D. A. Jarvis
Glass Industry Refractories and their Increasing Role in Energy Conservation
Thursday, February 21, 2019
Current analyst’s forecasts for global glass production in 2020 is for up to 120 Mt of end product with a value approaching USD 90 billion. These same estimates also indicate that the overall market size is once again exhibiting small growth. It should be noted however that at the time of preparation of this report stock markets and some major economies around the world were showing signs of some volatility from political as well as economic issues and so these forecasts may not be fully met. The entire population of Great Britain however should be commended as they most definitely appear to be fully committed to do-ing their best to increase glass production and use in its region.
Current analyst’s forecasts for global glass production in 2020 is for up to 120 Mt of end product with a value approaching USD 90 billion. These same estimates also indicate that the overall market size is once again exhibiting small growth. It should be noted however that at the time of preparation of this report stock markets and some major economies around the world were showing signs of some volatility from political as well as economic issues and so these forecasts may not be fully met. The entire population of Great Britain however should be commended as they most definitely appear to be fully committed to do-ing their best to increase glass production and use in its region. A recent study reported by The Econo mist from information first published in the British Medical Journal indicates that the size of wine glasses in Great Britain has increased from about 100 ml in the year, 1700 ml to about 450 ml today, with 50 % of the increase having occurred within the last 20 years.
Significant glass production in fact takes place in most industrialised countries around the world and Europe accounts for about a third of the total production. In Europe, total glass production peaked at about 35 Mt just after the millennium and then fell back by as much as 20 % although it is now recovering almost to the previous levels. Statistics show that the breakdown of product types is slightly different in other areas of the world and greatly different by country. The overall typical production with information derived from multiple sources is as shown in Tab. 1.
The European Economic and Social Com-mittee Report on Glass Industries pub-lished in April 2015, which of course is now slightly out of date, indicates that no new container glass plants were planned at that time while 13 new production units were announced outside the EU. In the flat glass sector, 12 out of the previous 62 plants had stopped while 9 new sites opened at EU borders. The fibre glass sector was worst af-fected with 1 out of 9 producers shut down, while in tableware and special glasses there had been serious competition from a high level of imports from Asia and Middle East. The Glass Alliance indicates that within Europe recently container glass comprised about 62 % of the glass produced while flat glass made up 29 % of the market, and other types of glass together the remaining 9 % of the total glass production. Members` estimates indicate that while tableware, such as large wineglasses, grew by 6 % the container, and flat glass sectors only grew by 2 %, and some special glasses declined by 10 % (Fig. 1).
According to an International Labour Or-ganisation report, Europe still accounts for about a third of the world’s total glass production and consumption in roughly all of the above types. At its peak, there were over 600 furnaces of all types and sizes pro-ducing a wide range of glasses in Europe. About 37,5 % of these were end-fired tank furnaces, and 24,0 % of these cross-fired tank furnaces, but since these were gener-ally much bigger, 47 % of all glass tonnage was produced from this type of unit and 38 % from end-fired tanks with the balance mainly special glasses such as crystal glass made in pot type units.
Today, there are generally fewer larger units in operation and this of course is due to economics and globalisation as much as from the many technical considerations. If readers will forgive, the pun the glass indus-try was and remains in a state of flux.
In Great Britain, old container furnaces were closed in the east of the country and new larger container furnaces were opened in the west. The two new furnaces built were designed to pull 600 t glass per day but are capable of producing 800 t glass per day. More new furnace capacity is now said to be planned. The plant in Cheshire which was originally Irish owned and Euro-pean funded is now part of a group head-quartered in Spain. The technology and logistic capabil ities of the plant are such that containers of wine arrive in through one gate while bottles of wine in cases on shrink wrapped pallets roll out of the other. Ironically, float glass production in GB has partly moved from west to east with the major original British producer, now J apanese owned, and older British furnaces closing with new ones being built closer to the European market in the east. These new furnaces are owned and operated by a USA-based company oper ating worldwide in float glass production.
