A dual-junction thermocouple probe for compensated temperature measurement in reacting flows
This project was conducted in an effort to develop an inexpensive and reliable means of acquiring spatially and temporally resolved temperature data in turbulent reacting flows. Such a system would allow for increased understanding of turbulent reacting flows without the need for a costly and complex optical system. The system under study uses the responses of two thermocouples of different sizes placed in close proximity in order to determine instantaneous temperatures. This is in contrast to previous work in which compensation is performed on a single thermocouple junction in order to correct for the error caused by its heat capacity. The compensation technique that was developed did not require knowledge of the physical properties of the flow or the physical properties of the thermocouples. It did require the measured junction temperatures, the temperature gradients, and the ratio of the time constants of each thermocouple, which is related to the ratio of the diameters of the two junctions.
Computer models were used to demonstrate the compensation technique itself and were used to show how this method is affected by such factors as diameter ratio, noise, size of the junctions, and the digital resolution of the voltage-to-temperature conversion. Dual junction probes were constructed and tested in non-reactive and reactive environments. Non-reactive experiments were used to calibrate the probe diameter ratio and compensation of thermocouples that were heated with a laser and then cooled appeared successful, with errors of 5% or less in the corrected temperatures. Data was taken in the exhaust duct of a step combustor and the compensated temperatures from this turbulent, combusting environment appeared realistic. Some non-physical temperatures were produced which resulted in the elimination of around 37% of the total data set. Non-physical temperatures were inconclusively attributed to a combination of spatial separation of the thermocouples, conduction losses, and to poor response of the junctions due to their size and heat capacity. Best results were obtained when the thermocouples were exposed to a non-reacting jet of heated air. In this situation, the response amplitude of the thermocouples was relatively large and the response frequency relatively low in comparison to the reacting experiments. In this case, the corrected temperature curve appeared to be physically realistic and properly in phase with the thermocouple signals. Around 10% of the data was discarded using error elimination techniques. It was decided that a workable system which was limited by the size and spatial separation of the thermocouples had been achieved.