essentially any combination of exhaust gas composition and temperature which makes it possible to estimate parameters over a wide range of conditions. However the applicability of these parameters in full scale models is often limited. Parameter estimation on full scale engine rig experiments on the other hand is limited by the exhaust compositions that are possible for the engine to produce. As a result, the fraction of CO is closely linked to the fraction of hydrocarbons and the fraction of NO is closely linked to the fraction of NO2. When switching between two engine operation points it generally takes several minutes before the properties of the emissions have stabilized. This does not only make the experiments time consuming, but it also complicates the transient modeling of the DOC since the changes in inlet properties are far from ideal step functions. In this study an experimental set-up is presented that makes it possible to change the inlet properties of the DOC without changing engine load point which results in much faster transients. The method also makes it possible to change the fraction of NO2 independently of the NO fraction. Method To achieve more controlled and faster changes in the inlet to the catalyst an extra DOC (DOC1) with the possibility for bypass flow and an SCR with urea injection are mounted before the catalyst. The fraction of exhaust gas flow through DOC1 allows variation in the conversion of HC and CO to CO2 and the conversion of NO to NO2. By injecting different amounts of urea the conversion of NO2 and NO to N2 is controlled. The SCR also makes it possible to obtain an inlet composition to the DOC that contains NO2 but is free of NO. Fast changes in inlet conditions are in other words possible and it is also possible to achieve compositions not achievable by only controlling the operation of the engine. Experiments have been performed at several engine conditions and using catalysts with different noble metal loading, lengths and washcoat thicknesses. To achieve high HC and CO concentrations the engine was tuned to run with late fuel injection. Significance A method to carry out engine rig experiments with a wider range of emission conditions makes it possible to more efficiently retune model parameters for a full-scale catalyst from literature data. This should result in faster model development which is of great importance in exhaust gas aftertreatment.