As can be seen, glass production around the world currently originates in many different types and sizes of furnace from 1000 t/d day float glass tanks through 500 t/d con-tainer glass tanks down to 1 t/d special glass batch furnaces.
Similarly, there are a vast range of refrac-tories product types used in the construc-tion and maintenance of glass furnaces. Pilkington Glass Technical Centre, now part of NSG, stated that in some of its new furnaces they often required over a thou-sand tons of various refractories at a cost in excess of a million pounds Sterling, plus of course the associated installation costs. A recent calculation made by RHI Magnesita indicated that 4–5 kg of refractories were consumed in the production of 1 t of glass. It should be noted however that there is an enormous range of consumptions of actual refractories recorded in specific areas of dif-ferent furnace designs in the manufacture of a wide variety and types of glasses. It is also necessary to differentiate between new builds, rebuilds and mid-term maintenance. Estimates of refractory materials consumed annually in the glass industry worldwide vary widely from just under 5 Mt to almost 8 Mt depending on which source is con-sulted and obviously also depending on the level of new furnace capacity constructed in any given year. No matter from what per-spective the glass industry is viewed, how-ever it is certainly a large and also a cur-rently expanding market for a wide range of refractories products and services and should certainly be viewed as providing good long term potential.
While this is the case, it is almost certainly also true however that the average con-sumption of refractories per ton of glass produced has halved in the last 25 years due to purer raw materials higher quality products more skilled installation as well as many operational factors. This has resulted in claimed lives of up to 15 years per cam-paign for a few furnaces although most are still significantly less than this. The thrust from furnace designers, refractories sup-pliers and end users has been not only to extend the life of older designs of furnaces but to get much better refractory per form-ance from furnaces with new cutting edge design and also to save energy.
One estimate of the energy required to melt and condition a ton of glass is that it re-quires an average of 8 GJ/t, but once again, it must be emphasised that this v aries widely. This has only more recently been achieved and shows a great reduction from previously much higher energy require-ments due mainly to furnace design and operational factors rather than to the re-fractories. Factors such as preheated batch, the increased use of cullet in the batch and various degrees of oxygen enrichment of the fuel have all played a part in greatly increased efficiency. It has been estimated that energy costs reduce between 2–3 % for every 10 % of cullet in the batch. There is a limit to how much suitable cullet can be collected, recycled and used however. It appears incontrovertible that there is still much greater scope to reduce energy con-sumption arising from better refractories materials and engineering than maybe from any other option currently available to glassmakers.
Currently, since most modern glass furnaces are fairly large structures and operate at very high temperatures for a considerable number of years then even the reduction of any heat losses by a small amount can be very valuable in terms of energy and cost savings.
A German source has estimated that for each gigajoule of energy used, perhaps 50 % might be subject to heat losses espe-cially after the products of combustion pass through the regenerator system. Perhaps the conclusion to be drawn from this is that the effective insulation of furnace regener-ator outer structures along with that of the flues, and heat recovery from the chimney are all very highly desirable objectives and effective in terms of cost reduction.
Many detailed studies have been made on the energy balances of all types of glass fur-naces. This is extremely complex because of the nature of the production process with some parts of the furnace such as the crown almost being exposed to steady state heat flow conditions, while regenerator pack-ings, for example, are subjected to much more transient heat flow conditions. These heat balances however can and do indicate in great detail where energy is introduced, where and how it is used, and perhaps most importantly of all, where it is lost.
This aspect of the reduction of heat losses is highlighted as being very important since it is difficult to make reductions in the current energy levels required to melt and condition the glass batch, especially since heat trans-fer from the fuels combusted to achieve this primary objective are already very efficient. It is standard practice in most furnace de-signs to insulate under the working hearth of a glass tank, perhaps, with firebrick on top of a layer of dense high duty calcium silicate slabs. It is not standard practice to apply any insulation to the lower sidewalls of the bath, where the hot face is fused cast refractory as this would greatly increase re-fractory wear reduce life, increase cost and affect glass quality. The regenerator structures them and the flues are all currently already externally insulated, but much more scope for even yet further energy savings appear to be possible in these areas and would be very cost-effective.
