Combustion and Emissions of a Small SI Engine with Buthanol Blend Fuels

The bioalcohols are an important alternative in the general efforts to replace the fossil fuels in transportation by renewable fuels. The global share of Bioethanol used for transportation is continuously increasing. Butanol, a four-carbon alcohol, is considered in the last years as an interesting alternative fuel, both for Diesel and for Gasoline application. Its advantages for engine operation are: good miscibility with gasoline and diesel fuels, higher calorific value than Ethanol, lower hygroscopicity, lower corrosivity and possibility of replacing aviation fuels. In the present work research with different Butanol portions in gasoline (BuXX) was performed on the 2-cylinder SI engine with variations of several parameters on engine dynamometer. In the steady state operation, it was found that Bu-blends generally reduce the emissions of CO, HC, NOx in untreated exhaust gas and have a very little influence on catalytic conversion rates of the 3-way-catalyst. At lower engine part load, “Bu” shortens the inflammation lag and reduces the cyclic dispersion of combustion. Nevertheless, this advantage disappears at higher engine loads and with higher “Bu” portions. The present paper shows some examples of the most important results.


Introduction
Butanol (CH 3 (CH 2 ) 3 OH) has a four-carbon structure and is a higher-chain alcohol than Ethanol, as the carbon atoms can either form a straight chain (n-Butanol) or a branched structure (iso-Butanol), thus resulting in different properties. Consequently, it exists as different isomers depending on the location of the hydroxyl group (-OH) and carbon chain structure, with Butanol production from biomass tending to yield mainly straight chain molecules. 1-Butanol, better known as n-Butanol (normal Butanol), has a straight-chain structure with the hydroxyl group (-OH) at the terminal carbon.
n-Butanol is of particular interest as a renewable biofuel as it is less hydrophilic, and possesses higher energy content, higher cetane number, higher viscosity, lower vapour pressure, higher flash point and higher miscibility than Ethanol, making it more preferable than Ethanol for blending with diesel fuel. It is also easily miscible with gasoline and it has no corrosive, or destructing activity on plastics, or metals, like Ethanol or Methanol.
The good miscibility, lower hygroscopicity and lower corrosivity make Butanol to an interesting alternative.

2
The trend of downsizing the SI-engines in the last years implies much higher specific torques and with it an aptitude of knocking and mega-knocking at high-and full load. The alcohols have a higher Octane Numbers (RON), are more resistant to knocking and are a welcomed solution for this new technology of engines, [1].
A basic research of butanol blends Bu20 & Bu100 was performed on mono-cylinder engines with optical access to the combustion chamber, [2,3]. One of the engines was with GDI configuration. It was demonstrated, that the alcohol blend improved the internal mixture preparation and reduced the carbonaceous compounds formation and soot.
Concerning the characteristics of combustion Bu100 was similar to gasoline. This research considered only little number of constant operating points.
Using n-Butanol in a optical port fuel injection (PFI) SI engine slightly higher combustion rates and lower formation of particulates was found compared to gasoline, [4,5]. Similarly [6] reported that the duration of the early combustion stage and length of combustion in an SI engine were, compared to gasoline, shortened with increased n-butanol share, and slightly lower variability of indicated mean pressure (IMEP) was observed when running on neat n-butanol. Shorter early combustion stage, faster combustion and better combustion stability were also observed by other researchers [7,8].
The alcohol blend fuels E85 & Bu85 were tested on a vehicle with 3WC in road application and with on-board measuring system for exhaust emissions, [9]. It was stated for butanol, that it has no significant influence on CO & HC, but it increases strongly NO x . Nevertheless, this is due to the limits of Lambda regulation and as effect of it to the production of too many lean Lambda excursions during the transients. The warm operation with Bu85 was with no problems, the cold startability and emissions were not investigated.
The presented tests were performed in the IC-Engines Laboratory of the University of Applied Sciences, Biel, CH within the framework of project GasBut (Gasoline + Butanol). The research objectives were:  full load (FL) characteristics,  variations of spark timing (αz),  research of lean operation limit at part load (λ-variations),  research of EGR limit at part load (EGR-variations),  research of knock limit at FL.
With this research, it was possible to investigate the influences of fuel quality on engine internal processes as well as on the standard exhaust aftertreatment (3WC).
The research was performed with Bu0, Bu30, Bu60 and Bu100. Figure 1 shows the engine on the engine dynamometer and Table 1 summarizes the most important engine data. The research was conducted on a Lombardini 2-cylinder SI-engine 0.5L. This engine is equipped with a programmable control unit, which allows a flexible parametrization of spark timing and equivalence ratio. There is a combustion chamber pressure indication with data acquisition and processing, which allows an accurate combustion diagnostic. The test bench with eddy-current dynamometer is equipped with analysis of limited exhaust gas components.

