Research Article

Design and application of air to fuel ratio controller for LPG fueled
vehicles at typical down‑way
Suroto Munahar1 · Bagiyo Condro Purnomo1 · Muji Setiyo1 · Aris Triwiyatno2 · Joga Dharma Setiawan3

Received: 5 August 2019 / Accepted: 3 December 2019 / Published online: 6 December 2019
© Springer Nature Switzerland AG 2019

Abstract
This article presents an investigation of air–fuel ratio (AFR) controllers applied to liquefied petroleum gas (LPG) fuelled
vehicles with second-generation LPG kits. When a vehicle is running on a down-way, fuel consumption tends to be rich
because of the increased vacuum in the intake manifold. Therefore, an AFR controller was developed that can work based
on a vehicle’s tilt sensor combined with an oxygen sensor. AFR controllers are employed to regulate injectors to form
leaner mixtures. We tested the performance of AFR controller at a typical down-way of 10°, 15°, and 20°. As a result, the
AFR controller was able to increase AFR value from an average of 14.5 (without controller) to 15.5–16.2, depending on
the gear position and down-way angle. Furthermore, a greater of road slope was observed to have produced greater
AFR. This AFR controller is very promising to be applied to vehicles operating in mountainous areas.

Keywords LPG vehicle · AFR controller · Down-way

1 Introduction high input power [8] while FCV is limited by infrastructure

needed to produce hydrogen. In the medium term, hybrid
In the past few decades, environmental factors have vehicle (HVs) is a reasonable choice and it involves com-
become the main orientation in technological develop- bining internal combustion engines (gasoline/diesel) with
ment, especially concerning health issues. In addition to electric motors [8, 9]. However, this technology is also not
the industrial sector, transportation is one of the sectors yet widely accepted due to the relatively high total cost of
targeted to reduce global warming, air pollution and emis- ownership (TCO).
sions [1–3]. Therefore, the design of vehicle technology Therefore, in the short term, controlling air to fuel ratio
needs to consider emission factors [4, 5]. From another (AFR) is an alternative method to reduce fuel consumption
perspective, there is also a potential global energy crisis and emissions. This system has progressed rapidly, even
and this call for the design of technology to improve fuel with the use of proportional–integral–derivative (PID) for
efficiency for new and operating vehicles [6]. stoichiometric purpose [10]. Neural networks as intelligent
Electric vehicles (EVs) and Fuel Cell Vehicles (FCVs) are control systems have also been applied to control AFR with
very promising to reduce fuel consumption and emissions, the concept of brain tissue [11]. Several other studies have
even to zero value, in the future. However, the implemen- been conducted by processing signals generated by oxy-
tation of EVs and FCVs is constrained in developing coun- gen sensors [12], applying genetic algorithms [13], fuzzy
tries due to uncompetitive prices and limited mileage [7]. logic controllers (FLC) [14–16], diagonal recurrent neural
In EVs, the battery requires a long time to charge with network (DRNN) [17], and brakes control system [18].

* Muji Setiyo, [Link]@[Link] | 1Department of Automotive Engineering, Universitas Muhammadiyah Magelang, Magelang,
Indonesia. 2Depattment of Electrical Engineering, Universitas Diponegoro, Semarang, Indonesia. 3Department of Mechanical Engineering,
Universitas Diponegoro, Semarang, Indonesia.

SN Applied Sciences (2020) 2:37 | [Link]

Vol.:(0123456789)
Research Article SN Applied Sciences (2020) 2:37 | [Link]