The methods of calculation of heat trans-fer into, through and from a refractory wall, roof or floor comprising of single or multiple layers over a wide range of temperatures have been defined in great detail in a num-ber of formulae published in thermodynam-ic textbooks and many of these, along with of course the refractories properties, have been incorporated into computer programs which are currently commercially available. These calculations cover radiation, conduc-tion and convection, in both steady state and transient heat conditions.
Consider a 250 mm thick vertical refractory wall constructed from high duty firebrick with a hot face temperature of 1000 °C and an ambient temperature on the cold face of 20 °C. Under steady state condi-tions, the heat loss at the cold face would be 4079 Wm2 and the actual cold face tem-perature of the layer would be 218 °C. In this example, the heat capacity of the layer would be 68 404 kJm2 (Tab. 2).
If 50 mm of efficient insulation were added to the cold face to insulate the slab and reduce heat losses under the same thermal conditions, then the interface temperature between the hot face layer and the insulation layer would be 764 °C while the cold face temperature would be 110 °C. The insulation would reduce the heat loss from 4079 Wm2 to 1273 Wm2 which is reduction in the heat loss of 2806 Wm2 equivalent to 69 % of the energy loss. In this exam-ple, the heat capacity of the layer would be 105 628 kJm2 which is of course higher due to the increased thickness and weight of re-fractories used in the construction.
If 50 mm of the same insulation were add-ed to the hot face to insulate the slab and reduce heat losses under the same thermal conditions, then the interface temperature between the hot face layer and the insula-tion layer would be 385 °C and the cold face temperature of the outer layer would be 115 °C. The heat loss would be reduced to 1365 Wm2 which is a reduction in the total heat loss of 2714 Wm2 equivalent to 66 % of the total energy loss. In this ex-ample, the heat capacity of the layer would be 36 126 kJm2.
Alternatively, if 50 mm of the same insula-tion were added to both the hot and cold face of the refractory layer, then the heat loss would be 754 Wm2 which is a reduction of 3325 Wm2 equivalent to an energy sav-ing of 82 %. In this case, the temperatures at the two interfaces would be 660 °C and 515 °C, while the temperature on the outer cold face layer of the wall would be 89 °C. In this example, the heat capacity of all three layers in total would be 81 356 kJm2 which is higher than the plain wall due to the add itional mass.
In the theoretical circumstances postulated above, it appears that there is a small ad-vantage to be gained by insulating the cold face compared to insulating the hot face of the refractory, but also an enormous ben-efit to be gained in insulating both the hot and cold faces in the furnace lining. The doubly insulated wall would cost more than the un insulated wall both to purchase and construct, but the payback time would be very much shorter, and it would save larger amounts of energy and money over the life of the furnace especially in steady state conditions.
These calculations are only a small part of the whole story. However, as cognizance must be taken of the compatibility of the materials and the mean temperature of each layer along with this relationship to the hot load stability of any individual layer. Attention must obviously be paid to any potential reaction of the hot face layer to the furnace atmosphere or batch. The thermal capacity of the wall could also be significant depending on whether the furnace was in a steady state condition or whether it was in a transient state in a cyclical batch unit. It is thus the optimum mix of good design, with premium mater-ials and quality construction to ultimately provide the end user with the entire essence of skilled refractories engineering, for any given set of operating parameters in any part of any furnace structure at all times.
The insulation products, most commonly applied to the top of the external walls of a glass regenerator, are ASTM 26 grade or ASTM 23 grade depending on the oper-ating conditions and ASTM grade 23 or ASTM grade 20 at the lower levels of the walls. Other materials are available how-ever, such as high density calcium silicate slab, insulating pouring mixes or more usu-ally light weight, medium weight gun mixes or insulation sprays (Fig. 2).
If typical thermal conductivities for these materials are compared from results found using the same conditions and test meth-ods as a basis, it can be seen that calcium silicate slabs and sprayed insulation are extremely efficient in reducing heat losses in comparison with for example insulation firebricks and insulation gun mixes of all types. Calcium silicate slabs, however, are slower and more difficult to fit and anchor as well as having multiple joints and being prone to damage over the life of the furnace.
