WO2014036979A1 - Method for continuous glass melting under controlled convection of glass melt and glass melting furnace for making the same - Google Patents

Abstract

Energy sources, such as industrial glass burners (11), heating electrodes (10) and other suitable heating energy sources operate on the molten glass melt (6) containing undissolved particles, especially glass sand and bubbles, in the longitudinal axis of the melting space, or in a direction parallel with this longitudinal axis, until creation of one or more longitudinal temperature barriers in the glass melt (6) and until the arising of the cross temperature gradient [K.m^-1] which generates spiral-type flowing of the glass melt (6) with a rotary movement across the melting space, and in fact perpendicularly to the longitudinal axis of the melting part. This spiral-type flowing proceeds in the direction from the front wall (2) to the submerged cross refractory barrier (7) in the glass melt (6), or in the direction from the front wall (2) to the cross row (9) of the energy sources. The transversal temperature gradient [K.m^-1] of each spiral-type flowing is always set as higher than longitudinal temperature gradient [K.m^-1] between the front wall (2) and the submerged cross refractory barrier (7) in the glass melt (6), or between the front wall (2) and the cross row (9) of the energy sources, as a consequence of which it is possible to utilise 0.6 to 0.8 multiple of the total melting space. The ratio of the cross temperature gradient [K.m^-1] to the longitudinal temperature gradient [K.m^-1] is higher than 1 and lower than 30, preferably it is within the range from 5 to 20. Energy sources, such as heating electrodes (10) and/or industrial glass burners (11) in the melting space are arranged for creation of one or more longitudinal temperature barriers in the glass melt (6) and for generation of the spiral-type flowing of glass melt (6) with a rotary movement across the melting part, in fact perpendicularly to the longitudinal axis of the melting part, and for the setting of the cross temperature gradient [K.m^-1] of the spiral-type flowing higher than the longitudinal temperature gradient [K.m^-1].

Description

Method for continuous glass melting under controlled convection of glass melt, and a glass furnace for making the same

Field of the Invention

The invention concerns a method for continuous glass melting under controlled convection of glass melt, containing undissolved particles, especially glass sand and bubbles. This method is implemented in a horizontally oriented, separated continual melting space of a glass furnace, where specified melting space is outlined by the height of glass melt, width between opposite side walls and length between the front wall and the submerged cross refractory barrier in the glass melt, or between the front wall and the cross row of energy sources. The energy sources, such as industrial glass burners, heating electrodes and other suitable heating energy sources^ are situated in the glass melting space.

The invention concerns also a glass furnace for a continuous glass melting process with controlled convection of glass melt, during which molten glass, containing undissolved particles, especially glass sand and bubbles, and situated in the horizontally oriented and separated continual melting space, is affected through operation of the energy sources, operating on the glass melt in the melting space. The melting space includes a bottom, front wall and rear bridged wall and the following items between them: opposite side walls, crown, and the submerged cross refractory barrier submersed in the glass melt.

Background of the Invention

The continuous glass melting process is a method used for preparation of traditional materials on a large-scale basis by melting, connected directly with the processing of the raw material in many ways. Its foundations date back to the 19^th century when the arising glasswork technology was inspired by another melting process, i.e. steel production. In accordance with this process, the first continuous glass melting systems known as glass furnaces were designed as covered horizontal spaces lined with refractory materials, into which glass melting charge was placed on the entry side - this charge usually being formed of a mixture of crystalline raw materials with a variable share of the cullet, while homogeneous glass melt was taken away on the exit side for high-temperature processing. The systems were heated with the help of burners installed above the glass melt level, the heating medium being gas or oil. This means that they were compact single-space systems heated from above, which were later divided, by through- flow or narrowing, into two parts - the melting part where glass forming - homogenisation processes took place and were completed, and the working part where the glass melt was especially temperature-homogenised for further processing. The through flow of the glass melt through these systems was basically horizontal, but as a consequence of the naturally arising differences in the temperature by dispensing the cold charge on the glass melt level, type and installation place of the heating system and due to different levels of heat losses on the interfaces, also a natural circulation flow was always stabilised in both the melting and working spaces of the furnace, which created - after superposition with the natural circulation currents- a relatively complex image of flowing in the furnace space.

Other types of glass melting spaces were developed later on; especially those with electrical heating and featuring in fact a vertical through flow of the glass melt through the melting space and with a similar composition of the glass melt flow. This invention registration application, however, does not concern any spaces with prevailingly vertical through-flows.

The tests of the so-called transient characteristics of melting spaces [1-3] made it possible to further clarify the issues of melting of unusable circulating volumes of the glass melt. During these tests, an indicator was applied to the entry of the melting system; the indication function of this indicator was not subject to change during the melting process and it was identifiable on the exit from the system. The most frequent indicator was formed of suitable radioactive isotopes. The monitoring of the nature of dependence of the indicator in resulting glass on time provided information about the nature of flowing in the melting device. It helped also for identification of the above mentioned space in which the glass melt circulated only. The identified share of such spaces (referred to as the dead space fraction) achieved quite often 0.5 in common melting devices, and in some of the devices it was even higher. That is why it was quite obvious that the flowing nature plays a significant role for the effectiveness of the melting process. After some time, Cooper [4] defined, on the basis of the data concerning the type of flowing and kinetics of melting processes, a characteristic which is the ratio between the rate constant of the control melting process and the rate constant representing the material flow rate, which provides a degree of completion of the process in a given place or on the space trajectory. The work highlights the importance of the nature of the flow for melting processes.

Searching for an effective glass melting method required very detailed knowledge of conditions under which the melting process takes place in the industrial space. Such knowledge was only available through mathematic modelling of the course of the melting process in the operated glass furnaces. Applications of mathematical models provided detailed temperature as well as velocity fields of the glass melt in melting spaces and thus enabled modelling of the melting processes directly during the through-flow of the mixture in particular melting space [5- 24]. The results confirmed conclusions acquired by previous experience and measurements. Melting spaces featured stabilisation of a relatively complicated type of flowing consisting of a combination of longitudinally as well as transversally rotating areas and of a through flow which was heavily affected by circulations of the glass melt. The dissolution process took place under quite inconvenient temperature conditions [20], the purifying process was in fact limited to a small area of the maximum temperature in molten glass [ 19-24] and the melting spaces indicated a high fraction of the dead space [12]. A significant problem of melting processes was the course of the so-called critical paths, i.e. the routes of the glass melt through the melting space, on which the relevant homogenisation process is completed from the viewpoint of time as close to the exit from the space as possible. It was often possible to observe short-circuit routes passing through the space with low residence time, moreover this occurred on temperature conditions which were not too convenient and which substantially limited the melting performance of the concerned space and increased specific consumption of energy for the melting process. Thus it was possible to register efforts of removing such short-circuit routes by affecting the flowing characteristics [25]. During this, it was very difficult to influence complicated natural flowing in melting units in a significant way in the right direction, because the optimum nature of flowing in melting spaces was not investigated at a more detailed level. Nevertheless, Cooper [26] characterised, already earlier on the basis of a simple mathematical model of flowing of the glass melt in a simple model space, by using separated types of longitudinal and cross flowing, some consequences of the given type of flowing for homogenisation of the glass melt (e.g. acceleration of homogenisation of processes as well as temperature through cross flowing). These results indicate the need of influencing the flows in the melting spaces on the basis of detailed knowledge of relations between individual types of flowing and knowledge of their influence on the corresponding homogenisation process. Recent global challenges relating to technologies with lower demands in terms of energy, materials and ecology - which include the glass making technologies as well - called out other activities including both changes concerning necessary times of the glass melt residence in the melting space and changes in the general arrangement of the melting process [27]. The need of such changes, however, requires application of new principles to the melting process, including also changes in space arrangement of the process, e.g. in the form of the so-called segment melting devices in which partial processes take place separately, alternative characters of heating, etc. It was therefore obvious that achievement of these challenges requires partial move from normally modelled existing industrial devices to simple model spaces working under precisely defined conditions, on which it is possible to sufficiently get acquainted with the principles, to find their general laws enabling safe application later on, and so on. This procedure includes also the efforts of removal of the disadvantages of the naturally stabilised flowing in glass furnaces. Summary of the Invention

The above mentioned disadvantages of classical melting of glass can be removed or substantially limited by using the method for continuous glass melting under controlled convection of glass melt according to this invention. The essence of this invention consists in the fact that energy sources, such as industrial glass burners, heating electrodes and other suitable heating energy sources operate onto the molten glass melt containing undissolved particles, especially glass sand and bubbles, in the longitudinal axis of the melting space, or in a direction parallel with this longitudinal axis, until a cross temperature gradient is created there. The temperature gradient created [ .mf¹] will call out a spiral flow of the glass melt with a rotary circular movement across the melting space. This spiral-type flowing proceeds in the direction from the front wall to the cross damming wall in the glass melt, in the direction from the front wall to the cross row of energy sources. The cross temperature gradient [K.m^"1] of each spiral- type flowing pattern is always set as higher than its longitudinal temperature gradient [K.m ¹] between the front wall and the cross damming wall in the glass melt, or between the front wall and the cross row of energy sources, which leads to the fact that 0.6 to 0.8 multiple of the melting space out of the total melting space is utilized.

The main advantage of the method for continuous glass melting under controlled convection of glass melt according to this invention is the significant reduction of space in the melting part, so-called dead spaces, i.e. the places where the melting process does not take place or where it takes place only with very small or even zero efficiency and the glass melt is heated in these places in an unproductive way, which leads to an increase in thermal losses as well. A significant strength of this invention is the possibility of the setting of the ratio of the cross temperature gradient with regard to the longitudinal temperature gradient which is created by means of suitable arrangement of the energy sources and which generates favourable conditions for spiral-type flowing of glass melt through the melting space. This ensures also an even temperature distribution in the melting space and an optimum homogenisation ability of the glass melt. This leads to a significant reduction of thermal losses and a significant increase in the melting capacity of the glass furnace, which may possibly lead to a proposal for reduction of the melting part and thereby also for savings of the glass furnace construction costs.

It is advantageous if the ratio of the cross temperature gradient to the longitudinal temperature gradient is higher than 1 and lower than 30, which represents the widest real limits. Optimally, this ratio is used within the range from 5 to 20. If this ratio is lower than 1 , then the spiral-type flowing of glass melt does not occur. If t is ratio was higher, the advantageous spiral- type flowing would occur, it is true, but the efficiency of adjustment of this ratio will not set off technical difficulties during implementation. The advantageous spiral-type flowing of glass melt passes through the space whose utilisation for dissolution of particles, especially glass sand, accounts for a 0.6 to 0.8 multiple out of the total space, outlined by the height h of glass melt, length / between the cross damming wall in the glass melt, or the cross row of the energy sources and the forward base wall, and width w between opposite side walls. The spiral-type flowing according to this invention brings the highest defined utilisation of the melting space exceeding all what has been known so far. The utilisation of the melting space represents the size of the used part of the melting space compared to the non-used space.

The glass melt flow features an excellent homogenisation function and it leads to the shortening of the process of dissolution of those parts which are difficu lt to melt, especially glass sand or alumina particles. The specified arrangement of the energy sources ensures effective achievement of the required spiral-type flowing of glass melt. The melting space states lower thermal losses and a high melting performance, which allows for a possible reduction of the melting part and as a consequence also construction savings.

