Abstract
To ensure that the engines to be used in LNG-fueled vehicles are matched with the expected variations in fuel composition, the knock resistance of the fuel must be determined unambiguously. Rather than rely on empirical methods using gas mixtures and “standard” engines traditionally employed for this purpose, we have derived a method based on the combustion properties of the fuel mixtures, and have developed a first-generation calculation tool for LNG fuels.
Engine knock is characterized by autoignition of the unburned fuel mixture, the so-called end gas, ahead of the propagating flame in the engine cylinder. The core of the new method described in this paper is the computation of the autoignition process during the compression and burn periods of the engine cycle. The method’s detailed chemical mechanism describing autoignition has been tested against experimentally determined autoignition delay times of the alkanes up to pentane (including the isomers of butane and pentane), hydrogen, carbon monoxide and carbon dioxide, measured in our Rapid Compression Machine (RCM). In addition to the effects on autoignition itself, the effects of fuel composition on the in-cylinder pressure and temperature conditions relevant for knocking, such as changes in heat capacity of the air-fuel mixture and changes in the phasing of the combustion process are also incorporated in the method.
We first report the testing and optimization of the method’s chemical model, based on measurements of the autoignition delay time in DNV GL’s RCM. We then present the method to predict the pressure and temperature history of the end gas upon changes in gas composition, using a two-zone thermodynamic model. Combining both methods into an “integrated” model, we derive a propane-based scale (Propane Knock Index, PKI) for ranking the knock resistance of different gases for DNV GL’s lean-burn, spark-ignited gas engine. We demonstrate that the prediction of autoignition of the end gas using the integrated model is adequate for predicting knock with varying fuel compositions by comparing the computed PKI with the Knock-Limited Spark Timing measured in the DNV GL engine, for hydrocarbon mixtures relevant for LNG and natural gas. Further, we compare the results with those obtained using a methane number method. We also report a first-generation algorithm, an easy-to-use, gas-input-only tool to predict the knock resistance for LNG fuels for our engine.
The tool is based explicitly on the performance characteristics of our engine and the fuels studied thus far. The extension of the tool to the LNG-for-transportation market will require analyses of other engine types, such as stoichiometric and dual-fuel engines.
While traditional knock resistance ranking tools rely on empirical methods using a single “standard” engine, our method is based on the unique approach of using the elementary physical and chemical processes to describe the macroscopic behavior of engines. This paper presents the first application of the method in a quantitative ranking tool, the PKI, and verifies its predictions against knock experiments in a practical engine.
Thus, a new, physically “correct” method for characterizing the knock resistance of gaseous fuels based on the physical and chemical properties of the fuel, and their effects on in-cylinder processes, has been developed and its first application in a calculation tool for ranking the knock resistance of LNG fuels is presented. The methodology being developed is readily extended to different engine types and fuels, and is therefore flexible (adaptable to several applications) and sustainable (able to develop together with new engine technologies).
KEYWORDS – Knock resistance; Autoignition; Non-empirical method; LNG; Transportation market