Tony Wheatley continues his examination of the global glass industry with an insight into the technology at the heart of glass production – furnaces.

Glass is described as a uniform amorphous solid material that results when the viscous molten raw material cools to below its glass transition temperature.

Raw materials for glassmaking are delivered to the plant as fine grained materials that are carefully blended into a batch before entering the first stage of the process – melting.

Although incremental improvements have been made to combustion systems, regenerators, refractory and batch handling that have extended furnace campaigns, decreased their size and increased energy efficiency, at the heart of the process is the regenerative melting furnace which remains essentially unchanged.

Regenerative furnaces
These enormous (over 100m2), capital intensive, energy guzzling and inflexible melting machines have served the glass industry for the past century, because ironically they still provide the most economical way to heat bulk materials to temperatures as high as 1650°C.

Fuelled with coal gas, pulverised coal or in recent years by natural gas, they belch out carbon dioxide in volumes that approach the mass of their output capacity while operating continuously for 11 to 15 years. The only energy optimising feature of their design is the incorporation of regenerators to absorb wasted energy from the flue gas and transfer it to incoming flows of cold material.

The position of the burners in conventional regenerative furnaces is their major disadvantage, because heat transfer is primarily by radiation on the surface of the batch and this is ineffective.

Theoretical calculations indicate that the specific energy required to melt glass should be around 2.56 GJ/tonne, but in practice when fired with conventional natural gas/air burners, the typical energy consumption of a regenerative furnace averages nearly double this value at around 5.00 GJ/tonne.

The massive waste of energy is attributed to system losses through furnace leaks, worn refractory and poor insulation, but the core of the problem is that only about 20% of the energy consumed is transferred from the burner flame to the charge. Over 30% of the energy is fortunately recovered by the regenerators, but the losses through the furnace walls and flue stack often exceed 40% of the primary energy input.

Process optimisation
Starting in the 1970s the idea of raising furnace performance by using pure oxygen to reduce the volume of air entering the furnace was tested exhaustively.

Nitrogen, which constitutes 78% of air, increases the gas flow through the furnace and absorbs considerable heat energy that ultimately escapes via the flue stack. Initially oxygen was introduced using auxiliary burners to direct energy into cold spots and production increases between 10% and 30% were achieved, leading to modestly improved specific fuel consumption.

Later oxy-fuel burners replaced some or all of the original burners and many configurations of burner and port position were evaluated. Burner technology proliferated, allowing oxy-fuel flames to be manipulated into specific shapes and lengths to suit many variations of fuel and furnace geometry.

During the 1990s full-scale oxy-fuel firing of large glass melting furnaces was tested with encouraging results. A secondary benefit that derives from replacing air with oxygen is reduced environmental pollution. Low NOx oxy-fuel burners reduce emissions substantially but depend on good furnace sealing to prevent air ingress.

SO2 emissions are around 50% lower with oxy-fuel technology, because the inevitably increased level of dissolved water in an oxy-fuel fired furnace enables effective fining reactions to be achieved with reduced sulphate content in the batch.

Oxy-fuel firing has also proved effective in reducing particulate emissions by over 50% where this is associated primarily with volatilisation of NaOH and the formation of Na2SO4. The reduced mass transfer rate of NaOH from the molten glass surface was attributed to lower gas velocities characteristic of oxy-fuel firing.

The fuel efficiency of oxy-fuel fired furnaces can be further improved by installing waste heat recovery systems like cullet pre-heating and waste heat boilers. The potential performance of a fully optimised regenerative glass furnace today is projected to be around 2.60 GJ/tonne, but often the capital investment to achieve this on existing plants is unjustifiable.

Submerged combustion melting (SCM)
Submerged combustion was developed for melting minerals by firing oxy-fuel burners directly into the batch of materials from underneath the melt chamber.

The combustion reactions take place inside the melt and the resulting large surface area of direct contact facilitates rapid heat transfer. The process tolerates wide variation in batch material sizes, multiple feeds and imperfect blending, so batch handling system can be greatly simplified.

Turbulent mixing occurs in the melting bath and this accelerates melting and chemical reaction rates, while homogenizing the molten product.

SCM was developed by the Gas Institute of the National Academy of Sciences of the Ukraine (GTI) and has been used there commercially for mineral wool production over the past 10 years.

A US Department of Energy sponsored project, initiated in 2006 in partnership with a consortium of glass manufacturers to design, demonstrate and validate the melting stage of a next generation glassmaking system, is still ongoing.

This development was recently referred to by Michael Greenman, Executive Director of the Glass Manufacturing Industry Council, as “what could be the most important new glass technology since the Pilkington float method.”

He summarised its features as:
* Low capital cost – 60% less than a traditional furnace of comparable capacity
* Small footprint – six to eight times the pull rate for the equivalent area
* High efficiency – 205 times better than tradition furnaces with low emission levels
* Very flexible in operation – start-up to full melt in just four hours, cool down the same duration

The first commercial application of SCM is in operation in Indiana in the US and several other prospects are under discussion including the production of sodium silicate. The final project phase involves developing refining techniques to suit all current glass applications.