This method for continuous glass melting under controlled convection of glass melt is implemented in the glass furnace whose melting space includes a bottom, front wall and a rear bridge wall, and between them opposite side walls, crown, and the submerged cross bridge refractory which is immersed in the glass melt. The essence of this invention consists in the fact that energy sources, such as heating electrodes and/or industrial burners are arranged in the horizontally oriented continual and separated melting space for creation of one or more longitudinal temperature barriers in the glass melt and for the calling out of the spiral-type flowing of glass melt with a rotary movement across the melting part in fact perpendicularly to the longitudinal axis of the melting part, and for adjustment of the cross temperature gradient of each spiral-type flowing higher than the longitudinal temperature gradient, advantageously with the ratio of the cross temperature gradient to the longitudinal temperature gradient being higher than 1 and lower than 30, preferably in the ratio from 5 to 20. The energy sources are, with an advantage, specifically situated vertical or side heating electrodes or industrial burners with transversal flames or vertical burners, or burners of the U-flame furnaces.

The main advantage of the glass furnace according to this invention is that the melting part makes it possible, with its arrangement of energy sources, to generate the spiral-type flowing which will markedly reduce dead spaces in the melting space, where the process does not take place and where glass melt uselessly takes energy. Th is enables an even distribution of glass melt temperature in the melting space. The flowing of glass melt features an excellent homogenisation function and it makes it possible to shorten the process of dissolution of the particles difficult to melt, especially glass sand and perhaps alumina particles. Specified arrangement of the energy sources ensures an efficient achievement of the required spiral-type flowing of glass melt. The melting space states lower thermal losses and high melting performance, which allows for a possible reduction of the melting part and consequently also construction savings.

An advantage of the presented glass tank using glass melting with controlled convection of glass melt is an essential increase in its melting performance by substantially reducing the average time of residence of the glass melt in the melting part of a glass furnace, as a consequence of the preset flowing and higher utilisation of the melting space. This will lead also to a proportional reduction of thermal losses through the walls and crown of the furnace, whereby energy savings are achieved. The high specific melting performance, controlled during application in the melting space, can be used at the designing of new melting spaces which can feature a substantially lower dimension and construction cost savings. This will proportionally reduce also thermal losses through the walls and crown of the furnace, whereby energy savings are achieved. The high specific melting performance, controlled during application of the melting space, can be used at the designing of new melting spaces which can feature a substantially lower dimension and during their construction it is possible to save a non-negligible part of capital investment costs.

The energy sources are arranged in one or two side walls of the melting space. If the energy sources are arranged in one side wall of the melting part, there arises one spiral-type flowing across the entire melting part, which is suitable for the narrow melting part. If the energy sources are arranged in both the walls, there arise two in fact symmetrical spiral-type flowing patterns separated by the longitudinal axis of the melting part, which is suitable for melting spaces with difficult operators' access to the bottom in the case of a shift or exchange of heating electrodes. In the case of larger glass furnaces, i.e. also with a larger melting space and width, it is possible to use installation of heating electrodes in the bottom in two or more rows in a longitudinal direction of the melting space. In this case the spiral-type flowing arises between each side wall and the row of heating electrodes, as well as between individual rows of electrodes.

It is also advantageous when energy sources are arranged in regular mutual distances. Regular separations ensure an even heating of the glass melt, and thereby also favourable even flowing of glass melt.

The energy sources are heating electrodes and/or in glass industry used burners. The most advantageous for spiral-type flowing are heating electrodes for the reasons of direct transfer of energy to glass melt and easy control of heating electrodes. Heating electrodes can be arranged in the bottom as vertical or inclined, or they can operate onto glass melt from above as vertical - top electrodes. Plate electrodes installed in the side walls seem to be suitable. There can be also other types of heating energy, e.g. microwaves or plasma burners. The melting part can be terminated with a cross row of heating electrodes installed in the bottom. This cross row of heating electrodes creates a cross thermal barrier which is an alternative to the cross damming wall, and besides this it enables better control of thermal performance, depending on the melting capacity of the glass furnace.

For very small melting spaces for special glass types it is possible to use, as an energy source, also other sources, e.g. microwaves, plasma burners, or resistance sources in the side walls, or indirect sources of heating energy in the side walls, Super anthal loops.

Brief Description of Drawings

The state of the art is illustrated by Fig. l a and Fig. l b, adapted from W. Trier, Advances in Glass Technology, Plenum Press, 1962, page 619, where

Fig. la shows a longitudinal axial section of the glass furnace at a constant temperature, without circulation of glass melt and

Fig. lb shows a longitudinal axial section of the glass furnace at presence of temperature gradients and therefore also with circulation of glass melt.

The invention is described at a detailed level in the following invention description which is explained at a general level in attached schematic drawings, of which the following figures illustrate respectively:

Fig. 2 an axonometric view of a simplified basic model of the melting space with a full input and output,

Fig. 3 an axonometric view of a simplified melting space with a simplified critical trajectory of glass melt and dissolving particles of glass sand at the preset optimum flowing,

Fig. 4a graphic dependence of the maximum utilisation of the model melting space, on the time of dissolution of particles of glass sand and on the length of the model channel of the melting space,

Fig. 4b graphic dependence of the maximum utilisation of the model melting space, on the bubbles growth rate in the model channel of the melting space. In a more detailed way, the model implementation 1 for the melting of flat glass is illustrated in Fig. 5 - 16 where it is possible to find schematic illustrations as follows

Fig. 5 the melting section of a typical shape and arrangement of vertical electrodes, with a lower input of glass melt and a lower output of glass melt,

Fig. 6 the melting space of Fig. 5 for melting of flat glass, with an upper input and a lower output of glass melt, Fig. 7 resulting velocity distribution, illustrated through sections of trajectories of the glass melt made for 30 sec, in a longitudinal axial section of the melting space from Fig. 6,

Fig. 8 projections of the critical trajectory and other trajectories which are the nearest ones to the critical trajectory, for dissolution of sand grains, into the longitudinal axial vertical section of the melting space from Fig. 6, for the reference case,

Fig. 9 resulting velocity distribution in a longitudinal axial section of the melting space, which is illustrated through sections of trajectories of the glass melt made for 30 sec through the melting space from Fig. 6, for an optimum case,

Fig. 10 projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains, into the longitudinal axial vertical section of the melting space from Fig. 6, for an optimum case,

Fig. 11 melting space for melting of flat glass, with a lower input and a lower output of glass melt,

Fig. 12 resulting velocity distribution in a longitudinal axial section of the melting space from Fig. 1 1, illustrated through sections of trajectories of the glass melt made for 30 sec, for an optimum case,

Fig. 13 projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains into the longitudinal axial vertical section of the space from

Fig. 1 1, for an optimum case

Fig. 14 resulting velocity distribution in a longitudinal axial section of the melting space, illustrated through sections of trajectories of the glass melt made for 30sec, for an optimum case for removal of bubbles,

Fig. 15 projections of the critical trajectory and other trajectories which are the slowest ones, of bubbles with a diameter of 0.1 mm, into the longitudinal axial vertical section of the space, for an optimum case for removal of bubbles,

Fig. 16 projections of the critical trajectory and other trajectories which are the slowest ones, of bubbles of the initial diameter of 0.1 mm, into the longitudinal axial vertical section of the space. The preferred embodiment 2 for production of spectacle mouldings is illustrated in Fig.

17- 23, where it is possible to see schematic illustrations for

Fig. 17 resulting velocity distribution in a longitudinal axial section of the melting space at dissolution of sand grains, illustrated through sections of trajectories of the glass melt made for 30 sec, for a reference case,

Fig. 18 projections of the critical trajectory and other trajectories which are the slowest ones, for dissolution of sand grains, into longitudinal axial vertical section of the melting space, for a reference case, Fig. 19 resulting velocity distribution in a longitudinal axial section of the melting space for dissolution of sand, illustrated through sections of trajectories of the glass melt made for 30 sec, for an optimum case,

Fig. 20 projections of the critical trajectory and other trajectories which are the slowest ones, for dissolution of sand grains, into a longitudinal axial vertical section of the melting space, for an optimum case,

Fig. 21 resulting velocity distribution in a longitudinal axial section of the melting space illustrated through sections of trajectories of the glass melt, made for 30 sec during removal of bubbles, for an optimum case,

Fig. 22 projection of the critical and other non-advantageous trajectories of bubbles, into a longitudinal axial vertical section of the space, for an optimum case,

Fig. 23 projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains, into the longitudinal axial vertical section of the space. The model implementation 3 for production of package glass mouldings is illustrated in

Fig. 24a - 30, where the melting space is presented in a schematic manner

Fig. 24a in a longitudinal vertical section,

Fig. 24b in a horizontal section below the glass melt level,

Fig. 25 resulting velocity distribution in a longitudinal axial section of the melting space, presented sections of the glass melt trajectories made for 30 s, for a reference case,

Fig. 26 projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains, into longitudinal axial vertical section of the melting space, for a reference case,

Fig. 27 resulting velocity distribution in a longitudinal axial section of the melting space, illustrated sections of glass melt trajectories made for 30 s, for an optimised case,

Fig. 28 projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains, into the longitudinal axial vertical section of the space, for an optimised case,

Fig. 29 resulting velocity distribution in a longitudinal axial section of the melting space, illustrated through sections of trajectories of the glass melt made for 30s, for an optimised case,

Fig. 30 projection of the critical and other trajectories (the most non-advantageous ones) for removal of bubbles, into the longitudinal axial vertical section of the space, for an optimised case. The model implementation 4 for production of domestic glass is illustrated, in a schematic manner, in Fig. 3 1 , 32, where it is possible to view a schematic illustration of the melting space in a longitudinal axial section, and in

Fig. 31 with heating by burners in the side walls, for a reference case

Fig. 32 with heating by vertical electrodes and side burners, for an optimum case.

Preferred Examples of the Invention

W. Trier, Advances in Glass Technology, Plenum Press, 1962, page 619, shows, in the attached Fig. l a, a typical view of flowing in the longitudinal axial section through the horizontal glass furnace without circulation at a constant temperature, and the attached Fig. l b shows a real picture of flowing of the horizontal glass furnace at the presence of temperature gradients, and therefore also circulations.

The naturally arising type of flowing of the glass melt obvious from Fig. lb had, especially in the melting parts of the glass furnaces, an essential influence on the course of the melting process, and therefore also on the effectiveness of the entire melting system from the viewpoint of its performance, as well as from the viewpoint of consumption of energy for the melting process. The type arising this way had certain advantages for the course of the melting process, consisting especially in the fact that the huge return flow of glass melt from the maximum temperature region to the glass batch layer in the introductory part of the furnace, as shown by the left circulation circuit in Fig. l b, brought the necessary heat to the area of reacting raw materials for both heating and raw material reactions. Another advantage of the natural flowing was represented by the frequently present return flow of glass melt from the throat to the area of the melting temperature, as illustrated in Fig. l b, in the lower part of the right circulation circuit. The return flow returned non-homogeneities back to the melting process, especially bubbles, which were a consequence of an imperfect course of the melting process. A disadvantage of the stabilised type of flowing was the arising downward flow of glass melt from the charge level to the bottom of the melting furnace, where it moved ahead at the lowest temperatures up to the area of maximum temperatures, where the circulation currents only brought it to the high temperatures at the level. A consequence of this flowing was a slow homogenisation of glass melt (dissolution of particles and chemical non-homogeneities) and a virtual absence of the purifying process (output of bubbles) in an essential part of the active furnace space at the bottom of the furnace. While the dissolution processes could be finished after ' the bringing of the glass melt from the bottom to the level, the process of removal of primary bubbles in the high-temperature area at the level only started. Both the processes thus took place in the area after reactions of the glass charge in fact in a series manner, and dissolution took place, for most time needed, at low temperatures at the bottom of the furnace. The follow-up of the processes (instead of their intensive parallel course) and the large fraction of unfavourable conditions during the dissolution process were reflected in high thermal losses of the melting process and in small specific melting capacity of the units, which required, at higher planned capacities, relatively huge devices. Besides this, it was shown that the stabilised circulation flowing creates, at low specific capacities of melting spaces, significant volumes of long-term circulating glass melt in the spaces, in which the melting process in fact does not take place (dead spaces - "over-processing" spaces). This glass melt is therefore maintained at the melting temperature without any actual effect and causes useless thermal loss.

