Keywords

1 Introduction

Greenhouse gas (GHG) emissions, mainly produced by the combustion of fossil fuels, are responsible for climate change [1]. Within these emissions, internal combustion engines contribute 17% of the total [2, 3]. In response, different cities and countries have agreed to replace internal combustion engines (ICE) with electric powertrains [4]; however, independent analyses [5] project that by 2050, 31% of passenger cars worldwide will be electric, while the rest of the fleet will continue to use traditional fuels. Therefore, to reduce emissions related to fossil fuels, it will be necessary to propose other options for the remaining fleet; some authors suggest alternative fuels, such as natural gas (NG), liquefied petroleum gas (LPG), or hydrogen (H2) [6, 7]. LPG is one of the leading alternative fuels used in the automotive industry. Its composition is essentially a mixture of propane (C3H8) and butane (C4H10), which varies according to technical requirements defined in standards such as N589 [8]. The main advantage of LPG is that, given its chemical characteristics, it can liquefy at low pressures, allowing it to maintain an energy density like traditional fuels [9]. However, the massive use of gaseous fuels presents critical challenges in transforming existing engines. One of the main problems is the control of knocking [10,11,12]. Its occurrence limits the engine’s thermal efficiency, restricting the exploitation of its advantages. LPG is sensitive to these effects because its resistance to knock will depend directly on its composition, so the calibration of an engine must consider this parameter. This work proposes a methodology to pre-design the transformation of a combustion engine from the original fuel, allowing it to speed up the process of fuel change and reduce production costs.

2 Methods and Materials

This paper proposes a methodology to understand the effects of a transformation from a base fuel to a new fuel. In this case, the conversion of an engine designed for NG to LPG will be used to apply the methodology. This type of transformation presents challenges; one of the most relevant, considering the properties of both fuels, is determining the admissible compression ratio (CR) for the new fuel. Therefore, the thermochemical effects of changing the fuel are analyzed to determine under what conditions knock combustion will occur.

2.1 Materials

The engine to be converted is a 4L Medium-Duty, a 4-Stroke SI engine with multipoint fuel injection and 12:1 compression ratio, maximum power of 100 kW at 3500 rpm, and maximum torque of 350 Nm at 1500 rpm. The engine was originally designed to run on NG; the objective is to convert the engine to run on automotive LPG, also called AUTOGAS. LPG automotive fuel is composed of propane (C3H8) at 38.42% and butane (C4H10) at 60.37%. Motor Octane Number (MON), in the case of NG (CH4), is 120, while LPG is 93.63. As LPG is composed of propane and butane, its boiling point varies between −42 ºC and −0.5 ºC, which allows it to liquefy at low pressures.

2.2 Methodology

To carry out the analysis, appropriate engine conditions must be defined to predict under what circumstances the auto-ignition will occur with the new fuel. The study reviewed 12 points of the engine and estimated the pressures to which the cylinder would be subjected according to the fuel. The analysis shows that when the maximum torque is reached, the highest pressures and temperatures appear inside the cylinder for both fuels. It is observed that the pressure for LPG is 5 bars lower than the pressure of NG; however, the antiknock characteristics of LPG are lower than those of NG. In addition, in lower-speed operation, the combustion time is longer, increasing the probability of auto-ignition. Considering the analysis, the engine critical condition was defined at maximum torque and minimum engine speed, i.e., at 350 Nm and 1500 rpm, corresponding to 55 kW of power. Once the critical condition has been determined, it is necessary to estimate the reference pressure and temperature at the bottom dead center (BDC). By establishing the critical condition, it is possible to calculate a compression to determine the mixture conditions at the top dead center (TDC), i.e., in the worst case and just before combustion occurs.

Equation (1), which corresponds to the effective power, is used to obtain the pressure and temperature at BDC, a reference density(ρref) is estimated for the critical condition.

$$N_{e} = \eta_{{e{ }}} { }\eta_{v} { }F_{r} { }F_{e} { }\rho_{ref} { }V_{{T{ }}} n{ }i{ }LHV$$
(1)

The same effective power is used for both fuels to maintain the original engine performance. At the same performance in NG and LPG, the reference density must change, i.e., the inlet conditions of the mixture must be different. Thus, it is possible to determine the conditions at the BDC. Using the ideal gas equation of state, Eq. (2), it is possible to calculate the pressure at BDC for a given temperature and CR; the TDC parameters can be obtained with Eq. (3) if a polytropic compression is assumed.

$$PV = m{ }R{ }T$$
(2)
$$PV^{n} = k$$
(3)

Pressure at BDC is obtained considering the reference density and assuming a fixed temperature. A range of 100 ± 20 ºC was defined as an appropriate temperature at BDC. Once BDC pressure and temperature conditions are calculated, Eq. (3) is applied to model a polytropic compressionFootnote 1. The result corresponds to the conditions at TDC of pressure and temperature according to CR. Once pressure and temperature are calculated for the given conditions, the method performs a predictive study of the combustion process. First, the auto-ignition delay (AID) is calculated, predicting the onset of combustion by detonation. A 0D homogeneous reactor model is used, assuming constant pressure conditions. In addition, laminar flame speed (SL) calculations are performed to estimate the combustion velocity. The CONVERGE v2.4 tool was used for both fuels. The expression ∆Knock, Eq. (4), is proposed to establish the existence of auto-ignition, which is expressed as follows:

$${\Delta }Knock = AID_{LPG} - { }\left( {\frac{{t_{combustion} }}{{{\raise0.7ex\hbox{${SL_{LPG} }$} \!\mathord{\left/ {\vphantom {{SL_{LPG} } {SL_{BASE} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${SL_{BASE} }$}}}}} \right){ }\left[ {ms} \right]$$
(4)

Equation (4) compares the LPG’s AID with the combustion time (tcombustion) normalized by the LPG’s SL and the base fuel SL ratio, in this case, NG.

