Design of a Suspension Burner System for Forestry and Agricultural Residues (NRI) Monitoring and analysis of data Monitoring Analysis of data

Design of a Suspension Burner System for Forestry and Agricultural Residues (NRI)

Monitoring and analysis of data

Monitoring

Chromel/alumel thermocouples were connected to a multipoint recorder to measure the flue gas temperature. Corrections were made for thermocouple radiation losses. Surface temperatures of the furnace, heat exchanger and cyclone were measured using a hand-held contact thermocouple connected to a digital indicator.

A number of methods were tested to analyse the flue gases, namely:

(i) gas chromatograph analysis for hydrogen, oxygen, carbon dioxide, methane and nitrogen balance;
(ii) infra-red analysis for oxygen, carbon dioxide and carbon monoxide;
(iii) orsat for carbon dioxide; and
(iv) electronic fuel efficiency monitor for oxygen.

Airflow velocity was measured with a rotating vane anemometer, calibrated against a pitot static tube.

Feedstock was weighed and turntable speed monitored to calculate rate of feed.

The flue gases from the exhaust stack were observed visually for signs of smoke. Measurements with a filtered smoke tester were made, but ash collecting on the filter obscured results.

Ambient air temperature, relative humidity and barometric pressure were regularly recorded for reference.

Analysis of data

The above data were considered for heat balance calculations. It was expected that, for given fuel feedrates, the massflow of the flue gases could be derived from: (i) airflow measurements; or (ii) flue gas analysis. However, the calculated massflows obtained by these two methods were inconsistent.

Whilst every effort was made to calibrate air flow measurement under hot and cold furnace conditions, it was not possible to conduct accurate measurements during the operation of the unit because of feedstock and ash particles fouling the instrument. Various contributive effects to explain anomalies in the gas analysis were considered as follows:

(i) small uncombusted particles were oxidizing in the sample lines thus producing localized and erratic results;
(ii) incomplete combustion;
(iii) egress of air into furnace or sample lines thus diluting/contaminating the flue gases;
(iv) accuracy of measuring equipment;
(v) stratification of flue-gases in region where gas sample is taken;
(vi) sample too small for proper representation; and
(vii) fluctuations in pressure in the furnace resulting in a pulsed-type flow caused, in part, by variations in particle size, and the flap valve on the line.

Each of these was investigated during trials with the cyclonic furnace and dual-chamber furnace, but, as no firm conclusion could be drawn from any singular or combined cause, alternative means for calculating flue gas massflow were adopted.

The method adopted was based on temperature measurement and accepted physical property data of the flue gases. A procedure was adopted to calculate flue gas composition, air flow rates and true flue gas temperature from a thermal balance standpoint. To arrive at a solution, the following were considered:

· heat loss from the furnace by natural convection and radiation;
· determination of the inside surface temperature of the flue from considerations of forced convection and conduction through a composite wall;
· heat gain to the flue gas thermocouple by forced convection; and
· radiated heat loss from the flue gas thermocouple to the flue walls.

Flue gas temperature was evaluated from the measurement reading corrected for convected and radiated heat gains and losses from the flue gas thermocouple. Air flow rate and flue gas composition were then calculated from the flue gas temperature, furnace heat losses and fuel flow rate. As some of the variables required to carry out the evaluation were not known initially, an iterative approach was used. The procedure converged in two iterations, with an accuracy compatible with the data used and uncertainties involved in the calculation procedures. To arrive at a solution the following assumptions were made:

· complete combustion of the fuel within the furnace - for this reason the method was adopted for those sawdust trials only where full combustion was observed;
· radiation contributions from the gases in the flue were small - this was a justifiable assumption in view of the complete combustion assumption above;
· conduction losses from the thermocouple stem were insignificant compared to the radiation losses;
· the furnace was surrounded by a perfectly absorbing enclosure at ambient temperature; and
· the inside of the flue behaved as a black body enclosure.