The inventors of the present application recently started to deal, at a detailed level, with the influence of the flowing nature on two most important melting processes, i.e. dissolution of non-reacted particles of the glass batch, which represents glass sand in most cases, and on removal of bubbles (refining), which is mostly the slowest part of the melting process [28-35]. They identified specific consumption of energy and specific melting capacity of the given melting space as quality criteria for the melting process performed. As the model melting space they selected a continual horizontal space in a block shape with both the input and output through the entire front walls or their parts. Its chart is shown in Figure 2, the shape of the melting space was simple, with an effort of achieving the results which are as general as possible, the dimensions of the device were small in accordance with the aim of miniaturisation of the next melting space - length 1 m, width and height of the glass melt level 0.5 m. Investigation was carried out with a possibility of applying the results to other dimensions of the real device. The tool of investigation selected by the authors was a mathematical model of the melting space with a possibility of a wide variability of conditions.

Fig. 2 shows a schematic illustration of the simplified model melting space with a full input and output. The appropriate nature of flowing was set with the help of linear temperature gradients inserted on the glass melt level; the bottom and walls of the space were insulated. Since the dissolution process and the process of removal of bubbles take place on different principles, the investigation was carried out for each process separately provided that during applications, a compromise between optimum conditions will be made or a separate segment of the device will be proposed for each process.

To understand the results of the monitoring process well, it is necessary to state a theoretical basis of investigation, concerning (for the present application) dissolution of particles of glass sand which is usually the main part of industrial glass and represents the glass raw material which is the most difficult and the last to melt in the glass melting process.

The criteria for the process without energy recycling, namely specific energy consumption and voluminous performance of the melting space, are described by the following equations including, already separately, the process time (its kinetic characteristics) and the influence of the nature of flowing. The influence of the nature of flowing is expressed by means of the newly introduced characteristic called "utilisation factor of the melting space", u_D [3 \ , 33- 35]:

V = u_D (2) where / J¾ is specific energy consumption [Jkg^"1]; is theoretical heat necessary for chemical reactions, phase and modification transformations and heating of the input mixture and arising glass melt to the melting temperature [Jkg ¹]; H^L is the total flow of heat through interfaces to the surroundings [Js^*1]; f_D is the mean time of dissolution of all particles of glass sand in various time and temperature modes arising in the space [s]; p is specific weight of glass [kg m³]; V is the volume of the melting space [m³]; V is volume flow rate of glass melt in the space on condition of completion of the melting process [mV], and κ_υ is the use of the melting space for dissolution. In this case, u_D expresses a relation between the mean time for implementation of the actual process, f_D, and the theoretical time of glass residence in the melting space, the so-called geometric residence time, which equals to the fraction of the volume and volume flow rate through the space r_c = V/V: u_D = ^ , _D e (0; l) (3)

It is clear from the equations (1 -2) that specific energy consumption is inversely proportional and volume flow rate is directly proportional to the utilisation of the melting space.

The formula for the use of the melting space in the case of dissolution of sand grains consists of two ratios of dead spaces, where nic, includes closed circulations and areas of almost static glass melt (e.g. corners). This value was explored already earlier, e.g. by using the transient characteristics method, as provided for in the text above. m_D refers to the area where all sand grains are already dissolved, but the corresponding glass melt is still heated in the space (over- processing). The use of the melting space for dissolution of sand is then given as follows: u_D = (1 - m_c)(l - m_D); m_G = m₀ = ~≥ (4)

Tc T

where f is the average residence time of the glass melt flowing through the space. The mean time of dissolution of sand grains on isothermal or almost isothermal conditions then meets the relation:

— — ^TDcrit- where τ_β is the time necessary for dissolution of all sand grains at the given or average temperature and T_Dcr;_t is the time of residence of glass melt on the fastest (critical) trajectory. The calculation of the use of the melting space works with the values of the times of dissolution of sand grains and with the results of model ling of the dissolution process in the model melting space in time-temperature modes on trajectories of glass melt (which means that the negligible buoyancy force operating on sand particles is not considered).

The formula for the use of the space during removal of bubbles consists of two parts as well, and its approximate value for small critical bubbles is as follows:

where T/~ is the time of removal of the smal lest critical bubble which reaches the level just on the output from the space, and m_vir, is the so-called virtual dead space. h_vir, is then the so-called virtual height expressing vertical distance which must be made by the critical bubble with regard to the molten mass in order to reach the level, and ho is the actual thickness of the layer of the glass melt in the space. Calculation of the use of the melting space works with experimentally identified bubble growing rates which are used for calculation of the reference time x_Fnj, and with the results of modelling of the dissolution process in the model melting channel at time- temperature modes on trajectories of the glass melt (which means that the negligible buoyancy force operating on sand particles is not considered) and on trajectories of bubbles.

Specifically, while investigating the influence of the nature of flowing on capacity and specific energy consumption, at a constant time of dissolution of sand grains or at a constant temperature dependence of the dissolution time on the temperature, the influence of the very slow convection of glass melt on the rate of dissolution of sand grains is not considered. The values of the times of dissolution of particles of glass sand are determined experimentally. In order to compare individual cases between each other, a reference state is specified. On achievement of the reference state, the last - critical - particle is dissolved just on the output from the melting space. Technologically, this state corresponds to the state of the process running "without any reserve" and real capacities would be (in fact) lower. The resulting particular value of the modelling process is the value of the use of the space ₀, which is, at a constant kinetics of dissolution, directly proportional to capacity of the space (equation 2) and inversely proportional to the specific energy consumption (equatio 1 ). For information purposes, the value u_D for the melting space (of the melting space type) is equal to 1 on presumption of a piston-type flow and for isothermal melting space with parabolic distribution of velocities of glass melt it is equal to 0.445. The higher value is, however, achievable only for ideal liquid, and the isothermal flow of the glass melt is then difficult to implement in high-temperature melting spaces where a horizontal difference of temperature values by several degrees already calls out efficient circulation flowing and reduction of UD. A detailed investigation of the influence of elementary and mixed types of flowing was carried out on the given model melting space, where the basic types of flowing were considered as pure longitudinal circulation flowing in both clockwise and anticlockwise directions (circulations take place in vertical planes parallel with the main work direction of flowing) and cross flowing, where circulations take place in the planes perpendicular to the main - work direction of flowing. Mixed flowing patterns then included flowing types whose intensity ratio was set by ratios between inserted horizontal temperature gradients. It was shown that both types of longitudinal flowing result in the arising of large dead spaces, where the use of the melting space achieved only the values from 0.1 to 0.2, while introduction of pure cross flowing provided the values of the use 0.4 to 0.5 as a consequence of disappearing of a part of dead spaces. The best results were, however, achieved for a mixed type of flowing, where the fast component of the forward flowing at the glass melt level was slowed down by the small temperature gradient with a higher temperature at the output from the space and where relatively intensive cross flowing was called out by the cross temperature gradient, and the absolute ratio between the sizes of the transverse and longitudinal temperature gradients was equal to 5 to 10. This optimum type of flowing was characterised by spiral trajectories of glass melt, as shown by Figure 3, and by a very low value of the dead space mc, low value of the virtual dead space m_virh partly reduced value of the dead space m_fl and values of virtual height, not exceeding the thickness of the glass melt layer ¾(see equations (4-5).

It was a positive fact that optimum conditions for high use of the space were very similar for both dissolution of particles of glass sand and for removal of bubbles, although both the processes feature a different nature. This means that if the two processes take place independently under the given conditions and if there is no nucleation of bubbles on the dissolving sand grains, it is possible to observe both the processes very well under advantageous conditions of flowing in the common space.

Other research showed how this optimum type of flowing and conditions of its setting change on changes in independent variables, which included the already mentioned ratio between the intensity of cross flowing to longitudinal flowing, overall intensity of circulation flowing, time of dissolution of particles of glass sand, bubble growing rate and length of the melting space. The influence of the glass characteristics and of the height of the glass melt layer can be included into the influence of the overall intensity of flowing. The width of the space will not have any essential influence on the use, if temperature gradients are maintained. For each couple of the time of dissolution of particles of glass sand or bubbles growing rate in the given glass melt and the length of the melting space, while maintaining the height of the layer of the glass melt, it is possible to find a maximum value of the use of the melting space for the dissolution process or for the process of removal of bubbles. These maximum values oscillated, in a wide range of the times of dissolution of sand, rate of the process of removal of bubbles (represented by the bubbles growing rate) and lengths of the channel of the melting space above 0.5; in most cases even between 0.6 to 0.8, which values were markedly better than even those that can be achieved in an isothermal melting space. The dependence of these maximum values of the use on both the above mentioned characteristics in a model melting space is provided for in Figure 4a, and a similar dependence of the use of the space for removal of bubbles on the bubbles growing rate can be seen in Figure 4b, Fig. 4a shows dependence of the maximum value of the use of the melting space on the time of dissolution of glass sand for different lengths of the melting space. Fig. 4b illustrates dependence of the maximum utilisation of the model melting space in a model channel on the bubbles growing rate. From the described dependencies it is clear that maximum values of the use of the melting space grow with both the sand dissolution time and the length of the melting space. Maximum values of the use of the melting space depend on the bubbles growing rate only little. At the same time it was possible to discover the rules governing the values of temperature gradients, which it is necessary to set for achievement of an optimum use of the melting space.

The results acquired showed that it is possible to find and specify n optimum type of the spiral-type flowing from the viewpoint of dissolution of glass sand particles as well as removal of bubbles, and to specify optimum conditions for its achievement. There were specified also rules for changes in the optimum use and of the optimum conditions at changes in input parameters. This means that there were acquired presumptions for the transfer of results to real melting spaces.

The glass furnace for the continuous glass melting under controlled convection of glass melt, during which the system operates on molten glass, has energy sources which are arranged in the horizontally oriented continual and separated melting space, such as heating electrodes JO and/or industrial burners H_. for creation of one or more longitudinal temperature barriers in the glass melt 6 and for generation of the spiral-type flowing of glass melt 6 with a rotary movement across the melting part, in fact perpendicularly to the longitudinal axis of the melting part. This enables setting of a cross temperature gradient [K.m ¹] of each spiral-type flowing higher than the longitudinal temperature gradient [K.m ¹] between the gable front wall 2 and the cross damming wall 7 in the glass melt 6, or between the front wall 2 and the cross row 9 of the energy sources, as a consequence of which it is possible to utilisation a 0.6 to 0.8 multiple of the melting space out of the total melting space available. The energy sources can be placed in various sites for creation of one or more longitudinal temperature barriers in the glass melt 6 of such a type.

Vertical heating electrodes 10 are installed, with an advantage, in one row in the longitudinal axial section of the melting space, or from both the sides of the longitudinal axial section of the melting space and/or in a manner parallel with this section and/or they are placed along one or both side walls 4 of the melting space. The most suitable vertical heating electrodes 10 are rod-type molybdenum electrodes 10 with constant or variable cross sections.

Side heating electrodes JO are installed in one or two side walls 4 of the melting space, whose tops reach the area of glass melt 6 situated along the longitudinal axial section of the melting space or reach, with their body, to the vicinity of side walls 4. Suitable side heating electrodes J_0 are also rod-type electrodes _10 or plate electrodes for side walls 4.

The transverse flame burners J_l are placed in the side walls 4 of the melting space and are directed in fact to the area of the longitudinal axis of the level of glass melt 6, or to the area of the level of glass melt 6 along one of the side walls 4 of the melting space. These burners Π_. can be gas burners or oxygen-gas burners.

