روش تنظیم احتراق توربین گاز صنعتی با استفاده از تجزیه و تحلیل حساسیت

This paper presents the results of combustion performance testing of a 5.25 MWe industrial gas turbine which features a conical counter-flow double-swirl stabilized, premixed combustor and the Combustion Tuning methodology using a Sensitivity Analysis (abbreviated to CTSA). The combustion performance test was conducted in an atmospheric pressure, optically accessible, real engine scale combustor. The atmospheric rig and real engine correlation was verified by comparing real engine data which were gathered from high pressure tests. NOx and CO emissions, combustor temperature at the fuel nozzle, dump plane and exhaust, dynamic pressure and flame structure, using planer laser induced fluorescence, were investigated with respect to power load and ambient temperature. To enhance the NOx and CO emission performances with stable combustion, the relative sensitivities of five control parameters were analyzed, and on the basis of sensitivity analysis data, combustion tuning testing was conducted. By using the CTSA, NOx emission in exhaust gas was reduced from 18 to 2.2 ppm at base load, with high combustion efficiency (>99.9%), and very little pressure fluctuation (Prms < 0.1 kPa).

In line with the increasing severity of certain environmental problems—including global warming, urban atmospheric pollution, and the depletion of energy resources—many governments, research institutes and businesses are striving to develop high-efficiency, eco-friendly energy resources. Such efforts include developing new and renewable energy resources, utilizing distributed power sources, and improving the efficiency of power plants. Within recent decades, there has been a significant increase in the use of micro or small-scale gas turbines, as well as heavy duty gas turbines due to their relatively low capital cost and low emissions profiles relative to other power generation systems,. The present paper deals with the development of small gas turbines (5.25 MWe), with the intention of utilizing high-efficiency cogeneration systems as a distributed power source. In particular, this paper discusses and analyzes the results from a performance test of a double-swirl combustor for a small gas turbine. Our aim is to improve the efficiency and reliability of the power plant through optimizing combustion. 1.2. Prior research on the tuning methodology of gas turbine combustors In recent decades, due to efficiency, reliability and simplicity, the lean premixed combustion method has become the industry standard to achieve low NOx emission. However since the Dry Low NOx (DLN) or Dry Low Emission (DLE) combustion method based on lean premixed combustion has relatively weak combustion stability, many fatal accidents caused by high combustion vibration were reported in many gas turbines [1]. Thus numerous researchers have studied about the combustion instabilities in gas turbines. Lieuwen et al. reported that the feedback loop among the equivalence ratio fluctuations, heat release oscillations and acoustic oscillations in combustor inlets and fuel lines are responsible for the combustion instabilities of gas turbines [2]. Seo also stated that the incompleteness of premixing is identified as significant perturbation source for inducing unstable combustion [3]. Kulsheimer et al. noted the relationship between vortex formation and Strouhal number, and gave a comprehensive understanding of the formation and reaction of large-scale coherent vortex structures in turbulent flames as drivers of combustion instabilities [4]. To attenuate these combustion instabilities, active, passive or their hybrid control methods have been derived. The active control methods are used to adjust some input parameters such as fuel distributions [5], or to modulate the acoustic boundary conditions of the combustion chamber [6]. On the other hand, passive control methods mainly consist of geometrical modifications of the combustion chamber [7] or sound-absorbing devices such as perforated plates [8] and Helmholtz resonators [9] and [10]. Some of these methods or ideas were applied to real engines and improved their efficiency. Oh et al. suggested a tuning methodology of a GE7FA +e DLN-2.6 gas turbine and reported that the tuning methodology effectively reduced both NOx emission and combustion vibration, which were significantly higher before tuning during the start-up mode [11]. In addition, Johnson et al. presented successful demonstrations of active control of combustion instabilities on a full-scale Siemens-Westinghouse gas turbine combustor [12]. This active control method attenuated the dominant acoustic modes by up to 15 dB and reduced NOx emission by approximately 10%. Hibshman et al. also designed and implemented active control systems on both sub-scale and full-scale combustors [13] and [14]. This control system through modulation of the fuel supply led to a 6.5-dB decrease in magnitude of the dominant instability frequency, while retaining comparable emission characteristics. Afgan et al. raised the concept of an expert control system for fault diagnosis and monitoring of gas turbine combustion chambers and explained the set-up procedure of such an expert system in detail [15]. Kelsall et al. suggested three potential solutions to instabilities, which are: 1) passive damping, 2) active control system based on high frequency actuator and low frequency fuel staging and 3) neural network based control system [16]. While Sen Li et al. also used a passive control method which requires retrofitting a power plant to improve its performance [17], we propose a new combustion tuning methodology of industrial gas turbines by using sensitivity analysis (CTSA) as an active control method which adjusts control parameters without modifying any part of a plant. Though the sensitivity analysis method is applicable for various purposes, it has never been applied to combustion tuning. In the combustion field, sensitivity analysis is usually used to evaluate the chemistry model of combustion calculations [18] and [19]. However, in this study the relative sensitivities of control parameters to combustion performance were analyzed in order to prioritize or rank them as control factors.

We successfully conducted a combustion performance test and tuning of a 5.25 MWe industrial gas turbine which features a conical counter-flow double-swirl combustor. From the results, we obtained the following conclusions: (1) The NOx emission in ambient pressure was judged to be satisfactory, at 25 ppm or less, at output power of 0.6 N or above. However, since NOx emissions at P3 = 1.43 MPa and T4 = 1100 °C were observed to be unsatisfactory and the concentration of CO at lower loads also needed to be improved, the CTSA technology is proposed to optimize combustion performance. (2) Priority in control parameters was decided by sensitivity analysis; from the results, we selected PFR for tuning the combustion. By adjusting PFR, NOx and CO emissions were decreased considerably (87.6% and 25%, respectively). Dynamic pressure and combustion temperature were not considered for tuning because the combustion oscillation was negligible and the nozzle temperature was predominantly controlled by the lift-off air temperature, without creating a localized overheated area. (3) By using CTSA, NOx emission in exhaust gas was reduced from 18 to 2.2 ppm at base load under cold-day conditions, with high combustion efficiency (>99.9%), and very little pressure fluctuation (Prms < 0.1 kPa). This highly efficient combustion tuning methodology was patented and will be applied to the industrial gas turbines. On top of that, this methodology can be provided for the algorithm of automated tuning system. (4) The proposed combustion tuning technique will contribute to further research into improving flame stability based on the development of combustors using multiple swirls and to the application of new fuels, including coal gas, di-methyl ether (DME), and biogas.