Aquametry in granulate materials

Abstract: On-line aquametry, or water content measurement, is of importance to several
branches of industry. In general, on-line measurements of water content are
needed in order to improve the performance of different algorithms used for
process control. The problem at hand might seem straightforward at first
glance. However, currently there exists no standardized "of the shelf"
solutions for on-line moisture measurements in granulate materials.

The focus of this thesis is on-line measurements of moisture in iron ore
concentrate. The moisture content of the concentrate is very important to the
quality of the manufactured pellets. The water concentration within the
concentrate might vary considerably compared to the water content at the
surface. This makes it difficult to use remote sensing technologies such as
IR, NIR etc. Hence, the long term objective of this project is to develop a
low price sensor that can be embedded into the concentrate and make on-line
measurements. The probe would most likely be an embedded system, preferably
communicating with Radio Frequency Identification technology (RFID). However,
this vision can only be realized, in conjunction with a robust measurement
method.

This thesis investigates two different methods that utilizes fundamentally
different physical properties of the material to detect water: both of which
might be suitable to incorporate into an RFID tag. The properties
investigated are changes in heat conduction and changes in electrical
capacitance of the system as the mositure content varies. The first method is
frequently referred to in literature as the Dual-Probe Heat-Pulse or DPHP and
is based on the principe how heat transfer in a granulate material. The
second method is named Capacitance Probe method (CP) where a parallel plate
capacitor is used to drive a clock circuit. The frequency in the resonant
circuit is determined by the electrical properties of the concentrate, which
in turn is dependent on the moisture content.

In order to thoroughly evaluate the two methods they were first analyzed
theoretically. The purpose of the theoretical study was to optimize the
design of the experimental equipment. With the acquired information a
measurement system was constructed. Some problems with the design did
eventhough occur, the scale of the heat probes were to large, leading to very
long response times an poor signal to noise ratio. To the capacitance probe,
there were problems to spread the concentrate evenly between the capacitor
plates.

Two sets of measurements were performed. First, a number of experiments in
the laboratory, with the water content kept at controlled levels, then to
test the performance of the developed system in an industrial environment a
second set of experiments were conducted at a concentrate silo at the LKAB
pellet plant in Kiruna.

The outcome of the experiments performed in laboratory showed that both
methods could detect water in iron ore concentrate. The standard deviation of
the measurements of the capacitance probe method was 0.35\% (mass percent)
moisture. The Dual-Probe Heat-Pulse method showed random disturbances, but a
clear correlation was detected. Unfortunately the capacitance probe method
measurements at LKAB were inconclusive. The Dual-Probe Heat-Pulse method
could, with measurements of the dry density of the material, show a
correlation between water content and changes in heat conduction. One of the
causes of the inconclusiveness at LKAB are discrepancies in packing density.
When water is added to the material and the material is subjected to
mechanical stress the particles tend to group together, the process is called
agglomeration. This affects the density of the concentrate.

In conclusion, both investigated methods are suitable when measuring moisture
content in concentrate. Minor modifications of the sensor design would most
likely improve the accuracy even more.