Fuels
Following base fuels were used for the research:  gasoline (RON 95) from the Swiss market  n-Butanol or i-Butanol from Thommen-Furler AG. As blend fuels were used: Bu30, Bu60 and Bu100 (30% vol, 60% vol Butanol and respectively neat Butanol 100% vol). Table 2 represents the most important data of the fuels (according to the literature sources). It can be remarked that with increasing share of Butanol the Oxygen content of blend fuel increases and the heat value and stoichiometric air requirement decrease.

Lubricant
For all tests, a special lube oil MOTUL 300V Le Mans 20W-60 was used. Fig. 2 represents the special systems installed on the engine, or in its periphery for analysis of emissions and for combustion diagnostics.

Engine dynamometer and standard test equipment
In the present work, an EGR-system (EGR-line, valve and cooler) was installed on the engine. The EGR-rate is estimated by means of CO 2 -measurement in exhaust and intake of the engine.   Table 3 shows the used laboratory equipment of the engine dynamometer. Different parameters are registered on-line via PC. The continuous registration of all parameters is possible.

Combustion diagnostics -pressure indication
During all tests, cylinder pressure was indicated, so that the combustion characteristics could be valued in each case. Therefore, following devices were used, see Table 4.

Test procedures on engine dynamometer
The stationary testing was performed at different constant operating points (OP's) of the engine. These OP's were chosen at different speeds and at different loads. One part shows the full load characteristics and the other part represents partial load. The operating points in the engine map for entire test program show Fig. 4 and Table 5.

Variations of spark timing αz
Variation of spark advance at engine part load can be performed in two ways: at constant OP (n/M), or at constant throttle position. Both variants of tests have been performed with all investigated fuels at different OP's Fig. 5 shows the gaseous emissions at higher part load and Fig. 6 represents some combustion characteristics at lower and at higher part load. These pictures represent mostly the advantages of Butanol blends. Nevertheless, the complete picture, which results from all tests (4 OP's not represented here) shows some limited or some neutral results. Following tendencies can generally be remarked with increasing share of nButanol in the blend fuel:  no effect on CO at low load, increased CO at higher load,  lowering of HCFID,  no effect on NOx at low load, clear reduction of NO x at higher load especially with nBu100,  lowering of CO 2 ,  αz for α50%@9°CA a TDC generally later for BuXX, For comparisons: nBu100 → iBu100 it can be remarked that iBu100 causes:  higher HCFID at low load and no clear differences (against nBu100) at higher load,  generally lower CO-and higher CO 2 values,  generally lower NO x values,  no differences of inflammation phase, combustion duration, COV and p max . Generally, the findings at part load could be confirmed: with increased share of Butanol there is lowering of NO x , HC and CO. The necessary spark timing (α z opt ) is nearer to the TDC, the maximum pressure rise is higher and the cyclic irregularities of combustion are lower. All these are signs of accelerated and improved inflammation phase. These effects of improved combustion are more pronounced at OP1 (lowest engine speed & torque) than at higher OP4 and OP6