Vertical burners H_ of a gas type or oxygen-gas type are advantageously installed in the crown 5 of the melting space and are directed in fact to the vicinity of the area of the longitudinal axis of the glass melt level 6 or to the area on the level of glass melt 6 along one or both side walls 4 of the melting space.

U-flame burners are advantageously situated in the gable wall 2 or the rear bridge wall 3 of the melting space and are directed to the area of the longitudinal axis of the level of glass melt 6 or to the area on the level of glass melt 6 along one or both side walls 4 of the melting space.

It is possible to expect that the most available and perhaps also the most frequently applied variant will be the use of vertical electrodes 10_^ or gas burners _U in a longitudinal axial section of the melting part. Depending on the design of the furnace and type of glass it is possible to use other heating types as well.

E x a m p l e 1

(Fig. 5 to 16)

For application of the results, it is however necessary to specify a particular melting space, particular energy sources, their allocation and heating performance enabling adjustment of the desirable type of spiral flowing, which is the subject matter of the present application for the invention registration.

The achievement of the objective consists in designing of a special glass melting space for dissolution of the parts of the glass charge, especially glass sand, and removal of bubbles, i.e. the designing of the shape and suitable dimensions of the melting space, its typical material composition, of the method of heating and allocation of heat sources in the melting space, and also definition of distribution of energy to individual heat sources in such a way that the melting process can take place at the mean temperature known beforehand and that it can be possible to achieve the type of the spiral flowing which ensures high utilisation of the melting space for the process of dissolution of glass sand as well as the process of removal of bubbles. The prescribed procedure of mathematic modelling of the melting space connected with calculation of its use will be used to achieve this situation.

During the process implementation according to the design of the present invention the process operates on molten glass, containing undissolved particles of glass sand and bubbles and situated in the horizontal continual melting space, energy sources, basically heating electrodes H) or industrial glass burners jj_ 'ⁿ such a way that the molten glass will feature the distribution of temperatures which generates the spiral type of flowing in the direction of the glass melt passing through the space, whereby it is possible to achieve a high use of the given melting space for the sand dissolution process and the process of removal of bubbles, and therefore also a high performance of the space and low specific losses of energy.

It seems that a suitable special space is a block whose length is basically greater than its width and the height of the glass melt level in the melting space of the glass furnace. The melting space is horizontally oriented, continual and separated, and it includes the bottom J_, gable wall 2 and the opposite rear bridge wall 3 and between them arranged opposite side walls 4, crown 5 and the cross damming wall 7 submersed into the glass melt 6. The energy sources are arranged in the melting space at least in one row in the longitudinal axis of the melting space, or in a manner parallel with this longitudinal axis, between the gable wall 2 and the cross damming wall 7 in the glass melt 6, or the cross row 9 of the energy sources in the melting part of glass melt 6 for generation of the spiral-type of flowing of glass melt 6 with a rotary circular movement across the melting part. The energy sources are preferably heating electrodes J_0 or industrial burners M_.

The mixture of the molten glass melt 6 with both solid and gaseous non-homogeneities enters the melting space through the upper input or lower input J_2, which may be narrowed or may occupy the entire width of the melting space and may feature different heights. The too narrow input V2 and output J_3_, however, affect the adjusted nature of flowing in the space, and therefore it is recommended for the input \ 2 or output Γ3 not to occupy an area smaller than about 20% of the charging front wall 2 or rear bridge wall 3.

The melting space is lined with a refractory lining which is normally used in glass furnaces. Concerning heating, according to the modelling results the best-proven method is electrical heating through electrodes J_0, which are usually of a molybdenum type. It is, however, possible to use heating with gaseous or oil burners J_L, especially in combination with electrical heating. A chart of the typical designed melting space heated electrically is provided for in Figure 5. The melting process takes place in the range of temperature values suitable and common during the melting of industrial glass, specific temperature depends on the glass type, but for common glass it is necessary to count on medium melting temperatures 1300 to 1500 °C. The continuous through-flow is ensured through the inflow of mixture of glass melt containing solid as well as gaseous non-homogeneities from the previous space, where the input mixture of glass raw materials is heated and melts, and through the off-take to another space, where the homogenisation process is eventually finished, bubbles are removed, or the temperatures only get stabilised and the finished glass melt is taken away. The melting space capacity can be adjusted, besides the setting of a suitable type of spiral flowing, which is the subject matter of this invention, through the grain size of solid particles (sand grains), adjustment of composition of resulting glass and the mean temperature in the dissolution space, addition of purification agents and possibly application of reduced pressure. The same quantity of glass melt 6 is maintained in the melting space of the glass furnace by means of inflow and off-take.

A typical shape and arrangement of heat sources, electrodes 10, in the designed module for dissolution is provided for in Fig. 5.

The grounds of the invention submitted consist in achievement of a certain type of the spiral-type flowing in the continual space through arrangement of heat sources, or through dimensioning of insulation of a glass furnace in such a way that cross circulation flowing can arise in the melting space due to the effect of both factors, which is to simulate an optimum type of spiral flowing, found out during investigation of the model space with inserted temperature gradients. This cross circulation flowing features a partially suppressed fast longitudinal component of the movement of the glass melt at the level, in order to reduce the differences between the times of residence of glass melt 6 on various trajectories. This can be achieved in the designed space, e.g. through a longitudinal row of electrodes JO installed basically from the bottom i of the glass furnace in such a way that there can arise a longitudinal thermal barrier, either in the longitudinal axis of the melting space, where electrodes JO are installed as well, or where a longitudinal area of the highest temperatures can arise at any of the side walls 4 of the melting space, where a longitudinal row of electrodes JO would be installed too. At melting spaces designed for large capacities it is possible to install also more than one longitudinal row of electrodes 10, but this arrangement seems to be less advantageous for the slowing down of the horizontal component of the velocity at the level. For the purpose of a better achievement of the desirable type of the spiral flowing it is reasonable (but not necessary) to place in the space also other sources, e.g. the cross row 9 of electrodes JO, which can ensure the slowing down of the fast forward component of flowing at the level in a better way or which can prevent the longitudinal circulation flowing of glass melt 6 from occurring. In a similar way it is necessary to arrange also burners JJ_. above the level of the glass furnace, either to place the burners JJ_. in the crown 5 of the furnace, or in the side walls 4_^ and to orient them in such a way that a temperature barrier can be created in the melting space in the longitudinal axis of the glass furnace or at one of the side walls 4. During the use of burners JJ_ as thermal sources it is possible to use at the same time electrodes K), prevailingly in the longitudinal axis of the melting space or at any of the walls which support creation of a longitudinal temperature barrier. Distribution of energy to heat sources or necessary insulation of the melting space or its part, ensuring an optimum nature of spiral-type flowing, must be set in a targeted mathematical modelling of the melting space. For modelling it is necessary to experimentally identify the dependence of dissolution of glass sand on the temperature in the presupposed temperature range of melting and temperature dependence of the bubbles growing rate. For glass of a similar composition and the same glass sand grain size it is possible to use the same temperature dependence. By modelling it is possible to identify, for both the processes separately, the melting performance and specific energy consumption of the melting space working in a critical mode, i.e. without reserve, i.e. the last dissolving particles will be dissolved just on the output 13 from the melting space, or the last bubble reaches the level just on the output. The distribution of energy on the sources is gradually adjusted in such a way that it can be possible to achieve a state when the calculated value of the use of the melting space achieves maximum achievable values, or when the value is the highest at the technically achievable arrangement for both the processes. While modelling, it is possible to make corrections also in the arrangement of heat sources, but always with a view of creating the required type of the spiral flowing. Required values of the use of the space should oscillate at least around the value of 0.4, however rather above the value of 0.5.

It is of course very useful to estimate how the spiral-type flowing and its corresponding use of the melting space behave at marginal values of settings, i.e. at very low or very high ratios of intensities of cross and longitudinal flowing with a potential return flow at the level of the glass melt 6, at low or very high intensities of circulations of glass melt 6, at very slow or very fast dissolution or removal of bubbles, or at short or (on the other hand) very long melting spaces.

The very low value of the cross flowing during the existence of longitudinal flowing will always lead to a drop of the utilisation and during the growing intensities of the longitudinal flowing it will approach to zero. And vice versa, in the case of a very high ratio of intensities of the cross-to-longitudinal flowing, the value of the use will be closer to 0.5 and then it will remain around this benchmark value, while the value of the use for the removal of bubbles drops to the values around 0.3 - 0.4 only at cross circulations of glass melt 6.

In the area of the ratio between intensities of flowing of glass melt 6, which is expressed through the ratio of cross to longitudinal temperature gradient equal to 5 to 20, the utilisation will achieve maximum values from 0.6 to 0.8. In this area, the technologically advantageous zone is characterised by the maximum value of the use of the space. In other ranges of the ratio of intensities of flowing it is virtually meaningless to work at any overall intensities of flowing, dissolution rates or melting space lengths.

Maximum values of the use of the melting space within this range of the ratio achieve the value 0.445 for dissolution of sand grains and 0.667 for removal of bubbles at a zero intensity of circulation flowing, when glass melt 6 flows through the isothermal rectangular melting space. With the growth of the overall intensity of circulation flowing the maximum values of the utilisation of the melting space achieve gradually the values from 0.6 to 0.8 at normally achievable values of intensity of circulations, and values from 0.5 to 0.6 at high intensities of circulations for dissolution of sand grains, or they drop to very low values for removal of bubbles. No very high intensities of circulations are, however, achieved under ordinary conditions. Maximum values of the utilisation in the area of an optimum ratio between intensities of cross and longitudinal flowing depend only little on the rate of dissolution of sand grains as well as on the bubbles growing rate, and oscillate within a wide range from 0.6 to 0.8. The very short channels of the melting space state low values of the utilisation, because spiral shape of flowing is unable to develop there, the values drop to 0.4 at the length of the melting space 0.5 m, and they will approximate to the value 0.4 at dissolution of sand grains and 0.6 at removal of bubbles for still shorter spaces. For very long channels of the melting space, the value of the utilisation for both the processes will be retained relatively high, around 0.6 and more, nevertheless the necessary optimum ratio between the intensity of the cross and longitudinal flowing will inadequately rise.

A specific objective is then to transfer the nature of the temperature field created in the model melting space with the help of temperature gradients to the designed particular melting space, differing from the model one especially by stating temperature losses through interfaces and is heated by particular sources, e.g. electrodes U) or burners JJ. '^η such a system and distribution of the energy supplied that a temperature field creates inside the space, which generates the required spiral type of flowing of glass melt 6, characterised by the high utilisation of the melting space and corresponding to the nature of flowing in the model device. This objective is achieved through the design of the first arrangement of the melting space, based on the knowledge acquired from the original model device by means of mathematical modelling of the temperature and velocity fields of the glass melt and the course of dissolution of sand grains or removal of bubbles on trajectories of glass melt created by the energy inserted into the arranged heat sources and loss in the variant designed. Mathematical modelling will then be made for each process separately. The capacity of this arrangement is adjusted after the first calculation in such a way that - at a preset and maintained average temperature in the melting space - the last glass sand grain can be dissolved just on the output 13_. from the melting space, or the last bubble of the critical value can achieve the value of the glass melt level 6 just on the output, i.e., so that this space can work without any melting reserve. This state is referred to as critical.