The current range of sprays which are avail-able are quick and easy to install in large panels with no joints. This means that they are not only effective insulators but, that they also seal any joints in the under lying brickwork arising from construction or movement of the structure on heating from ambient. They also have a very uniform low density and high strength with no require-ment to apply anchors unless a thicker outer layer is calculated to be more beneficial. Large companies, such as RHI Magnesita, Vesuvius and HWI, amongst many others supply a wide range of refractories for the glass industry. Other companies around the world supply more specialist niche products. There are some specialist installers who can build and commission glass furnaces of all of the different types and sizes. There are extremely few refractories producers who can engineer supply, and install specialist insulation packages including bricks, emis-sivity coatings and insulation sprays. TNCR in China, working in collaboration with western partners, has been a lead-ing exponent of energy conservation. They have carried out a project involving almost 4000 m2 of insulated surfaces on a flat glass tank which feeds multiple production lines, and which may be the largest project in the world to date along with many others. Many other refractories suppliers and in-stallers can offer gunned insulation linings, based on lightweight or medium weight products, either from within their own or-ganisation or sub-contracted to other com-panies. In practice, these external refractory veneers are much heavier and often subject to variable density and thickness. They thus require substantial anchorages to ensure they stay in place.
The negative PLC of such materials is always significant at the interface tem-peratures created. These factors usually mean that any initial cost savings are soon negated by a higher tonnage of material that is required but even this is al-most insignificant compared to the lower energy and cost savings in use which this these types of products can achieve Tab. 3.
To illustrate this, consider a theoretical wall with 250 mm of firebrick with a hot face of 1000 °C and an ambient temperature of 20 °C insulated externally with 50 mm layers of either insulating gun mix or in-spray as alternative options. The density of the insulating gun mix would be around 1000 kg/ m3 while the inspray would be 250 kg/m3. 100 m2 of wall would therefore require 5 t or perhaps 6 t of insulating gun mix after rebound losses are taken into account and assuming that the wall was trimmed properly. The inspray alternative requires 1,25 t it has no rebound and would require minimal trimming.
Some readers may have doubt about insula-tion on the hot face of a furnace refractory lining but it is not uncommon in steel heat treatment furnaces, aluminium homogen-ising furnaces, ceramic kilns and petro-chemical fired heaters as well as of course most glass annealing lehrs.
In some large glass furnaces there is now a proven technique of applying a high emis-sivity coating on the hot face of refractories in some parts of the structure to effectively reradiate heat back into the furnace. These coatings work on the principle that signifi-cant amounts of energy are radiated back into the furnace rather than being absorbed in the lining. These new materials can have emissivity factors of around 90 % com-pared to a theoretical perfect black body. The coatings need to be thermally stable, tough and strong as well as to adhere well to whatever refractory they are applied to. In one fibreglass production unit where the coating was applied to the inside of the crown energy savings of 5 % were claimed over a three year operation (Fig. 3).
In a different example a lead crystal glass fossil fuel fired or electrically heated pot fur-nace whose refractory lining configuration before and after modification to improve the final insulation in service was as de-scribed (Fig. 4).
Before modification, the hot face was a 1625 °C, the cold face 424 °C, the heat loss 2907 Wm2 and the heat capacity 323 625 kJm2. After modification the hot face was still 1625 °C but the cold face was 362 °C the heat loss was 868 Wm2 giving a reduction in heat loss of 60 % with a heat capacity 393 436 kJm2. This provided a very short payback period and substantial energy and financial savings over the life of the furnace. Such an engineered solution executed job with premium materials and skilled personnel is not greatly more in cost than a cheap but never cheerful solution. There is almost never energy saving reasons to justify an inferior installation resulting in lower energy savings and certainly a re-duced operating performance with greater risk of shorter life and higher production costs (Fig. 5).