Moreover, other methods to reduce emissions have Table 1  Vehicle specification

been researched to include the application of alternative Descriptions Specification
fuels such as ethanol, methanol, compressed natural gas
(CNG), and LPG [19, 20]. Ethanol produces good efficiency Engine manufacturer Toyota
and reduces emissions, but it cannot be produced in large Engine code 5A-FE
numbers except a country has a reliable policy on agricul- Cylinders Inline 4
tural land for food and energy [21]. Therefore, LPG is con- Capacity 1498 cc
sidered an alternative and choice of several countries due Bore × Stroke 78.7 × 77 mm
to many advantages such as high octane, lower exhaust Valve mechanism DOHC, 4 valves per cylinder
emissions, and availability. Maximum power output 77 kw @ 6000 rpm
Research on several variables of LPG as an alternative Maximum torque 135 Nm @ 4800 rpm
fuel has been conducted by different researchers. For Compression ratio 9.8:1
example, Morganti [22] conducted a study to test the Fuel system EFI
research octane number (RON) and motor octane num-
bers (MON) for iso-butane, propylene, n-butane and pro-
pane, followed by observations of auto-ignition from a Table 2  Injector specification
mixture of propane and butane. In another study, Chikhi Descriptions Specification
[23] investigated CO, HC, NOx and ­CO2 emissions produced
by 17 units of bi-fuel vehicle using LPG to replace gasoline Voltage range 12–15 Volt
and diesel. Moreover, it is also possible to control the sulfur Maximum pressure 16 bar
and toxic gases produced by LPG vehicles to achieve bet- Controlled Solenoid
ter emissions [24, 25]. Other studies focus on iso-octane
and air mixtures [26], performance characteristics of LPG,
CNG and LNG vehicles [27], direct-injection application This control system works based on the primary informa-
with lean combustion methods [28], and risk analysis of tion from the tilt sensor.
the safety of LPG-fueled cars [29].
Meanwhile, several research works have also been con-
ducted on the control of LPG. In 2015, Erkus [30] developed
2 Materials and methods
an LPG control system to be applied in carburetor-based
engines. The results of this study confirm an increase in
2.1 Vehicle specification
engine performance and better exhaust emissions com-
pared to the carburetor system. Others study, including the
This research was conducted on gasoline cars modified
fuel cut-off method to cut off LPG flow to the engine during
into LPG with the Vapor Phase Injection (VPI) system. Vehi-
deceleration by controlling the solenoid on the vaporizer
cle and injector specifications are presented in Tables 1
[31, 32], emission comparison using a control system on
and 2, respectively.
liquid phase injection (LPI) and direct injection (DI) [33], as
With a natural suction system, to achieve a stoichio-
well as the characteristics of injection duration and con-
metric mixture, the amount of injected LPG depends on
trol [34–36]. This has led to the development of intelligent
the air entering the combustion chamber. Therefore, the
control systems to support fuel efficiency. However, the
stoichiometric mixture (ṁstoich) on the cylinder is highly
studies conducted have not considered the contours of the
dependent on the mass of LPG injected, the number of
land, such as up-way and down-way. When a vehicle passes
cylinders (icyl), and the engine speed (n). The formula to
through a down-way, kinetic force, and gravity affect its
obtain the stoichiometric mixture (𝜆 = 1) is presented in
movement. Meanwhile, when the vehicle accelerates on a
the following equation.
down-way, the fuel is reduced or even cut off.
Furthermore, even though LPG kits technology is now 2ṁ air
equal to GDI technology, in fact, more LPG vehicles use
𝜆=
icyl ⋅ MLPG ⋅ mLPG ⋅ n (1)
second-generation LPG kits (vapor phase injection, VPI)
without strict AFR and emission settings [37]. With second- In the VPI system, LPG is divided into two phases, liq-
generation LPG kits, AFR stoichiometry is only obtained uid and gas. LPG in the tank until the vaporizer is a liquid
in partial conditions. When the vehicle accelerates in the phase, while LPG injected into the intake manifold from
down-way, the tendency for low AFR is greater than the the vaporizer is a gas phase. The injected LPG depends
high AFR. Therefore, we developed the AFR control system on the effective flow area—μA ­(mm2), injection dura-
on LPG vehicles that pay attention to the slope of the road. tion—tLPG (s), gas pressure—ΔPLPG (Pa) at intake manifold

Vol:.(1234567890)
SN Applied Sciences (2020) 2:37 | [Link] Research Article

temperature—T0 (K) and pressure—P0 (Pa). In addition, 2.2 AFR controller

LPG has cp of 1750 J/kg K and RLPG of 161.26 J/kmol K,
therefore, the mass of the injected LPG is presented in the The AFR controller developed in this study worked as a
following equation [36]. signal manipulator generated by the throttle position sen-
sor (TPS). The controller circuit was paired between the
TPS and ECU which was regulating the AFR based on the

( ) √ [ ( ) ]
p0+ΔPLPG cp−RLPG √
√
p0 p0 RLPG
mLPG = (𝜇A)LPG ⋅ ⋅ ⋅ 2 ⋅ cp ⋅ T 0 1 −
√ ⋅ ⋅ tLPG (2)
RLPG T0 p0+ΔPLPG cp p0+ΔPLPG cp

