Coal Preparation » Process Control
This project has developed and field tested a simple, robust flow rate measuring technique based on signal cross-correlation. It was conducted in two stages. The first, for which a report was published in October 2002, conducted a proof of concept investigation at pilot scale over a wide range of velocity, particle topsize, solids concentration, pipe diameter and pipe orientation. The experimental data showed that the cross correlation flow meter can provide precise (less than 10%) and accurate (5% of mag flow readings) measurements of slurry velocities for diverse feed types at solids concentrations as low as 0.3% w/w.
The second stage developed a robust prototype and carried out on-line testing of a DMC feed to determine the slurry velocity in an operating plant. New robust and environment protected circuits were designed and built. While the circuits built for stage 1 were similar in design they were not suitable for use in the plant environment. The new circuits were designed to give small robust boards which could be mounted in a VersaTainer® ruggedised enclosure. Considerable time and effort went into producing this part of the instrument to ensure a reliable operation in a plant environment. The tuning circuit, which is used to inductively tune the circuit to optimise the signal output from the demodulating circuit, was redesigned from scratch and built for the plant tests to give a high signal to noise ratio in a compact robust enclosure, separate from the generation and demodulation circuits in order to keep it as close to the electrodes as possible for optimum tuning. This new circuit was capable of reducing the system noise by a factor of 10 compared to the old tuning circuit used.
The instrument outputs two analog voltage signals which are collected via a National Instruments PCMCIA data acquisition card operating in a laptop. LabView® software was used to collect and process the signals, which were filtered, passed to a cross correlation function and the velocity determined based on the average time of flight as determined by the cross correlation function and the distance between the two electrodes.
A basalt lined pipe spool piece with a 300mm inside diameter was constructed with flanged ports through which the measurement sensor could be inserted. The measurement sensor consisted of four 316 stainless steel electrodes 0.9mm in diameter set 3mm apart and cast into a wear resistant epoxy to resist abrasion. This sensor was constructed from materials designed to last the duration of the project only. The sensor was inserted into the pipe spool piece until the face of the electrode was flush with the inside basalt pipe lining. Two of the electrode ?wires? were used to transmit the excitation signals and the other two used as measurement electrodes. Data was collected over two trips; the first trip involved testing the hardware and developing the software configuration which may best suit the site. On examining the FFT of the processed signals two close peaks not present in the pilot plant data were found at this plant. During the first trip in June these peaks were quite small compared to the signals generated by the cross correlation events, however during the August trip these peaks were large compared to the signal of interest and careful selection of band pass filters were needed to ensure that the signals did not interfere with the measured cross correlation signals found at the lower end of the frequency spectrum.
During the two months between trips the sensor was left in the pipeline to determine the effect of wear on the signal, examination of the wear resist tip showed some wear of the matrix material around the wear resist particles. The measured cross correlation signal strength was reduced by a factor of four however the signal was still sufficiently strong to produce credible results albeit with a slightly higher raw data standard deviation then the June results. It should be noted that this probe was a prototype and was designed to last the length of the test period only. Materials selection would considerable decrease the rate of wear on this probe particularly if the electrodes were embedded in a ceramic or basalt material.
Velocity data collected during the second trip was compared to calculated flow rates based on the feed pressure and cyclone nominal feed rate. No correlations were found in the data variations between the two measurements or in spikes or trends. These variations in the order of 5% to 10% of the flow velocity to the DMC was unexpected and has obvious implications to the operation of the DMC. The velocity measurements were taken at a point 4m below the cyclone inlet from where the pressure measurements were made. The pressure measurement therefore may be influenced by different factors which influence the cross correlation flow meter. Calculations of the medium to coal ratio at the same point in the DMC feed line using EIS technology(Hu and Firth 2005) found that the coal to medium ratio was also varying in a similar manner to the cross correlation velocity measurements indicating a potential solids surge effect. The cross correlation flow meter will measure particle events crossing the two electrodes while the pressure cell measures an average head or energy not related to the particles within the slurry. For this reason it is not unexpected that the pressure and cross correlation readings while measuring similar velocities would not necessarily have a high correlation with each other in terms of spikes and trends.
This reports shows that a direct, reliable and cost effective method for the measurement of pipe slurry velocity in coal preparation plants can be achieved. This method is applicable to a broad range of slurry types such as tailings, flotation feed and DMC feed. This system could easily be adapted to connect to the plant control system in either a hardwired format or wireless giving an analog signal proportional to the velocity in the pipeline. The method is not inferred from other properties but is a direct measurement of the transit time of resistivity fluctuations between two electrode pairs referenced to the computer clock, and as such is a primary calibration source.