The same technology can be applied to even the largest furnaces although the re-fractories engineering capability employed needs to be extensive and the application of the insulation itself extremely precise and well organised as this phase of the project is executed after the furnaces reach working temperature. In a large tank furnace crown, the refractories would be primarily super duty silica bricks made from pure, very low flux raw materials because the aim would be to operate the furnace with the hot face temperature close to the melting point of the bricks to maximise production. The roofs can be and are insulated, however, only after very careful calculations are made regarding extra weight, increased interface temperatures, increased heat capacity and potential energy savings.
Since the crown operates under conditions approximating to steady state conditions, careful engineering can obtain a signifi-cant reduction in heat losses. Sprayed and pumped insulation is frequently used in the top layer, but since the spray has lower con-ductivity and lower density it has an overall advantage.
In larger furnaces, a study is made and a diagram of the regenerator structure includ-ing the crown is compiled showing the types of refractory and the thickness of each layer at a minimum of 27 significant points on a medium container tank, and at least 36 points on a large float glass tank. The hot face interface and cold face temperatures would be computed for the structure at each point before, and after the application of some specific insulation, and the results compared with the figures for the safe work-ing temperature of each layer as well as the calculated heat savings usually with one or more options of insulation ma ter ials and the installation method to apply it. Results from the top of a regenerator wall of a small to medium container glass tank would typically show benefits as illustrated in the example. One small container glass furnace instal-lation, monitored by the Energy Efficiency Office of the GB Department of the En vir-onment, showed that a GBP 10 000 invest-ment gave a reduction in energy costs of GBP 24 000 which was a payback period of 23 weeks so that the project paid for itself 20 times over during the life of the furnace. Calculation of the way which materials and labour prices have escalated since the trials in 1991, compared with the rise in energy costs indicates that this financial benefit would be greatly increased today (Fig. 6). The crowns of the regenerators are sub-ject to more transient heat flow conditions when the firing is reversed but the span of the arch is less and the hot face tempera-tures are generally lower than in the main crown and thus the risk of failure is less al-though still a factor.
Efficient insulation of the regenerator side and end walls as well as of the flues is as has already been stated, extremely bene-ficial in terms of reducing heat losses in all circumstances. Here again, however, careful calculation and skilled installa-tion maximises the benefits while min im-ising the risks and optimising the costs. Anyone doubting that there are risks in this, has probably not seen a crown col-lapse or calcium silicate panels breaking and separ ating from the walls and falling off in slabs. Even insulation gun mixes can separate and fail perhaps taking part of the front face of the underlying refractory layer with it. Overspray of any monolithic insula-tion can buckle the adjacent furnace steel- work.
Separate calculations are also additionally made for the flue as it is highly beneficial to do so. From these results, it is possible to estimate the likely overall energy sav-ings along with the payback time and the cost effectiveness for each project. It has been found that there is good correlation in practice between the results predicted from these studies and the results actu-ally achieved from many different types of furnaces over a long period of time. Larger float glass furnaces have paybacks of around a year and lower percentage overall energy savings of perhaps 10 % which can still amount to huge reductions in energy costs from the reduction of major heat loss-es after the port necks. It is difficult to argue that savings in energy costs equivalent to perhaps a year’s total energy consumption every 10 years is are not extremely well worth aiming for (Fig. 7).
Competent and experienced refractories engineers can evaluate the heat losses, and reduction in energy costs for a refractory structure at any point at any defined oper ating temperature even on multi-com-ponent walls which may be 500 mm thick and have as many as five layers of differ-ent materials. An overall projection of total savings and cost reductions can then be ar-rived at and experience in several hundred installations suggests that the findings are well borne out in practice as measured by reduced energy costs over the life of the furnace.
These studies and their competent imple-mentation have been proven over many projects to provide significant energy and cost savings as well as significant reduc-tions in CO2 and NOX and so are not only of economic benefit but also of consider-able environmental benefit also to all. The cost of carrying out a properly designed and installed furnace lining with optimum insulation saves itself many times over in service. The cost of getting the wrong en-gineering competence inferior materials or substandard installation is too high to even contemplate.
David A. Jarvis