Through calculation it is possible to obtain time T_Dcr„, of dissolution of sand grains, or time Tfcrii, of removal of bubbles on the most disadvantageous critical route through the space, when glass sand grains are dissolved just on the output from the space or the bubble rises towards the level just on the output from the space. The given critical trajectory determines maximum melting performance V (based on the critical state) and appropriate geometric residence time c = V/V. From the modelling, the times of dissolution of sand on other monitored trajectories (their number exceeding 10⁵ and more), the mean value f₀ of the dissolution time of glass sand particles, and from the distribution of residence times of glass melt on monitored trajectories, the mean residence time f in the space are determined. With the help of equation (4) it is then possible to acquire the corresponding utilisation of the space and then the data of capacity and energy consumed is used for calculation of the specific energy consumption for thermal losses. While monitoring the removal of bubbles, it is necessary to use the reference time for removal of bubbles, calculated from the thickness of the layer of the glass melt in the glass melting space and the bubbles growing rate at a mean temperature in the melting space [30].

The first arrangement can be (optionally) the reference one and simulates an arrangement corresponding normally to a classically heated device, stating a low value of the use of the space. The results of calculations at expected advantageous type of flowing with a high use of the melting space (hereinafter referred to as "space" only) are then compared with reference ones and their comparisons serve as a basis for other modified variants. As it has already been mentioned, it is not always necessary to make the calculation of reference settings and while searching for an optimum spiral-type flowing it is possible, in next calculations, to build upon the acquired values of the use of the space only, where good values start approximately about 0.4. The aim is to get to the values uo and W/ as large as possible, for the use of the melting space, approximately between 0.6 and 0.8.

The second arrangement for both the processes is based also on results of modelling on the melting space with preset temperature marginal conditions, nevertheless it is necessary to set already such arrangement and dimensioning of selected heat sources (or dimensioning of interface insulation) that it can be possible to better achieve the described advantageous type of spiral flowing. This setting will be similar for both the processes because both of them require a similar type of flowing for their effective course. In the first variant for both the processes, the above described procedure will be used for calculation of the use of the melting space in this variant, space capacity and specific energy consumption and it is eventually compared with reference values of the first arrangement. According to the results it is possible to adjust the arrangement of sources or insulation and distribution of supplied energy in the direction where increased use for each process is expected. The modification direction is determined with the help of already acquired general rules through which it is possible to control the use of the melting space. In the second variant it is possible to obtain the second set of the use, specific energy consumption and capacity. The process of designing of improved variants possibly continues until achievement of the best achievable value of the use of the melting space for each process separately.

This procedure leads, however, to the obtaining of quite different conditions as well as arrangement for each process. Another task will be the setting of the same conditions for both processes in such a way that the slower process will be the control process, i.e. it will only be terminated on the output from the device and the faster process will be terminated before the output, which means that there will be a slight reserve. In this step it is possible to use optimum settings of flowing of the slower process for a faster process, and it is possible to calculate its values of the use of the space, capacity and energy consumption under these conditions. If the capacity found out for the originally faster process is, even on these conditions, higher than the one for the slower process, these conditions are suitable for operation of both processes simultaneously in one space. If the capacity is lower, it is necessary to additionally compute, through changes in the space capacity and modelling of both processes, the state when the slower process is terminated just on the output from the space and the faster process inside the space.

The procedure leading to the design of the melting space with a high use is therefore divided into the following steps:

1. Identification of the dependence of the time of dissolution of glass sand grains on temperature in the laboratory and determination of the temperature dependence of medium dissolution rate on temperature. Finding out of medium bubbles growing rate depending on the temperature.

2. Design of basic dimensions and method of heating of the new melting space with the use of ordinary construction knowledge and results from the original model space on the basis of requirements for capacity and while maintaining the necessary conditions of melting, i.e. the same mean temperature for both the processes.

3. The calculation of the use of the melting space for each process separately at a reference arrangement corresponding to conditions in a classically operated system. This step is not inevitable or it is carried out, in an informative manner, for one process only.

4. The calculation of the use in the melting space with proposed parameters corresponding to the expected desirable type of spiral flowing for each process separately. 5. Repeated modifications of conditions and calculation leading to the arrangement and conditions, on which it is possible to achieve the highest possible use of the melting space and corresponding values of the melting performance and specific losses from the viewpoint of theoretical as well as practical aspects, for each process separately.

6. Identification of a slower control process on the basis of melting performance achieved and calculation of the use of the melting space of the faster process under optimum conditions valid for the slower process. The calculation ends if the melting performance achieved is higher than the melting performance of the slower process.

7. If the melting performance achieved is lower than the melting performance of the slower process, it is necessary to reduce the performance of the melting space to the state when the faster process is terminated just on the output from this area and the slower process inside this area.

A special melting space for dissolution of glass sand grains and removal of bubbles in flat glass of the float type working at a medium temperature of 1400 °C and with required melting performance of at least 20t/24h. With regard to identified high specific capacities obtained on the previous model space with preset temperatures, a relatively small melting space was designed, in the shape of a block with a length of 2 m, width 1 m and the height of the level of glass melt 6 being 0.5 m. A detailed chart of this space is in Figure 6. The melting space is preceded by the space for creation of glass melt, which is not the subject matter of this invention. The walls of the melting space are composed of refractory material layers, the crown 5 is low, slightly vaulted. The input 1_2 is placed in the entire width of the melting space, either at the bottom I, or at the level of the glass melt 6 and occupies 20 % of the charging front wall 2. The output 13 is then formed of a through-flow at the level of glass melt 6 in the entire width of the melting space, whose area represents also 20 % of the rear through-flow wall. The melting space is heated by means of a longitudinal row 6 of molybdenum electrodes JjO installed from the bottom i in the melting space axis. The electrodes K) are connected in the way when the row of the electrodes is divided into three couples and each couple is assigned one transformer. Thus it is possible to divide the electrical capacity of the electrodes 1_0 into three zones and to set the required capacity, and therefore also temperature gradients. This approach makes it possible to regulate energy inlet into the glass melt in individual parts of the furnace and to adjust the nature and intensity of natural flowing in the melting space.

Figure 6 represents a detailed chart of the designed space for melting of flat glass with the upper input J_2- In the first phase of modelling, the distribution of energy set on electrodes _10 simulated the type of flowing usual in classical glass furnaces (reference setting). The input 12 into the melting space was at the level of glass melt 6. Through concentration of energy on the electrodes no. 5 and 6 it was possible to simulate the cross thermal barrier, which calls out usual longitudinal circulation flowing. Distribution of energy on electrodes in % is given in Table 1.

Table 1

Energy distribution on electrodes J_0 according to their numbering in Figure 6 at a reference setting of heating and with the upper input Y2 of glass melt 6.

The nature of flowing was checked so that it can correspond to the nature known from classical glass furnaces.

Figure 7 shows resulting velocity distribution in a longitudinal axial section of the melting space, illustrated through sections of trajectories of glass melt made for 30 s, representing the reference case. The process of dissolution of sand grains was selected as the process used for behaviour of the space under conditions simi lar to the classical type of flowing (reference case).

Also projections of the critical trajectory and other disadvantageous trajectories for dissolution were obtained - these projections into the longitudinal axial vertical section of the melting space are provided for in Figure 8 which illustrates the reference case.

The resulting distribution of melt velocities and shapes of trajectories correspond to the results obtained on classical devices, the reference case is therefore correctly set from a qualitative point of view.

For sand dissolution, experimental data was obtained through repeated laboratory melting of glass and calculation of undissolved sand grains in resulting samples of glass acquired in the time sequence at temperatures from 1300 to 1500 °C and in intervals by 50 °C.

The inventors of this invention newly found an empiric equation for the mean rate of dissolution of sand grains with the initial maximum grain diameter of 0.5 mm as follows:

v,j,,_s = 1.56xl 0^"18exp(0.0 l 53T) [ms ¹] [36]

where T is temperature in K.

Concerning removal of bubbles, a value of the mean bubbles growing rate depending on the temperature was found out in an experimental way. For the mean rate of bubbles diameter growth, the inventors of this invention newly found, for this invention and by extrapolation experimental points, an empiric equation as follows:

= exp(120.34 - 3.80xl0⁵/T + 2.608xl0⁸/T²) [m/s]

where T \ ^'s temperature in .

Both the equations were applied to the sand dissolution process and to the process of removal of bubbles taking place in the melting space with temperatures and flowing set for the reference case and later to other cases as well.

The following values were obtained by modelling the final variant describing the classical setting, complying with the average temperature of 1400 °C and by subsequent processing.

Glass sand dissolution process in the reference case:

Average temperature: 1400 °C.

Melting space dimensions: 2x1 x0.5 m - length, width, height.

input 12: from the upper side.

Electrical heating power input E: 1 18.6 kW.

Critical melting performance V: 6.79x l 0^"5 mV = 13.65 t/24h = 6.82 t/(24h m²).

Average glass melt residence time in the space f: 13 888 s.

Average sand grain dissolution time f_D : 245 1 s.

Geometric glass melt residence time T_C = V/V: 14724 s.

Fraction of the dead space in the glass melt circulation process 0.050.

Fraction of the dead space after dissolution of sand grains W/ 0.825.

Fraction of the use of the melting space u_D: 0.166.

Specific heat losses H^L = ET_G/(Vp): 751 kJ/kg.

The reference specific melting performance of the melting module found out at 6.82 t/(24h m²) is high in comparison with specific melting performance of ordinary glass furnaces, it is however necessary to take into consideration that in the given case the preset medium temperature of 1400 °C is relatively high, the matter concerns the highest possible melting performance - critical - when the space does not have any melting reserve available and the given melting performance concerns only completion of dissolution of sand grains. With the 50 % reserve, the specific melting performance is only 3.41 t/(24h m²) and this value is already realistic. The same situation applies to specific heat losses, which would make - at a 50 % reserve - 1502 kJ/kg, which is also a realistic value with regard to the high average melting temperature of 1400 °C.

The identified fraction of the dead space me is at a given type of flowing much lower than expected and it is often found out e.g. through measurement of the transition characteristic of the furnace. The inspection of the curve of distribution of the time of residence of glass melt in the furnace, however, did not reveal any irregularities on this curve which could reveal recycling of some ongoing trajectories for the reason of numerical errors of calculations (the value τ is increasing and the calculated value ma is decreasing). The reason for low values of the dead space m_G is the fact that the cross temperature barrier was set only by means of the electric heating power input on axially situated electrodes J_0. This already called out a certain portion of cross flowing, which led to a decrease in the values of w_c. The utilisation of the melting space would be still lower in the case of creation of the classical cross temperature barrier.

Besides this, energy was set on individual electrodes J_0 in such a way that it can be possible to achieve the required type of the spiral-type flowing.

During the final setting, see Table 2, it is possible to observe essential changes in the nature of flowing of glass melt and a typical spiral nature of critical and close trajectories of glass melt, as shown by Figures 9 and 10. Figure 9 shows the resulting velocity distribution in a longitudinal axial section of the melting space illustrated by sections of trajectories of glass melt made for 30 s, and represents an optimum case. Figure 10 illustrates projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains in the longitudinal axial vertical section of the space and represents an optimum case.

Table 2

Distribution of energy on electrodes J_0 according to their numbering in Figure 6 at the optimum variant of the setting of heating and at the upper input 12.

The end variant with distribution of electrical heating power inputs stated in Table 2 provided an optimum result:

Average temperature: 1400 °C.

Dimensions of the melting space : 2x1 x0.5 m - length, width, height. Input 12: upper input.

Electrical heating power input E: 1 19.5 kW.

Average cross temperature gradient 45 K/m, average longitudinal gradient 8 K/m, average ratio of the cross gradient to longitudinal gradient 5.6.

Critical melting performance V: 2.63x1ο^"4 mV = 52.86 t/24h = 26.43 t/(24h m²).

Average residence time of glass melt in the space f: 3760 s.

Average time of dissolution of sand grains f_D: 2498 s.

Geometric residence time of glass melt T_C = V/V 3807 s.

Fraction of the dead space of glass melt circulation ma-' 0.012.