Fig. 1  Set up LPG controller

and AFR measurement

Fig. 2  Wiring diagram of AFR controller

Vol.:(0123456789)
Research Article SN Applied Sciences (2020) 2:37 | [Link]

information from the TPS, camshaft position sensor, and the throttle valve position that sends signals to the ECU.
signals from other sensors. The injector sprayed LPG in the Moreover, the controller consists of power supply, angle
gas phase to be regulated using the pulse wide modula- sensor, transistor, relay, capacitor, and a variable resistor.
tion (PWM) method. The angle sensor, as a road tilt angle The power supply, as a voltage source, activates the relay
detector, was paired with the AFR controller while the and angle sensor while the NPN transistor activates the
AFR meter obtains data from an oxygen sensor attached relay when it gets triggered from the sensor angle. Fur-
to the exhaust manifold. The concept of the LPG controller thermore, the capacitor controls the relay just after the
designed with the AFR data retrieval is presented in Fig. 1. trigger from the angle sensor is sent to the transistor and
the LPG controller becomes activated by adjusting the
sensitivity of the sensor angle. Relay contact points at
3 Results and discussion normally closed terminals are connected to ECU and TPS,
while normally open terminals are connected to ECU and
3.1 Prototyping variable resistors.
AFR controller works based on the position of the
The parts of the AFR controller circuit work in an integrated vehicle. When the vehicle is running in the down-way,
manner, as shown in Fig. 2. For example, the TPS detects the angle sensor produces a signal which triggers the

Fig. 3  Installation of oxygen

sensors (a), clinometer (b),
LPG injectors (c), testing LPG
controllers on down-way (d)

Fig. 4  Block diagram of LPG controller

Vol:.(1234567890)
SN Applied Sciences (2020) 2:37 | [Link] Research Article

Table 3  AFR measurement data 3.2 AFR measurement

Slope angle Speed gear Without AFR With AFR control-
position controller ler The AFR was measured at varying road slopes includ-
AFR Deviation AFR Deviation
ing 10°, 15° and 20°. Therefore, the AFR data on speed
gear 1, 2 and 3 positions without and with LPG control-
10° 1 14.5 ± 0.4 15.4 ± 0.3 ler at a slope angle of 10°, 15°, and 20° are presented in
15° 1 14.6 ± 0.4 15.6 ± 0.3 Table 3. Tests carried out in down-way each of 10 s. For
20° 1 14.6 ± 0.35 15.9 ± 0.3 each road-slope, the vehicle is set to start running at 0 s
10° 2 14.5 ± 0.2 15.4 ± 0.3 and the AFR controller will work at 2 s and end at 9 s
15° 2 14.6 ± 0.4 15.6 ± 0.2 after the vehicle starts. Vehicle speeds are left naturally
20° 2 14.6 ± 0.2 15.9 ± 0.2 based on the level of the road slope. The AFR controller
10° 3 14.4 ± 0.3 15.5 ± 0.3 is intentionally set at the 2 s after start running to ensure
15° 3 14.6 ± 0.2 15.6 ± 0.2 the slope of the vehicle is in the down-way, not at the
20° 3 14.6 ± 0.2 15.9 ± 0.2 speed-bump or when the vehicle is stuck in a road pit.
Therefore, the AFR value after the 9th second is the same
transistor to activate the relay. This, however, reduces the
as the initial value, caused by the vehicle being stopped.
voltage from TPS to ECU through a variable resistor. Fur-
If the vehicle is travelling in a long down-way, a high AFR
thermore, the ECU also reduce the flow of LPG entering
will last as long as the vehicle is still detected tilt. The
the engine through the injectors. Oxygen sensor meters
graphs of the test results are presented in Figs. 5, 6, and
are used to determine the amount of LPG injected as
7, respectively.
shown in Fig. 3a. Finally, the Clinometer installed to
From Table 3 and Figs. 5, 6, and 7, the result shows
measure the angle of inclination is as shown in Fig. 3b.
that the greater slope of the road produced higher AFR.
We use clinometer because it is more practical with high
This indicates that the kinetic force and vehicle gravity
accuracy (maximum deviation is only 0.2° compared to
can be used as input parameters to control the AFR. In
measurements using bevel gauge).
previous studies [31, 32], the AFR controller system in
ECU controls LPG injectors based on information from
LPG vehicles were applied based on decelerations with
sensors such as CMP for engine speed, TPS for throttle
input parameters from engine, brake, and vehicle speed.
valve opening positions, and an AFR meter attached
In this study, the AFR controller developed has the abil-
inside the exhaust manifold to measure mixed quality
ity to work on reduced road conditions at low vehicle
by monitoring the LPG/air mixture entering the engine.
and engine rotation speeds. Based on the data during
The LPG controller block diagram is presented in Fig. 4.
observation, this AFR controller has a great potential
to be applied to modified LPG vehicles that operate in

Fig. 5  AFR profile on various road slope at 1st speed gear Fig. 6  AFR profile on various road slope at 2nd speed gear

