15 - 19
On a Test Arrangement
of a Glow-Discharge
Probe for Fluid Bed Diagnostics
L. EGIZIANO. Dipartimento
G. FABRICATORE Elettrico
and B. hlACCIIIAROL1 Naples
in revised form September
29, 1983) --.--
A glow-discharge probe has b‘een developed for measuring bubble-phase parameters of gasfluidized beds. A low-cost, multichannel electronics has been set up and preliminary tests for probe calibration have been carried out. Probe performances appear to be comparable with those of capacitance probes, while its disturbing effect on the bubble flow pattern is at the minimum level for intrusive ?neasuring techniques, due to the slender, needle-shaped detecting electrode.
The use of intrusive probes for the measurement of local density of gas-fluidized beds has long been recognized as a valuable technique in analyzing fluid dynamics of such systems. In particular, capacitance probes have been widely employed with various geometries and sizes of the electrodes [l - 43 _ More recently, optical probes have also been developed [ 5]_ The question arises, in the use of local probes of any type, whether they significantly perturb the flow conditions, with particular reference to the bubble phase properties_ Recently, Rowe and Masson  analyzed the disturbing effect of various probes on the bubble flow pattern and concluded that the probe stem tip must be as slender as possible in order to minimize the disturbance. Even if needle-shaped capacitance probes have been developed (71, their size cannot be reduced below certain limits, due to the increasing noise/signal ratio. Also, the size of the 0032~5910/84/$3_00
tip of optical probes cannot be reduced indefinitely. A different technique has been developed in recent years, based on the use of needleshaped probes, with diameters of the tip down to less than one millimetre IS. 9). This measurement technique is based on highvoltage glow discharges arising from sharpedge bodies due to dielectric inhomogeneities. A needle, placed vertically in the bed. is connected to a high a-c. voltage source_ The other electrode is a metal plate, placed along the bed wall. When a bubble surrounds the needle tip, glow discharges start and the discharge current can be detected and recorded_ The interaction between the needle and the bubble is at the minimum level with intrusive methods. The drawback of this method, when it x-as developed, was the high cost of the signalrecording electronics, based on ERA discharge detectors, as well as the difficulty of obtaining a multichannel device_ In this paper, a development of this probe is presented, with very effective, low-cost. multichannel electronics. Preliminary tests have shown that the measuring set is able to detect small bubbles, down to 1 cm in diameter, with high time resolution_
In Fig_ 1, a block diagram of the esperimental apparatus is repcrted. The probe is a sharp steel needle having a diameter of OS mm and a length of 200 mm, supported by a stem, 3 mm OD, which also acts as an electric wire. The stem is fised in the bed by means of 0 Elsevier SequoialPrinted in The Setherlands
Fig. 1. Block diagram of the experimental apparatus_ BIA, bubble injetction apparatus; A, amplifier; CU. control unit; LC, limiticg circuit; HVT, high-voltage transformer; OA, output adaptor; IU, input unil :; R, recorder_
1 Fig. 2_ Detecting
supports. The probe is connected to a suitable measurement voltage transformer, regulated by a continuously variable autotransformer to obtain on the probe a voltage of about 6 - 7 kV. The receiving electrode, which is a steel plate placed at the fluidized bed wall, is connected to the detecting circuit. When a ‘bubble’ crosses the needle, a glow discharge occurs in ax. systems; a train of voltage pulses is present and, in order to detect this train, the electronics has to perform the following features: a) large-band, high-gain amplification (>60 dB); b) output at low impedance, in order to feed directly a galvanometric recorder, if necessary; c) overvoltage protection, by means of a liiiting c:zcuit; d) accurate grounding of the circuitry. The scheme of the detecting device is presented in Fig. 2_ A low-noise IC dual insulated
amplifier is used (LM 381 or similar), with open-loop high gain (112 dB) and wide range of supply voltage (9 - 40 V). The bandwidth at unit gain is 15 MHz and the amplifier is completely compensated and protected against short circuits_ In the single chip of the IC, each of the two amplifiers is completely independent within the limit of a separation of 60 dB between the two channels. This allows to set up a two-channel base unit. The gain is continuously adjustable by regulating the resistor R i, while by varying the capacitance of the capacitors C1 and C, it is possible to regulate the values of the upper and lower comer frequencies. The rejection is about 100 dB at a frequency of 100 Hz_ The output adapter has been designed in order to supply a high-current (200 mA) output signal on a load of 12 fl with the following
characteristicsr bandwidth of 10 - 15 MHz, voltage gain of O-99 and current gain of 30000. These performances remain almost the same, irrespective of the qualitative level of the components. A simple but effective limiting circuit, consisting of 4 silicon diodes. with a clipping voltage of about 1.5 V, performs the protection of the galvanometers_