Fraction of the dead space after dissolution of sand grains m_D: 0.336.

Fraction of the use of the melting space u_D: 0.656.

Specific heat losses H^L = Er_G/(Vp): 196 kJ/kg.

This result already indicates an essential improvement for the reason of setting of a more efficient type of flowing in the melting space. The use of the melting space WQ increased, in comparison with the reference case 3.95x, the melting performance of the space increased 3.87x and specific heat losses dropped 3.83x. While considering a 50 % melting reserve, the space performance would be 26.43 t/24h and the specific dissolution capacity would be 13.22 t/(24h m²). The conditions for the required capacity of the melting space were met.

While operating the system with the optimised flowing of glass melt, there will be a problem consisting in occasional leakage of non-molten charge from the level of previous space through the upper inflow into the dissolution space. This issue was resolved by placing the input 12 in the area at the bottom 1 of the dissolution space, as shown by Figure 1 1. Figure 1 1 illustrates a detailed chart of the proposed space for melting of flat glass with a lower input.

The achievement of the optimum setting of flowing will be demonstrated for this case with the help of two variants, the first variant and the last one - optimum. During the first setting it was necessary to build upon results achieved on the original model device with the preset temperature field [33-34]. In the first variant, virtually the same electrical power heating input was set for all electrodes 10, as shown by Table 3. Table 3

Energy distribution on electrodes _10 according to their numbering in Figure 1 1 during the first variant of setting of heating and at the upper input Y2.

The results of the calculations of the first variant are as follows:

Average temperature: 1400 °C.

Dimensions of the melting space 2x 1 x0.5 m - length, width, height.

Input 12: lower input.

Electrical heating power input E: 122.9 kW.

Average cross temperature gradient 40 K/m, average longitudinal gradient 8K/m, average ratio of the cross gradient to longitudinal gradient 5.

Critical melting capacity V: 2.49x 1 0^"4 mV = 50. 1 1 t/24h = 25.05 t/(24h m²)

Average residence time of glass melt in the space f: 3988 s.

Average time of dissolution of sand grains f_D : 2531 s.

Geometric residence time of glass melt r_c = V/V: 401 1 s.

Fraction of the dead space of glass melt circulation m_G: 0.006.

Fraction of the dead space after dissolution of sand grains m, 0.365.

Fraction of the utilisation of the melting space u_f . 0.631 .

Specific heat losses H^L = ET_G/(Vp) 212 kJ/kg.

The nature of flowing with a change in placement of the input J_2 has not changed.

Modelling continued until achievement of the optimum case.

The optimum case is illustrated with the help of Table 4 and Figures 12 and 13. Figure 12 shows resulting velocity distribution in a longitudinal axial section of the melting space illustrated by sections of trajectories of glass melt made for 30 s, as an optimum case. Figure 13 illustrates projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains into the longitudinal axial vertical section of the space, which represents an optimum case:

The setting of the flow of energies in an optimum case is in Table 4. Table 4

Energy distribution on electrodes J_0 according to their numbering in Figure 6 when the optimum heating is set and the input J_2 is at the bottom I .

The calculation results for the last - optimum - variant are as follows:

Average temperature: 1400 °C.

Dimensions of the melting space 2x1 0.5 m - length, width, height.

Input 12: lower input.

Electrical heating power input s 124.2 kW.

Average cross temperature gradient 45 K/m, average longitudinal gradient 7 K/m, average ratio of the cross gradient to longitudinal gradient 6.4.

Critical melting performance V: 2.54xl0^"4 m ¹ = 51 .05 t/24h = 25.53 t/(24h-m²).

Average residence time of glass melt in the space f: 3915 s.

Average time of dissolution of sand grains f₀ : 2545 s.

Geometric residence time of glass melt T_C = V/V: 3943 s.

Fraction of the dead space of glass melt circulation m_c: 0.007.

Fraction of the dead space after dissolution of sand grains m_D: 0.350.

Fraction of the utilisation of the melting space u_D: 0.646.

Specific heat losses ^L = Ez_G/(Vp) : 210 kJ/kg.

A comparison of the first and optimum variants shows that already the first presupposed variant differs only little from the optimum variant. The results are therefore little sensitive to small changes if there is a substantiated idea of what can be achieved.

The comparison was made with the use of the values from the reference case with an inflow situated at the level. According to the calculation, the use of the space increased 3.89x, the melting performance increased 3.74x and specific heat losses dropped 3.58x. The performance would achieve, at a 50 % melting reserve, 25.53 t 24h and the specific performance would be 12.76 t/(24h m²). The values are only slightly worse than the ones in the case with the inflow at the level, the space performance exceeds the 20t 24h required and the given last arrangement can be used with both the inputs. In the following steps, a calculation of conditions and outputs of the optimised type of flowing for removal of bubbles was carried out. The critical state corresponding to a maximum melting performance of the space for removal of bubbles was found. Energy distribution on electrodes J.0 is provided for in Table 5. Resulting velocity distribution in the central longitudinal section is then stated in Figure 14, and the critical trajectory with several other similar trajectories of the smallest bubbles with a diameter of 0.1 mm is shown in Figure 15. At a more detailed level, Figure 14 illustrates resulting velocity distribution in a longitudinal axial section of the melting space, illustrated through sections of trajectories of the glass melt made for 30 s, which represents an optimum case for removal of bubbles. Figure 16 illustrates projections of the critical trajectory and other trajectories which are the slowest ones for bubbles with a diameter of 0.1 mm to the longitudinal axial vertical section of the space, which represents an optimum case of removal of bubbles from the glass melt 6.

Table 5

Energy distribution on electrodes _10 in the optimum case of removal of bubbles according to their numbering in Figure 6 at the input V2 at the bottom.

The results of calculation of the optimum variant for removal of bubbles are as follows: Average temperature: 1400°C

Dimensions: 2x1 x0.5 m.

Input 12: lower input.

Electrical heating power input E: 126.4 kW.

Critical melting performance V: 3.44xl0^"4 m³/s = 69.12 t/24h = 34.56 t/(24h m²)

Bubbles residence time on the critical trajectory T_F : 2168 s.

Geometric residence time of the glass melt r_G = V/V: 2908 s.

Fraction of the dead space of the glass melt circulation m_virl: 0.254.

Virtual height h_virl: 1.18 m.

Fraction of the utilisation of the melting space »_/. : 0.561.

Specific heat losses H^L = Er_G/(Vp) : 1 58 kJ/kg. The results of calculations of removal of bubbles indicate that the optimum melting performance of 3.44x l0^"4 m /s is substantially higher than the optimum melting performance for dissolution of sand grains, which was 2.54x10^"4 m³/s. In the event of setting the melting performance and nature of flowing to the case with a melting performance value of 2.54x10⁴ m³/s, the bubbles will be probably removed still before the output from the melting space, and dissolution of sand grains will be the controlling operation of the melting process at a simultaneous course of both the processes, see Figure 16.

This means that the setting of the flowing valid for dissolution of sand grains in Table 4 will apply as the final setting for the simultaneous course of both the processes. The results then open the possibility of increasing the melting performance of the space by means of a measure which could accelerate kinetics of dissolution of sand grains, but this goes beyond the framework of this application.

E x a m p I e 2

(Fig. 1 7-23)

Special space for dissolution of glass sand grains and removal of bubbles in glass for production of spectacle mouldings operating at an average temperature of 1300 °G and with required melting performance of at least 8t/24h.

The melting process can use an identical special space with the length of 2 m, width of 1 m and level layer height of 0.5 m, as in example 1 , the input 12 into the melting space is placed at the bottom I (see Figure 14), heating is again electrical, the heat sources, however, make it possible to achieve the average temperature for dissolution of only 1300 °C. With regard to a similar composition of glass and identical raw materials used, the same dependence of the time of dissolution of sand grains on temperature as in Example 1 was used. A reference case was dealt with as well, even though it was supposed that the decrease in the average temperature would not have any essential influence on the utilisation of the melting space [34]. The resulting reference variant brought the following distribution of electrical heating power input on electrodes 10.

Table 6

Energy distribution on electrodes 10 according to their numbering in Figure 6 at the reference setting of heating.

Electrodes 1 +2 3+4 5+6

Electrical heating power

0 0 106.5

input [kW] The resulting type of flowing and the course of critical trajectories in fact did not differ from flowing at a temperature of 1400 °C at the reference setting, as shown by Figures 17 and 18. Figure 19 illustrates resulting velocity distribution in a longitudinal axial section of the melting space, illustrated through sections of trajectories of glass melt made for 30 s, as a reference case. Figure 20 shows projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains to the longitudinal axial vertical section of the space, as a reference case.

The results of the solution of this reference case are as follows:

Average temperature: 1300 °C.

Dimensions of the melting space: 2x1 xO.5 m - length, width, height.

Input 12: lower input.

Electrical heating power input £. 106.3 kW.

Critical melting performance V: 3.78xl 0^"6 mV = 0.76 /24h = 0.38 t/(24h m²)

Average residence time of glass melt in the space τ: 63206 s.

Average time of dissolution of sand grains f_D : 101 1 s.

Geometric residence time of glass melt r_G = V/V: 264355 s.

Fraction of the dead space of glass melt circulation m_c: 0.761.

Fraction of the dead space after dissolution of sand grains m_D: 0.840.

Fraction of the use of the melting space u : 0.038.

Specific heat losses H^L = Ez_G/(Vp): 12080 kJ/kg.

The case is hig ly disadvantageous, a large negative role being played by the lower input into the space. There is already a large part of the dead space ntc, the losses is extraordinarily high as well, the critical trajectory and other trajectories are near the bottom \ at low

temperatures, and the melting performance is consequently very low. Due to its extreme disadvantageousness, this case cannot serve as a reference case, nevertheless it shows an extraordinary influence of an unsuitable type of flowing on effectiveness of the dissolution process.

In the next steps, the electrical heating power inputs on electrodes jO were set, which can create the required type of optimum flowing in the melting space. Distribution of electrical heating power inputs is provided for in Table 6, while velocity distribution and critical trajectory together with similar trajectories are provided for in Figures 16 and 1 7. Table 7

Energy distribution on electrodes JO according to their numbering in Figure 6 at the optimum setting of heating.

Figure 16 illustrates the resulting velocity distribution in a longitudinal axial section of the melting space, illustrated through sections of trajectories of glass melt made for 30 s, which is the optimum case.

Figure 17 illustrates projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains to the longitudinal axial vertical section of the space, which represents the optimum case.

The optimum variant solution results are as follows:

Average temperature: 1300 °C.

Dimensions of the melting space: 2x1 xO.5 m - length, width, height.

Input 12: lower input.

Electrical heating power input E: 105.2 kW.

Average cross temperature gradient 50 K/m, average longitudinal gradient 7 K/m, average ratio of the cross gradient to the longitudinal gradient being 7.1 .

Critical melting performance V: 5.59xl0^"5 mV = 1 1.23 t/24h = 5.62 t/(24h m²)

Average residence time of glass melt in the space r. 17025 s.

Average time of dissolution of sand grains f_D : 10599 s.

Geometric residence time of glass melt T_G = V/V: 1 7895 s.

Fraction of the dead space of glass melt circulation m_G: 0.049 s.

Fraction of the dead space after dissolution of sand grains m_D: 0.378.

Fraction of the utilisation of the melting space u_D: 0.592.

Specific heat losses H^L = ET_G/(Vp) 809 kJ/kg.