Vol.:(0123456789)
Research Article SN Applied Sciences (2020) 2:37 | [Link]

References
1. The Clean Air Initiative for Asian Cities Center (CAI-Asia Center).
Improving Vehicle Fuel Economy in the ASEAN Region (2010).
https​://clean​airas​[Link]/, https​://clean​airas​[Link]/impro​ving-
vehic​le-fuel-econo​my-in-the-asean​-regio​n/. Accessed 10 Oct
2016
2. Santos G (2017) Road transport and ­CO2 emissions: what are the
challenges? Transp Policy 59:71–74
3. Colvile R, Hutchinson E, Mindell J, Warren R (2001) The
transport sector as a source of air pollution. Atmos Environ
35(9):1537–1565
4. Karagiorgis S, Glover K, Collings N (2007) Control challenges in
automotive engine management. Eur J Control 13(2–3):92–104
5. Michalek JJ, Papalambros PY, Skerlos SJ (2004) A study of fuel
efficiency and emission policy impact on optimal vehicle design
decisions. J Mech Des 126(6):1062–1070
6. Tverberg GE (2012) Oil supply limits and the continuing financial
crisis. Energy 37:27–34
7. Messagie M, Lebeau K, Coosemans T, Macharis C, Van Mierlo
J (2013) Environmental and financial evaluation of passenger
vehicle technologies in Belgium. Sustainability (Switzerland)
Fig. 7  AFR profile on various road slope at 3rd speed gear
5(12):5020–5033
8. Deilami S et al (2011) Real-time coordination of plug-in electric
vehicle charging in smart grids to minimize power losses and
mountainous areas, although this only makes a small improve voltage profile. IEEE Trans Smart Grid 2(3):456–467
contribution to areas with the majority of flat roads. 9. Setiawan IC (2019) Policy simulation of electricity-based vehicle
utilization in Indonesia (electrified vehicle—HEV, PHEV, BEV and
FCEV). Automot Exp 2(1):1–8
10. Iliev S (2015) A comparison of ethanol and methanol blend-
4 Conclusion ing with gasoline using a 1-D engine model. Procedia Eng
100:1013–1022
The results showed that the kinetic force, gravity, vehicle 11. Zhai Y-J, Yu D-L (2009) Neural network model-based automo-
weight, and road slope have the potential to be used as tive engine air/fuel ratio control and robustness evaluation. Eng
Appl Artif Intell 22(2):171–180
input signals by the AFR controller to improve fuel effi- 12. Cavina N, Corti E, Moro D (2008) Closed-loop individual cylinder
ciency. The AFR controller developed was able to increase air–fuel ratio control via UEGO signal spectral analysis. IFAC Proc
the AFR value from an average value of 14.5 to 15.5–16.2, Vol 41(2):2049–2056
depending on the down-way angle while the gear position 13. Zhao J, Xu M (2013) Fuel economy optimization of an Atkinson
cycle engine using genetic algorithm. Appl Energy 105:335–348
has no measurable effect. Furthermore, a greater slope of 14. Wu T, Karkoub M, Chen H, Yu W, Her M (2015) Robust track-
the road was observed to have produced greater AFR. In ing observer-based adaptive fuzzy control design for uncer-
conclusion, the AFR controller has the ability to increase tain nonlinear MIMO systems with time delayed states. Inf Sci
AFR and it is very suitable for modified LPG vehicles with 290:86–105
15. Bouarar T, Guelton K, Manamanni N (2010) Robust fuzzy Lyapu-
first and second generation of converter kits that are not nov stabilization for uncertain and disturbed Takagi—Sugeno
yet equipped with lambda sensors, especially for LPG vehi- descriptors. ISA Trans 49(4):447–461
cles operating in mountainous areas. 16. Jansri A, Sooraksa P (2012) Enhanced model and fuzzy strategy
of air to fuel ratio control for spark ignition engines. Comput
Acknowledgements This research is part of an environmentally Math Appl 64(5):922–933
friendly vehicle development project at the Automotive Laboratory 17. Zhai Y, Yu D, Guo H, Yu DL (2010) Engineering applications of
of Universitas Muhammadiyah Magelang. The researchers appreciate artificial intelligence robust air/fuel ratio control with adaptive
the technicians involved in this study. DRNN model and AD tuning. Eng Appl Artif Intell 23(2):283–289
18. Triwiyatno A, Sinuraya EW, Setiawan JD, Munahar S (2016) Smart
controller design of air to fuel ratio (AFR) and brake control sys-
Compliance with ethical standards tem on gasoline engine. In: 2nd International Conference on
Information Technology, Computer, and Electrical Engineering,
Conflict of interest The author(s) declare that there is no conflict of pp 233–238
interest regarding the publication of this article. 19. Masum BM, Masjuki HH, Kalam MA, Palash SM, Habibullah M
(2015) Effect of alcohol-gasoline blends optimization on fuel
properties, performance and emissions of a SI engine. J Clean
Prod 86:230–237
20. Elfasakhany A (2015) Investigations on the effects of etha-
nol–methanol–gasoline blends in a spark-ignition engine:

