Virtual sensors for spark ignition engines using neural networks

Hanzevack, E.L.; Long, Ting; Atkinson, Christopher M.; Traver, Michael

doi:10.1109/acc.1997.611885

Cited by 26 publications

(17 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the many attempts to estimate AFR using NNs the overwhelming number present results only on the 'hold-out' data set [20,16,21,15], and no final experimental validation data are presented. The absence of final experimental engine validation highlights the difficulties in establishing effective online NN systems.…”

Section: Article In Pressmentioning

confidence: 98%

“…This use of NNs for AFR prediction has now been extensively researched [15,13,16,12,6,17]. Networks sizes and complexities have been determined through trial and error and estimation errors as low as 2% for 90% of test transients have been achieved [18].…”

Section: Article In Pressmentioning

confidence: 99%

“…Results using the UEGO sensor data which are not experimentally validated, since they are presented in simulation only, show a robust AFR virtual sensor accurate to within 1:2%. Earlier work attempted a virtual sensor to find the stoichiometric index, l, for a general aviation aircraft SI engine [15]. With l as an output, oil temperature, MAP and throttle have been used as inputs to simple two layer NNs.…”

Section: Article In Pressmentioning

confidence: 99%

See 2 more Smart Citations

A transient virtual-AFR sensor using the in-cylinder ion current signal

Rivara¹,

Dickinson²,

Shenton³

2009

Mechanical Systems and Signal Processing

View full text Add to dashboard Cite

Section: Article In Pressmentioning

confidence: 98%

Section: Article In Pressmentioning

confidence: 99%

Section: Article In Pressmentioning

confidence: 99%

See 1 more Smart Citation

A transient virtual-AFR sensor using the in-cylinder ion current signal

Rivara¹,

Dickinson²,

Shenton³

2009

Mechanical Systems and Signal Processing

View full text Add to dashboard Cite

“…Most frequently applied are linear observer-based techniques [4], [11], parity-space methods [5], [12]- [14], and parameter-estimation schemes [6], [15]. The use of nonlinear virtual sensors in aerospace applications has not been extensively investigated yet, although this technique has been successfully applied in other domains like process control and engine control [16], [17].…”

Section: A Analytical Redundancymentioning

confidence: 99%

Soft computing applications in aircraft sensor management and flight control law reconfiguration

Oosterom

Babuška

Verbruggen

2002

IEEE Trans. Syst., Man, Cybern. C

View full text Add to dashboard Cite

Abstract-A sensor management system based on soft computing techniques has been developed and implemented in the flight control system of a small commercial aircraft. Unlike in the conventional sensor management system, the signals from sensors are assigned weights based on fuzzy membership functions and the consolidated signal is computed as a weighted average. This approach improves the quality of the consolidated signal and reduces transients due to sensor failures. This soft voting is extended to soft flight control law reconfiguration. In addition, a virtual sensor has been introduced as an arbitrator which enables the isolation of the failed sensor in the duplex operation and the detection of a sensor failure in the simplex operation. The effectiveness of the proposed methods is demonstrated by using an extensive simulation model of a small commercial aircraft, developed by airframe and control system manufacturers on the basis of an existing business jet. Furthermore, the system has been successfully evaluated and compared to standard techniques by means of pilot-in-the-loop simulations on the Research Flight Simulator of the National Aerospace Laboratory in The Netherlands. This application, developed within a Brite/EuRam research project, is characterized by the effective combination of novel soft computing techniques with standard, well proven methods of the aircraft industry. The properties of the conventional sensor management system have been retained, with the additional advantage that the quality of the consolidated signal is improved, the failure-induced transients are reduced, and the consolidated signal remains available up to the last valid sensor.

show abstract

“…The AFR estimation is used to calculate the fuel in AFR control, which results in AFR = 14.7: 1 ± 0.5. In another study, it is presented AFR virtual sensor-based neural network (NN) [11]. Results of the study are the observer open-loop with feedback sensors on SI engine to generate a new variable, namely Power (kW), fuel pressure (kPa), lambda (0.3--1.3, rms: 0.0232), HC (g/s), CO (g/s), CO2 (g/s) and NOx (g/s).…”

Section: Iatmentioning

confidence: 99%

Parameter Identification of Nonlinear System on Combustion Engine Based MVEM using PEM

Badmianto

Firmansyah

Cahyadi

2018

IJITEE

View full text Add to dashboard Cite

Abstract-In four-stroke engine injection system, often called spark ignition (SI) engine, the air-fuel ratio (AFR) is taken from the measurement of lambda sensor in the exhaust. This sensor does not directly describe how much AFR in the combustion chamber due to the large transport delay. Therefore, the lambda sensor is used only as a feedback in AFR control "correction", not as the "main" control. The purpose of this research is to identify the parameters of the non-linear system in SI engines to produce AFR estimator. The AFR estimator is expected to be used as a feedback of the main "AFR" control system. The process of identifying the parameters using the Gauss-Newton method, due to its rapid computation to Achieve convergence, is based on prediction error minimization (PEM). The models of AFR estimator is an open-loop system without a universal exhaust gas oxygen (UEGO) sensors as feedback, called a virtual AFR sensor. The high price of UEGO sensors makes the virtual AFR sensor can be a practical solution to be applied in AFR control. The model in this research is based on the mean value engine models (MVEM) with some modifications. The research dataset was taken from a Hyundai Verna 2002 with the additional UEGO type of lambda sensors. The throttle opening angle (input) is played by stepping on the gas pedal and the signal to the injector (input) is set to a certain quantity to produce the AFR (output) value read by the UEGO sensor. This research produces an open loop estimator model or AFR virtual sensors with normalized root mean square error (NRMSE) = 0.06831 = 6.831%. (model) is required to obtain an accurate mathematical model. A good model is a model that more accurately reflects the dynamics of the system or plant. The dynamics of air-fuel ratio (AFR) on SI injection system Engine is very non-linear and multivariable [1], makes the process of the parameter identification difficult compared with the linear system and a few variables. Keywords--ParameterThe AFR is an important indicator that affects exhaust emissions, engine power, and fuel consumption in Spark Ignition (SI) Engine. The influence of lambda (normalized AFR) on power and fuel consumption are presented in Fig. 1. Maximum power will be achieved if the lambda value is in the range of 0.82 (AFR = 12:1) [2], while the fuel consumption will be achieved if the minimum number lambda is in the range of 1:09 (AFR = 16:1) [2].

show abstract

Virtual sensors for spark ignition engines using neural networks

Cited by 26 publications

References 3 publications

A transient virtual-AFR sensor using the in-cylinder ion current signal

A transient virtual-AFR sensor using the in-cylinder ion current signal

Soft computing applications in aircraft sensor management and flight control law reconfiguration

Parameter Identification of Nonlinear System on Combustion Engine Based MVEM using PEM

Contact Info

Product

Resources

About