Two main problems must be solved in order to give physical meaning to the voltage pulse trains recorded. They are: a) possible delays, of the signal recorded with respect to the bubble passage; b) low signal/noise ratio. -4s far as point a) is concerned, it is known that glow discharges can start with a certain delay on the instant when the bubble meets the needle_ As mentioned in a previous paper , this delay varies, as a function of the breakdown voltage of the dielectric, between l/20 and l/3 of the period of the applied voltage. As the breakdown voltage of the fluidized bed is not perfectly defined, the necessity of calibration tests arises to assure a correct time response of the probe. For calibration purpose only, a twodimensional bed with transparent walls has been set up_ Its cross-section is 12 X 250 mm, its height 700 mm. Different solids, like silica sand, glass ballotini and alumina, have been used as test materials_ The probe is mounted vertically from the top of the bed, its tip being at 20 cm from the distributor_ As to the distance of the metallic receiving plate from the needle tip, this did not prove to be a critical parameter_ As the glow discharges occur when a bubble is very close to the needle tip (less than 1 mm), the overall performance of the probe apparatus is not affected by the distance between the electrodes; also, when the needle is placed far from the plate, a high signal/noise ratio can be achieved by means of a limited regulation of the voltage. A movie camera, operated at 50 fps, is used in order to follow the bubble motion. Single bubbles are injected from a nozzle. Synchronization of the film with the electric signal is
performed by the lighting of a lamp, in correspondence with the bubble injection_ The same pulse of the lamp is also fed to the oscillographic recorder used for the probe signal. A typical film sequence compared with the recording of the electric signal is presented in Fig. 3. Comparisons made for 50 bubbles have shown that the relative time error of the glowdischarge probe with respect to optical detection has an average value of 5%. The standard deviation of the relative error is O-1. The signal/noise ratio is, in general, high. Further improvement of the signal amplitude could be obtained by applying a higher voltage; but this is only true to a limited extent, since the noise generation is related to the amplitude of the applied voltage_ The proper choice of this voltage must be performed taking into consideration the solids characteristics and ‘;he fluidization conditions of the bed as well as the distance between the needle and the detecting electrode_ For the calibration esperimental set, optimum values have been found to lie between 6 and 7 kV_ Experiments on the multichannel device have been also carried out,‘using a three-dimensional bed. 15 cm ID, with two probes inserted horizontally, with tips vertically aligned at a distance varying between 5 and 20 cm. The purpose of these esperiments was to verify to what estent the two channels can interfere one with the other. The possible interference phenomena are the following: a) electromagnetic coupling betxeen the probes; b) the possibility that the glowdischarge phenomena from a needle can invest both receiving electrodes_ These problems have been completely solved by accurately shielding the detecting circuitry as well as by determining esperimentally the optimum spacing between the probes_ In the geometry tested, the spacing was of the order of about 10 cm.
..t the end of the set-up stage of the esperimental device, the following conclusions can be drawn: - the measuring principle appears to be
Fig. 3. Examp!e of comparison between optical and electrical recording of glow-discharge gram is labelled with its sequence number and the corresponding time value.
sound, and its performance is comparable with that of capacitance probes; - the sizes of the needle probes are smaller than other intrusive probes, and can be further reduced. The level of fluid-dynamic disturbances should be held at the minimum value among the types of probes proposed in last years; - the circuitry is inexpensive and very reliable, even if built with components not expressely selected for the purpose, and the multiplication of the measuring channels can be performed as desired at a very low cost. Tests on huger scale apparatus wilI be performed in order to define the application limits of this technique.
REFERENCES 1 R. D. Morse and C. 0. Ballou, Chem. Eng. Z’wp-. 47 (1951) 199. 2 K. P_ Lameau, Trans. Inst. Chem. Engrs.. 35 (1960) 12.5. 3 S. Crescittlii. A_ Ginnasi, L. Massimilla and G. Maviglia, La Chimicn e Z’Industri&. 55 (19i3) 9_ -1 J_ Werther and O_ %Iolerus, Int. -J_ _IfuZtiphase FZoom. 1 (1973) 122. 5 D. L. Fenton and J. J_ Stukel, Int. -1. Jfultiphase Flow. 3 (1976) 141. 6 P. N_ Rowe and H_ Masson, Chem. Eng. ScL, 35 (1980) 1113. i J. Werther, Trans. Inst. Chcm. Engrs.. -53 (197-I) l-%9_ 8 F. Gasparini and B. Macchiaroli. ISH 1972. Miinich, F.R.G. inp. 9 S. Crescitelli. L. Egiziano and B. Ilacchiaroli. Chim. Ital.. IO (19F-I) 23.