3 Results

The combination of both effects (AID and SL) is relevant for the analysis; even though LPG has a higher SL, its AID is lower than NG, so in the same thermodynamic condition, there is likely undesired auto-ignition in the engine. To contextualize this difference, it is necessary to establish a reference corresponding to the combustion time. The reference combustion time was estimated according to previous studies. A combustion duration of 40 CAD at a speed of 1500 rpm was defined. The combustion time for the engine under these conditions is 4.44 ms. The proposed hypothesis is that knocking will occur if the AID time is shorter than the combustion time or the combustion time normalized by SL, i.e., if ∆Knock < 0.

3.1 Knocking Zone

To define a zone of knocking appearance, the analysis considers the calculation of a 5 ×  5 matrix for each CR (from CR 10:1 to 12:1) by imposing a range of 5 BDC temperatures (80 ºC, 90 ºC, 100 ºC, 110 ºC, and 120 ºC) and five pressures related to the temperatures. For each matrix, AID and SL are calculated in each cell; then, it is possible to evaluate AID − tcombustion and ∆Knock. Finally, the percentage of cells that present knocking conditions (<0) over the total of each CR matrix is established as an indicator, i.e., if 25 out of 25 conditions present negative values, the probability of knocking is 100%. Results of this calculation are shown in Table 1, AID − tcombustion situations vary from 0% at CR 10:1 to 96% at CR12:1; in the case of ∆Knock, from CR 10:1 to CR 11:25, the probability is 0%, and CR 12:1, it is 28%.

Table 1. Probability of knocking by CR for LPG
Full size table

In Fig. 1, the upper blue curve corresponds to the percentage of cells, per CR, where AID − tcombustion < 0, and the lower blue curve is the percentage of cells where ∆Knock < 0; the area between both curves is defined as the Knocking Zone. It is observed that at CR 12:1, the range of auto-ignition probability is 28% to 96%, while at CR 11:1, the content is 0% to 48%; however, at a compression ratio of 10:1, the knocking probability is 0%.

Fig. 1.
figure 1

Knocking Probability (Knocking Zone) for LPG according to compression ratio.

Full size image

3.2 Experimental Results

Considering the theoretical analysis, it was decided to perform experimental tests on two cylinder heads with different CR, first with a volumetric CR of 11.05:1 and the second with a volumetric CR of 9.86:1. The engine performance was evaluated in an asynchronous dynamometer where the instantaneous brake power and torque were measured. The engine is also instrumented with a pressure sensor inside the cylinder, allowing the analysis of combustion behavior. The conditions of each measurement were at 1500 rpm and maximum torque, which according to the configuration, reached 300 Nm for the 11:1 compression ratio and 280 Nm for the 10:1 compression ratio. Figure 2 shows the in-cylinder pressure for the two proposed arrangements. In the case of CR = 11.05:1, it is observed that there is a dispersion of 25 bar between combustion cycles. In addition, it is possible to observe instability in the signal for a significant number of cycles. However, in the combustion at maximum torque with CR = 9.86:1, it is observed that the dispersion between cycles is 10 bar, with stability in the signal for every measured cycle.

Fig. 2.
figure 2

Average pressure and its dispersion (10 cycles) inside the cylinder for the modified cylinder head with CR = 11.05:1 (left) and CR = 9.86:1 (right), using LPG.

Full size image

4 Conclusions

This research presents a methodology to promote the transformation of traditional engines to run on zero or low-carbon fuels. This methodology is the first step to defining the conditions where a transformation becomes feasible. The Knocking Zone obtained shows that for LPG, for CR = 12:1, the probability of auto-ignition is between 28% to 96%; for CR = 11:1, the range is 0% to 48%, however, in CR = 10:1, the probability of auto-ignition is 0%. When performing the experimental tests, it was found that with a cylinder head of CR = 11.05:1, at 1500 rpm and maximum torque, there is knocking; however, for a CR = 10:1, this phenomenon does not appear. In the literature, it is defined that CR for engines powered by LPG would be between 10:1 and 11:1 [13, 14], arguing that the optimum point of thermal efficiency would be between CR of 10:1 and 10.5:1. However, other authors analyze that if the auto-ignition level can be controlled, the maximum efficiencies in LPG would be reached at CR close to 12:1 [15]. Indeed, the calibration process of a converted engine considers an essential number of variables to be adjusted. However, it is relevant to define a priori the feasibility that the base conditions of combustion are favorable. Finally, the methodology allows for predicting the occurrence of knocking in the transformation of an internal combustion engine. The prediction defines a range of circumstances because there are strategies that will enable reducing the occurrence of knocking under certain conditions, such as ignition delay, intake conditions, detection strategies, etc. Despite the uncertainty, the tool allows for designing a pre-conversion strategy based on thermodynamic calculations without having complex combustion analysis models, which is its main functionality.