Variations of Lambda 
These variations were also performed with all fuels at different engine operating points.   Increasing of Lambda was performed up to the lean operation limit, which was attained at strong increasing of cyclic irregularities (high values of COV) and increasing of HC.
The lean limit for this engine was: at OP2: λ = 1. The reason for this tendency is the lowering of the internal residual gas content with the increasing engine load.
The diagrams of results in function of λ show the comparisons between the fuels. With increasing of Butanol content following tendencies can be remarked:  lower HC-values and lower HC-increase at lean limit,  lower maximum values of NO x ,  shorter inflammation phase (IP = α 5% -α z ), especially with Bu60 & Bu100,  lower cyclic dispersion (COV) at lean limit.
Comparisons of fuels at λ  1.10 and α zopt confirm these statements. With increasing BuXX there are:  reduction of HC  shortening of IP (except OP2) and reduction of COV.
There are also tendencies of reducing NO x and lowering T exh with the higher Butanol content. Summarizing: the present results of Lambda variations confirm the statements from previous tests. Butanol blended to gasoline slightly shortens the inflammation phase and lowers the cyclic irregularities of combustion at part load operation of the engine. It moves the lean operation limit to higher λ-values and it has positive influences on lowering NO x and HC.

Variations of EGR
The variations of EGR at part load were initially performed at OP4 with all fuels (Bu 0/30/60/100).
General tendency was found, that the higher Bu-content enables higher EGR-rate at the same COV (cyclic dispersion). This is a result of improved inflammation with Butanol.
At OP12 there was only a limited possibility of realizing EGR (gasoline up to 1%, Bu 100 up to 6%), but the effects of increasing Bu-content were well visible.  The findings are confirmed: with increasing Butanol share at part load there is an improved inflammation, the IP-duration is shortened, and higher EGR-rates can be attained (at COV = idem). The combustion duration is only slightly shortened with higher Bu60 and Bu100. The gaseous emission components CO, HC, NO x are generally reduced with higher BuXX.
Summarizing: there are positive effects of Butanol on inflammation at part load, which enable application of higher EGR-rates. There are also positive influences of Butanol on emissions and on the specific energy consumption.

Knocking
The objective of this part of tests was to confirm the potentials of iButanol (with higher RON) concerning knocking. It was necessary to approach slightly the knock limit and indicate the knocking with a very low intensity to avoid damaging the engine. The chosen OP was WOT at 2100 rpm with variation of spark timing and the compared fuels were: gasoline and iBu100. Fig. 11 represents cyclic dispersion of indicated pressure traces and samples of cycles without and with weak knocking.
To recognize weak knocking (weak oscillations, or irregularities on the indicated pressure signal) methods with differentiation of pressure (dp/dα) or with ROHR (dQ/dα) are applied. The second one, according to [2], was applied in the present tests. Fig. 12 confirms the advantages of iBu concerning knocking: advancing spark timing (α z ) the very weak knocking starts to be recognized with iBu at α z , which is more than 10°CA b.TDC earlier than with gasoline. Until the end of α z -variation range (70°CA b.TDC) the knocking with iBu stays very weak (K i = 0.4%), while with gasoline the knock probability increases (up to K i = 3.6%). In other words: the use of iBu moves the knock limit at FL to the higher values of spark advance. This can offer clear advantages of power and of fuel consumption in modern engines with higher compression ratio and with electronic knock control system.

Conclusions
The most important statements can be summarized as follows:  The operation with Butanol blended to gasoline is possible without any problem. With neat Butanol (Bu100) nevertheless the cold start is problematic (with engine motoring).  The lower overall heat value of BuXX-blends leads to a respectively lower full load torque without corrections of fuel dosing.  The α z -variations at part load of the engine show lowering of HC, NO x & σpmi with increasing Butanol rate.  The improvements of combustion at part load are not observed at full load and with higher Bu-content there is even longer inflammation phase and longer combustion duration.  IsoButanol causes lower CO-, higher CO 2 -and lower NO x values than nButanol, the development of combustion is affected by isoButanol, in the same way as by nButanol.  The λ-variations at part load of the engine show lowering of HC, NO x & COV with increasing Butanol rate.  Butanol blended to gasoline slightly shortens the inflammation phase and lowers the cyclic irregularities of combustion at part load operation of the engine.  With higher Bu-content the lean operation limit at part load is moved to higher λ-values.  Higher Bu-content enables higher EGR-rate at the same COV (cyclic dispersion).  There are positive influences of Butanol on emissions and on the specific energy consumption.  Concerning knocking: the use of iBu moves the knock limit at FL to the higher values of spark advance.