The resulting values meet the melting performance presumption, which achieved, at the presupposed 25 % melting reserve, the value of 8.42 t/24h for the entire system, i.e. 4.21 t/(24h m²). Specific heat losses in kJ/kg is again very low according to expectations. This was followed by the searching for a type of flowing, during which it could be possible to find the highest use of the space and therefore also the largest melting performance for the process of removal of bubbles. The capacity distribution into individual electrodes 10 in the identified optimum case is given by Table 8. Figure 21 then brings resulting velocity distribution in a longitudinal axial section of the melting space illustrated through sections of trajectories of the glass melt made for 30s, for an optimum case. Figure 22 illustrates an optimum case for projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains to the longitudinal axial vertical section of the space.

Table 8

Energy distribution on electrodes 10 according to their numbering in Figure 6 at the optimum setting of heating at removal of bubbles.

The results of solution of the optimum variant for removal of bubbles are as follows: Average temperature: ! 300°C

Dimensions: 2x1x0.5 m.

Input 12: lower input.

Electrical heating power input E: 102.9 kW.

Critical melting performance V: 4.10 l 0^"5 mVs= 8.29 t/24h = 4.15 t/(24h m²)

Time of removal of critical bubbles: r_/. = 17248 s.

Geometric residence time of the glass melt T_c = V/V: 24377 s.

Virtual height h_vin: 1.733 m.

Virtual dead space m_virl: 0.292.

Fraction of the utilisation of the melting space u_/. : 0.532.

Specific heat losses H^L = Er_G/{Vp) : 1072 kJ/kg.

The comparison of capacities for an optimum case of dissolution of sand and removal of bubbles illustrates that the slowest process is removal of bubbles. For this reason, the flowing set for the case with an optimum removal of bubbles was used for dissolution of sand grains, see Figure 23. Figure 23 illustrates projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains to the longitudinal axial vertical section of the space.

E x a m p l e 3

(Fig. 24 - 30)

It is necessary to suggest a special space for dissolution of glass sand and removal of bubbles in white package glass melt, which is to achieve a melting performance of at least 300 t/24h at the achievable average melting temperature 1425 °C. Necessary energy was to be supplied according to the needs by both gas burners JJ_ and electrodes H), but it was proven that the use of electrodes J_0 only was sufficient to cover the heat losses. The melting space with a length of 6.57 m and a width of 2 m at the thickness of the layer of the glass melt of 1 m, with a total volume of 13.14 m³ will be designed for dissolution. The chart of the space with the designed heating with electrodes 1 0 in both reference and optimised cases and with the input 12 in the lower part of the input front is in Figure 24a (vertical section of the longitudinal axis) and 24b (horizontal section below the level). At a more detailed level, Fig. 24a, 24b represent a chart of the proposed melting space with the heating with electrodes J_0 and placement of electrodes K) in sections XZ (a) and XY (b).

Table 9 illustrates energy distribution of electrodes J_0 according to their numbering in the Figure, with regard to their sources in a reference arrangement in the first optimised case.

Table 9

Energy distribution on electrodes J_0 according to their numbering in Figure 24a at the reference and the first optimised variant of setting of heating.

Resulting velocity distribution in a longitudinal axial section of the melting space illustrated through sections of trajectories of glass melt 6 made for 30 s in the reference case is in Figure 25 and projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains, to the longitudinal axial vertical section of the space, also in the reference case, are in Figure 26. Figure 25 illustrates resulting velocity distribution in a longitudinal axial section of the melting space illustrated through sections of trajectories of glass melt made for 30 s, as a reference case.

Figure 26 illustrates projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains to the longitudinal axial vertical section of the space, as a reference case.

The results of the reference case are as follows:

Average temperature: 1425 °C.

Input J_2: lower input.

Dimensions of the melting space 6.57x2x 1.0 m - length, width, height.

Electrical power input of heating systems: 471 .0 kW.

Critical melting performance V: 2.20x10^"4 mV = 44.1 t/24h = 3.35 t/(24h m³)

Average residence time of glass melt in the space f: 36450 s.

Average time of dissolution of sand grains f_D : 2433 s.

Geometric residence time of glass melt T_C = V/V: 59875 s.

Fraction of the dead space of glass melt circulation ma'. 0.391.

Fraction of the dead space after dissolution of sand grains m_D: 0.933.

Fraction of the utilisation of the melting space u_D: 0.041

Specific heat losses H^': 924 kJ/kg.

The resulting values characterise a classical glass furnace well. The relatively high specific melting performance is again given by the fact that the matter concerns a critical (maximum possible) performance of the space and the process includes dissolution of sand grains only.

In the following steps, new distribution of energy supplied to electrodes 10, which is provided for in Table 9 for an optimum case, was designed on the basis of previous calculations on a model melting space and on the basis of the experience obtained. The aim was to call out again the required spiral-type flowing of glass melt. Resulting velocity distribution in a longitudinal axial section of the melting space illustrated through sections of trajectories of glass melt made for 30 s in the first optimised case is in Figure 27 and projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains to the longitudinal axial vertical section of the space also in the first optimised case are provided for in Figure 28. The results of the first optimised case are as follows:

Average temperature: 1425 °C.

Input 12: lower input.

Dimensions of the melting space 6.57x2 1 .0 m - length, width, height.

Electrical heating power input of heating systems: 492.3 kW.

Average cross temperature gradient 55 /m, average longitudinal gradient 10 K/m, average ratio of the cross gradient to longitudinal gradient being 5.5.

Critical melting performance V: 2.97 10^"3 mV = 597 t/24h = 45.4 t/(24hm³).

Average residence time of glass melt in the space τ: 3994 s.

Average time of dissolution of sand grains f_D : 1 795 s.

Geometric residence time of glass melt r_c = V/V: 4419 s.

Fraction of the dead space of glass melt circulation m_G: 0.096.

Fraction of the dead space after dissolution of sand grains m_D: 0.551 .

Fraction of the utilisation of the melting space u_D: 0.406

Specific heat losses H¹ 71 .2 kJ/kg.

Resulting values provide an unexpectedly high dissolution melting capacity of the melting space. This is caused by the increase in the use of the melting space by one order against the reference case. Lower values of the average time of dissolution of sand grains show that dissolution in the optimised case took place also on more advantageous temperature conditions. The consequence of the high melting performance of the device is a very low specific heat losses. If an approximately 16 % melting reserve is admitted, the given space would provide a dissolution capacity of 500 t/24h. The value of the use 0.41, however, illustrates that it is still possible to optimise flowing in the space.

In the following steps, the distribution of the melting capacity on electrodes JO was modified in such a manner that it can be possible to extend the glass melt residence residence time which corresponds to the critical trajectory. Another two cases with an optimised energy distribution to individual electrode couples were calculated. By means of the second variant, an essential improvement of the use was achieved, and the third variant differed only little from the previous one, which means that the use was near to an optimum value. The third variant was therefore considered as the final one. The distribution in the third optimised variant is stated in Table 10. Table 10

Percentage energy distribution on electrodes J_0 according to their numbering in Figure 24a at the reference and optimised variants of the setting of heating for dissolution of sand grains.

Resulting velocity distribution in a longitudinal axial section of the melting space and projections of the critical trajectory and other trajectories which are the slowest ones for dissolution of sand grains to the longitudinal axial vertical section of the space for the third optimised variant are very similar to the first optimised case.

The results of the third optimised case are as follows:

Average temperature: 1425 °C.

Input 12: lower input.

Dimensions of the melting space 6.57x2x 1 .0 m - length, width, height.

Electrical heating power input of heating systems: 507 kW.

Average cross temperature gradient 55 K/m, average longitudinal gradient 8 K m, average ratio of the cross gradient to longitudinal gradient 6.9.

Critical melting performance V: 4.39x10 3 m V = 881 t/24h = 67 t/(24h m³)

Average residence time of glass melt in the space f: s.

Average time of dissolution of sand grains f_D : 1 782 s.

Geometric residence time of glass melt T_G = V/V: 2994 s.

Fraction of the dead space of glass melt circulation m_G: 0.164.

Fraction of the dead space after dissolution of sand grains m₀: 0.288.

Fraction of the utilisation of the melting space u_D: 0.595

Specific heat losses: tt: 49.7 kJ/kg.

Resulting values state an extraordinarily high melting performance and very low specific heat losses. The proposed space with optimised spiral flowing therefore features a very high capability of dissolving the glass sand even after considering a significant melting reserve. The case illustrates that the dissolution of glass sand in the given space under given conditions with a high probability cannot become a process limiting intensity of the entire melting process. In the following steps, an optimum case of flowing for removal of bubbles was calculated. The resulting energy distribution on electrodes is provided for in Table 1 1.

Table 11

Percentage distribution of electrical heating power input on electrodes K) for an optimum case at removal of bubbles.

The resulting velocity distribution in a longitudinal axial section of the melting space illustrated through sections of trajectories of the glass melt made for 30 s in the first optimised case is in Figure 29, and projections of the critical trajectory and other trajectories which are the slowest ones for removal of bubbles to the longitudinal axial vertical section of the space, also in the first optimised case, are in Figure 30.

The results of the optimised case for removal of bubbles are as follows:

Average temperature: 1425°C

Dimensions 7.83x2x1.0 m.

Input 12: lower input.

Electrical heating power input E: 507 kW.

Critical melting performance V: 4.39x10^"3 m Vs = 881 t/24h = 67 t/(24hV)

Time of removal of critical bubbles: T r = 2228 s. Geometric glass melt residence time r_G = V/V: 3590 s.

Virtual height h_virt: 1.328 m.

Virtual dead space m_v/rl: 0.379.

Fraction of the utilisation of the melting space «_/.: 0.602.

Specific heat losses H¹: 46.6 kJ/kg.

In both the cases it was possible to achieve the same melting performance under the same conditions, only the target point for removal of bubbles was moved to a distance of 7.83 m, which was permitted for the arrangement of the furnace, without the glass quality being endangered. The setting of the melting space is therefore suitable for both the processes.

E x a m p l e 4

(Fig. 31 , 32)

This example concerns a melting space, heated with three couples of cross gas burners _Π for the melting of soda-lime glass and processing by moulding for domestic glass products. In its melting part, the system features a length of 6 m and a width of 2 m. The height of the level of glass melt 6 is 0.6 m. The required melting performance is to increase substantially to a value about 30t/24h. The average melting temperature oscillates about 1400 °C. A chart of one half of the original system is shown in Figure 3 1 . Figure 3 1 illustrates an original glass furnace in the longitudinal axial section. In the original arrangement, the heating energy was distributed in individual burners JJ_ according to Fig. 3 1. The numbering in the Table is according to the Figure from left to right.

Energy distribution in three couples of gas burners J_l according to Figure 3 1 from left to right at the original setting of burners J_l and the reference case is as follows:

Electrical heating power input of the first couple of burners _Π : 637 kW.

Electrical heating power input of the second couple of burners JT : 700 kW.

Electrical heating power input of the third couple of burners JJ.: 750 kW.

Total electrical heating power input of burners: 2087 kW.

Electrical heating power input of electrodes: 0 kW.

The results of the reference case are as follows:

Average temperature: 1400 °C.

Input 12: With a charge on the level.

Dimensions', length 7.2 m in total, length of the melting part 5.2 m, width 1.65 m, depth of the glass melt 6 being 0.6 m.

Critical melting performance V: 6.47x l0^"5 mV' = 2.52 t/(24h m³) ==13 t 24h.

Geometric residence time of glass melt T_G = V/V: 79598 s.

Fraction of the utilisation of the melting space u_D 0.059.

Specific energy consumption tf¾: 13866 kJ/kg.

Resulting values indicate a small use of the space, to which a small melting performance of the system corresponds as wel l. In the current arrangement of burners FL however, it was not possible to influence the nature of flowing in a desirable direction in a more essential way. For this reason, a renovation of the heating system was made, consisting in placement of 14 vertical electrodes J_0 from the bottom in the longitudinal axis of the melting space. Thus it was possible to create the necessary longitudinal temperature barrier in the molten glass mass and the desirable type of spiral flowing of glass melt. Dissolution of sand grains was considered as a critical process. Resulting arrangement of the melting process in the optimised case is provided for in Figure 32, and the energy distribution in burners ϋ and electrodes K) is provided for in the following text. The necessary transverse spiral-type flowing was achieved through the new setting in the field of electrodes J_0.