Vol:.(1234567890)
SN Applied Sciences (2020) 2:37 | [Link] Research Article

performance and emissions analysis. Int J Eng Sci Technol 31. Setiyo S, Munahar M (2017) Modeling of deceleration fuel cut-
18(4):713–719 off for LPG fuelled engine using fuzzy logic controller. Int J Veh
21. Hulwan SVJDB (2018) Multizone model study for DI diesel Struct Syst 9(4):261–265
engine running on diesel ethanol biodiesel blends of high etha- 32. Setiyo M, Munahar S (2017) AFR and fuel cut-off modeling of
nol fraction. Int J Automot Mech Eng 15(3):5451–5467 LPG-fueled engine based on engine, transmission, and brake
22. Morganti KJ, Foong TM, Brear MJ, Da Silva G, Yang Y, Dryer FL system using fuzzy logic controller (FLC). J Mechatron Electr
(2013) The research and motor octane numbers of liquefied Power Veh Technol 8:50–59
petroleum gas (LPG). Fuel 108(2013):797–811 33. Myung CL, Kim J, Choi K, Hwang IG, Park S (2012) Comparative
23. Chikhi S, Boughedaoui M, Kerbachi R, Joumard R (2014) Scien- study of engine control strategies for particulate emissions from
ceDirect On-board measurement of emissions from liquefied direct injection light-duty vehicle fueled with gasoline and liq-
petroleum gas, gasoline and diesel powered passenger cars in uid phase liquefied petroleum gas (LPG). Fuel 94:348–355
Algeria. JES 26(8):1651–1659 34. Jaworski A, Kuszewski H, Lejda K, Ustrzycki A (2016) The effect
24. Cho CP, Kwon OS, Lee YJ (2014) Effects of the sulfur content of of injection timing on the environmental performances of the
liquefied petroleum gas on regulated and unregulated emis- engine fueled by LPG in the liquid phase. In: Lejda K, Woś P (eds)
sions from liquefied petroleum gas vehicle. Fuel 137:328–334 Internal combustion engines system, vol. i, no. tourism. IntechO-
25. Myung C, Choi K, Kim J, Lim Y, Lee J, Park S (2012) “Comparative pen, London, p 13
study of regulated and unregulated toxic emissions character- 35. PradeepBhasker J, Porpatham E (2016) LPG gaseous phase elec-
istics from a spark ignition direct injection light-duty vehicle tronic port injection on performance, emission and combustion
fueled with gasoline and liquid phase LPG (liquefied petroleum characteristics of Lean Burn SI Engine. IOP Conf Ser Earth Envi-
gas). Energy 44(1):189–196 ron Sci 40(1):0–11
26. Dimitris Assanis MSW, Wagnon Scott W (2015) An experimental 36. Mitukiewicz G, Dychto R, Leyko J (2015) Relationship between
study of flame and autoignition interactions of iso-octane and LPG fuel and gasoline injection duration for gasoline direct
air mixtures. Combust Flame 162(4):1214–1224 injection engines. Fuel 153:526–534
27. Raslavi L, Mockus S, Ker N, Starevi M (2014) Liquefied petroleum 37. World LPG Association (2017) Autogas incentive policies, 2017
gas (LPG) as a medium-term option in the transition to sustain- edition, Neuilly-sur-Seine
able fuels and transport. Renew Sustain Energy Rev 32:513–525
28. Kim J, Kim K, Oh S (2016) An assessment of the ultra-lean com- Publisher’s Note Springer Nature remains neutral with regard to
bustion direct-injection LPG (liquefied petroleum gas) engine jurisdictional claims in published maps and institutional affiliations.
for passenger-car applications under the FTP-75 mode. Fuel
Process Technol 154:219–226
29. Van Den Schoor F, Middha P, Van Den Bulck E (2013) Risk analysis
of LPG (liquefied petroleum gas) vehicles in enclosed car parks.
Fire Saf J 57:58–68
30. Erkus B, Karamangil MI, Surmen A (2015) Designing a prototype
LPG injection electronic control unit for a carburetted gasoline
engine. Uludağ Univ J Fac Eng 20(2):141–153

Vol.:(0123456789)