Fuel Economy Gains through Dynamic-Skip-Fire in Spark Ignition Engines

Pumping losses are one of the primary energy losses in throttled spark ignition engines. In order to reduce fuel consumption, engine manufacturers are incorporating devices that deactivate the valve-train in some cylinders. In the operating strategies currently implemented in the market, fixed sets of cylinders are deactivated, allowing 2 or 3 operating modes. In contrast, Tula Technology has developed Dynamic Skip Fire (DSF), in which the decision of whether or not to fire a cylinder is decided on a cycle-by-cycle basis. Testing the DSF technology in an independent certified lab on a 2010 GMC Denali, reduces the fuel consumption by 18% on a cycle-average basis, and simultaneously increases the ability to mitigate noise and vibration at objectionable frequencies. This paper outlines the results of the experiments that have been conducted on an eight cylinder engine over a wide range of conditions to investigate the fuel consumption gains and emissions impact when incorporating DSF technology. The experiments have been carried out over a wide range of engine speeds, loads, and DSF strategies and significant improvements have been observed. Introduction Pumping losses are one of the major sources of thermal efficiency losses in spark ignition engines. Cylinder deactivation reduces the pumping losses by deactivating cylinders during every engine cycle based on torque requirements. As a means to improve engine efficiency, cylinder deactivation has a long history. It has been employed by General Motors’s Active Fuel Management (AFM) system in eight cylinder engines [1]. In their approach, four of eight cylinders are deactivated in a 5.3L or 6.2L OHV V8 engine; meaning that a fixed pattern of deactivation is applied and fully implemented in each engine cycle. Other OEMs have also used cylinder deactivation [2-3]. For instance in VW’s 1.4-liter TSI 4cylinder engine, cylinders 2 and 3 are deactivated [2]. Similar to [1], in VW’s engine a fixed pattern of deactivation is implemented and completed in each engine cycle. In this firing pattern, a deactivated cylinder is always followed by a firing cylinder event in each engine cycle. Although firing every other cylinder is the most common approach, some deactivation strategies have used other approaches. For instance, Honda has fired two out of every three cylinders in a 3.5L i-VTEC engine [3]. At Tula Technology, we have introduced Dynamic Skip Fire (DSF) an evolved version of deactivation which is capable of deactivating the cylinders without any limitation. DSF deactivates cylinders in a manner that achieves the load demanded while avoiding objectionable noise and vibration. For example, Figure 1 displays a DSF pattern of 1 fire followed by 2 skips for each cylinder. Figure 1 shows that three cycles are required to achieve this pattern. Figure 1: DSF operation of 1 fire 2 skips for each cylinder in an eight cylinder engine Along with rewarding achievements of DSF, there are also some challenges. Choosing the firing density based only upon fuel economy would induce undesirable NVH characteristics. Each fire induces a torque which creates an acceleration exerted to the crankshaft. The acceleration causes vibrations whose frequency is a function of firing density and firing pattern. Due to the variable nature of DSF in terms of firing frequency, great care should be taken to avoid frequencies in which perception of vibration or resonance is encountered. More details about evaluation and mitigation of NVH in Tula DSF can be found in [5]. One of the benefits of DSF is the wide range of firing pattern selections to minimize the resonance modes. This capability of DSF provides us the opportunity to choose firing patterns which produce surprisingly lower NVH than V8 at equivalent power. Besides NVH considerations, our FTP cycle data for L94 engine (employed in 2010 GMC Denali) show fuel economies of 19.92 mpg at DSF versus 17.34 mpg at V8 mode. Meaning, DSF reduces fuel consumption by 14-18% over V8 [6]. In Tula’s previous publications (e.g. [5, 6]), the fundamental concepts of DSF, have been discussed. In this paper we are presenting a comparative study of DSF advantages versus V8 mode. We will discuss the results in a wide range of operational conditions (listed in