Abstract
There are only a few means whereby man produces the abundance of power with which he dominates his environment, and almost all of them, at present, involve the combustion of fossil fuels. The "energy crisis" is first and foremost a "combustion crisis". If engineers could burn fuels much more efficiently, then mankind's energy shortages would be greatly alleviated. Flame control, or tailoring a flame to its associated combustion chamber (Moles 1974), is the first requisite for efficient combustion, and, over the last decade, the Fuels and Energy Research Group at the University of Surrey have pioneered the application of such techniques to rotary kilns, incinerators, flares and melting furnaces. The second requirement for highly efficient combustion is the use of some kind of flame sensor which can monitor the rapidly fluctuating conditions within an average combustion chamber, to produce signals which can control and then optimise combustion efficiency. The ultimate aim of this ten year research programme is to control the gigantic pulverized coal flames used in the world's power and cement industries. However, for the initial stages of the work described in this thesis, the simpler problem of controlling a gas flame in a purpose designed, pilot plant scale, recuperative furnace was tackled. Although the furnace was constructed within the University's laboratories, it was sufficiently large to be representative of industrial scale practice. Any transducer for combustion plant use has to be universally applicable, relatively simple, robust, reliable, and above all, easily fitted to an existing burner installation. To achieve these criteria, a novel, spherical gross heat flow meter (GHFM) was developed. The water-cooled stainless steel shell was relatively small, 70 mm in diameter, and the unit was designed to generate up to nine control signals for investigatory purposes. Initial experiments on the performance of this spherical heat flux meter were carried out, using a 3 KW laboratory, electric, boxlike muffle furnace and an electrical network analogy was developed to yield a simple theoretical analysis of this system. The effects of different radiative emission levels and thermal masses of the furnace were investigated at this stage and found to be of some significance. It was possible to generate nine signals from the heat flow meter. The signal which was generated by the thermopile of the GHFM's receiving surface was found to be suitable for all modes of operation in furnaces -namely Signal 8. All signals were found to Be suitable for steady-state operation. The principle of the operation of our meter is based on Signal 8. The variation in the generated output millivolts of Signal 8 was dictated by the heat energy movement of the system in a gross manner . A linear relationship was found to govern the gross heat energy movement and the generated millivolts of Signal 8. The measured response of Signal 8 to a step input signal was found to compare very well with its corresponding values, which were predicted from the electrical analogy simple model. This approach was based on gross manners. The range of this response was 2.0 ≤ taug ≤ 6.5 minutes approximately. When the effect of the thermal mass of the furnace was minimized, an average gross time constant of 1.623 minutes (including effects of the thermal delay and conduction loss of the meter), and an average net time constant of 1.005 minutes (excluding thermal delay of our meter) were obtained approximately. The meter was not found suitable for monitoring the generation of heat energy level during the first 37 seconds after its introduction to a furnace nor from a cold start. This is caused by the initial thermal delay of the instrument. It is believed that this delay causes no serious drawback in the performance of the GHFM. (Abstract shortened by ProQuest.).