The energy distribution in burners II according to their arrangement in Figure 32 in the optimised setting of burners J_I and electrodes J_0 is as follows:

Power input of electrodes 10: 1400 kW:

Electrodes 1 -2 -^308 kW;

Electrodes 3-4 - 266kW;

Electrodes 5-6 - 210kW;

Electrodes 7-8 - 168kW;

Electrodes 9- 10 -> H 2kW;

Electrodes 1 1 - 12 ^70 kW;

Electrodes 13- 14 - 70kW;

Electrodes 15- 18 -»98kW.

Power input of burners: 16.8 kW

The results of the optimised arrangement are as follows:

Average temperature: 1425 °C

Input 12: With a charge on the level.

Dimensions of the melting space: length 7.2 m in total, length melting part 5.2 m, width 1 .65 m, depth of glass melt 0.6 m.

Electrical heating power input of heating systems: 1400 kW on electrodes 10, 16.8 kW on burners _Π .

Average cross temperature gradient 55 K/m, average longitudinal gradient 12 K/m, average ratio of the cross gradient to longitudinal gradient 4.6.

Critical melting performance V: 1 .8 1 l 0^"4 mV = 7.05 t/(24h m³) = 36.3 t/24h

Geometric residence time of glass melt r_G = V/V: 28453 s.

Fraction of the utilisation of the melting space »_/> 0.239.

Specific energy consumption H : 3335 kJ/kg.

The melting performance of 30t/24h can be achieved at the given setting according to Figure 32 at about a 1 7% reserve. This reserve is not too high, but at a stabilised operation it is sufficient. Specific melting performance 3.02 t/(24h m²) is 2.8x higher than in the case of reference setting, it is true, but it is relatively low in comparison with the cases provided for above. A problem concerns the existence of a glass batch layer on the level area, which does not permit full development of the requested spiral-type flowing. The furnace, however, works in both cases inefficiently from the viewpoint of energy consumption; in the optimised case it is possible to exclude burners U_. and to operate the furnace as an all-electric device.

Extraordinarily high dissolution capacities of the designed space were achieved in most cases stated in the examples. Nevertheless, the results cannot be automatically related to the entire melting process. This concerns implementation of only one melting process, dissolution of sand grains, in the situation when the melting space does not have any reserve. The melting performance of such special melting space can then be further reduced, but not substantially [31], if also the refining process is considered as another process. In the most substantial way, however, it is possible to reduce the melting space melting performance in the case when also the conversion of the batch to glass takes place in the same space, which is finally obvious from results in Example 4. The fact is that as soon as limitations of the melting performance are cancelled by introduction of the optimised flowing of glass melt, it is just conversion of the charge that becomes the control part of the entire melting process, and the entire dissolution melting performance of the given space cannot be used unless the rate of the charge conversion is increased. And yet it is possible to expect higher specific melting capacities (or even significantly higher than 10 t/m 24h). The main effort should then be devoted to an increase in the capacity of conversion of batch into glass.

Reference markings

1 Bottom

2 Front wall

3 Rear bridge wall

4 Side walls

5 Crown

6 Glass melt

7 Submerged cross refractory barrier

9 Cross row 9 of energy sources

10 Electrodes

1 1 Burners

12 Input

13 Output

Claims

C L A I M S

1. Method for continuous glass melting under controlled convection of glass melt, containing undissolved particles, especially glass sand and bubbles, taking place in a horizontally oriented continual melting space of a glass furnace, where the defined melting space is outlined by the height of glass melt (6), width between opposite side walls (4) and length between the front wal l (2) and the submerged cross refractory barrier (7) immersed in the glass melt (6), or the length between the front wall (2) and the cross row (9) of energy sources; at the same time the glass melting space includes energy sources, such as in glass industry used burners (1 1 ), heating electrodes ( 10) and other suitable heating energy sources, characterised in that

energy sources, such as heating electrodes ( 10) and/or in glass industry used burners ( 1 1 ) or other suitable heating sources operate onto the molten glass melt (6), containing undissolved particles, especially glass sand and bubbles, in the longitudinal axis of the melting space, or in a direction parallel with this longitudinal axis until creation of one or more longitudinal temperature barriers in the glass melt (6) and until arising of a cross temperature gradient [K.m ¹], which generates the spiral-type flowing of the glass melt (6) with a rotary movement across the melting space, and in fact perpendicularly to the longitudinal axis of the melting part,

this spiral-type flowing continues in the direction from the front wall (2) to the submerged cross refractory barrier (7) in the glass melt (6), or in the direction from the front wall (2) to the cross row (9) of the energy sources,

and the cross temperature gradient [K.m ¹] of each spiral-type flowing will always be set as higher than longitudinal temperature gradient [K.m^"'] between the front wall (2) and the submerged cross refractory barrier (7) in the glass melt (6), or between the front wall (2) and the cross row (9) of the energy sources, as a consequence of which 0.6 to 0.8 multiple of the melting space is utilized from the total melting space.

2. Method for continuous glass melting under controlled convection of glass melt according to Claim I , characterised in that the ratio of the cross temperature gradient [K.m ¹] to the longitudinal temperature gradient [K.m ¹] is higher than 1 and lower than 30.

3. Method for continuous glass melting under controlled convection of glass melt according to Claim 2, characterised in that the ratio of the cross temperature gradient [K.m ¹] to the longitudinal temperature gradient [K.m^"'] is within the range from 5 to 20.

4. A glass furnace for continuous glass melting under controlled convection of glass melt, during which energy sources affect on molten glass containing undissolved particles, especially glass sand and bubbles, and energy sources situated in the horizontally oriented continual and separated melting space; where the defined melting space is outlined by the height of glass melt (6), width between opposite side walls (4) and length between the front wall (2) and the submerged cross refractory barrier (7) immersed into the glass melt (6), or the length between the front wall (2) and the cross row (9) of energy sources; and the glass melting space includes energy sources, such as in glass industry used burners ( 1 1 ), heating electrodes ( 10) and other suitable heating energy sources, according to some of the Claims 1 to 3, characterised in that energy sources, such as heating electrodes ( 10) and/or industrial burners (1 1 ) or other suitable heating sources are arranged in the horizontally oriented continual and separated melting space, between the front wall (2) and the submerged cross refractory barrier (7) in the glass melt (6) or the cross row (9) of the energy sources in the melting part of the glass melt (6), for creation of one or more longitudinal temperature barriers in the glass melt (6) and for generation of the spiral-type flowing of glass melt (6) with a rotary spiral-type motion across the melting part, essentially perpendicularly to the longitudinal axis of the melting part, for the setting of the cross temperature gradient [K.m ¹] of each spiral-type flowing higher than the longitudinal temperature gradient [K.m ¹], with an advantage in the ratio of the cross temperature gradient [K.m^"'] to the longitudinal temperature gradient [K.m ¹] being higher than 1 and lower than 30, preferably in the ratio from 5 to 20, as a consequence of which 0.6 to 0.8 multiple of the melting space is utilized from the total melting space.

5. A glass furnace according to Claim 4, characterised in that vertical heating electrodes (10) for creation of one or more longitudinal temperature barriers in the glass melt (6) are placed in one row in the longitudinal axial section of the melting space,

or from both the sides of the longitudinal axial section of the melting space

and/or in a direction parallel with this section

and/or are installed along one or two side walls (4) of the melting space.

6. A glass furnace according to Claim 4, characterised in that the side heating electrodes (10) for creation of one or more longitudinal temperature barriers in the glass melt (6) are placed in one or both side walls (4) of the melting space, whose tips reach the area of glass melt (6) situated along the longitudinal axial section of the melting space or reach into vicinity of the side walls (4) with their body.

7. A glass furnace according to Claim 4, characterised in that flame burners ( 10) for creation of one or more longitudinal temperature barriers in the glass melt (6) are placed transversally with the side walls (4) of the melting space and are directed substantially within the area of the longitudinal axis of the level of glass melt (6), or to the area of the level of glass melt (6) along one of the side walls (4) of the melting space.

8. A glass furnace according to Claim 4, characterised in that vertical burners (1 1) for creation of one or more longitudinal temperature barriers in the glass melt (6) are placed

in the crown (5) of the melting space and are directed substantially to the vicinity of the longitudinal axis of the level of glass melt (6) or to the area on the level of the glass melt (6) along one or both the side walls (4) of the melting space.

9. A glass furnace according to Claim 4, characterised in that U-flame burners for creation of one or more longitudinal temperature barriers in the glass melt (6) are placed

in the front wall (2) or rear bridge wall (3) of the melting space and are directed to the area of the longitudinal axis of the level of the glass melt (6) or to the area on the level of the glass melt (6) along one or both the side walls (4) of the melting space.

References

1. Danckwerts, P.V.: Chem. Eng.Sci. 2, 1 (1953).

2. Cooper, A.R.: J. Am. Ceram. Soc. 42, 93 (1959).

3. Smrcek, J.: Sklaf a Iter. 23, 270, 304 (1973).

4. Cooper, A. R.: J. Am. Ceram. Soc. 43, 97 ( 1960).

5. Moult, A.: Glass Technol. 23, (2), 106- 1 12 (1982).

6. Mase, H., Oda, K.: J. Non-Cryst. Solids 38,39, 807-812 (1980).

7. Layens, G.: Glastechn. Ber. 47, ( 1 1 ), 251 -270 ( 1974).

8. Carling, J.: Glastechn. Ber. 49, ( 12), 269-277 ( 1976).

9. Patankar, S.V.: Numerical heat transfer and fluid flow, Hemisphere

Publishing Co. McGraw-Hill 1980.

10. Ungan, A., Viskanta, R.: Glastechn. Ber. 60, (4), 1 15- 124 ( 1987).

1 1. Murname, R.A., Moreiand, N.J.: Ceram. Eng. Sci. Proc. 9 (3-4), 192-202 ( 1988).

12. Simonis F. De Waal H. and Beerkens R.C.G., XIV Int. Congr. on Glass -

Collected Papers Part III ( 1986), p. 1 18- 127.

13. Ungan, A. Viskanta, R.: Glastech. Ber. 60, 71 ( 1987).

14. Ungan,A., Viskanta, R.: Glass Technol. 28 (6), 252-260 ( 1987).

15. Choudhary, M. .: J. Non-Cryst. Solids 101, 41 -53 ( 1988)

16. Viskanta, R.; J. Non-Cryst. Solids 177, 347-362 ( 1994) .

17. Goldberg, A.J.: Glastechn. Ber. 46 (4), 67-70 ( 1973).

18. Simonis F., Proceedings of the 2nd Int. Conf. "Advances in the Fusion and Processing of Glass", Glastech. Ber. 63K, 1 (1990).

19. Ungan, A., Turner, W.H., Viskanta, R.: Glastech. Ber. 56K, 125 ( 1983).

20. Nemec L., Miihlbauer M.: Glastechn. Ber. 56K, 82 (1983).

21. Beerkens R., Muijsenberg H., van den Heiden T.: Glastechn. Ber. Glass Sci. Technol., 67 (1), 179 ( 1994).

22. Ungan, A: Glastechn. Ber. 63K, 19-28 (1990).

23. Matyas, J., Nemec, L.: Glass Sci. Teechnol. 76 (2), 71-80 (2003).

24. Oda, K, aminoyama M.: J. Cer. Soc. Japan 117 (6), 736-741(2009).

25. asa, S., Lisy, A., Novotny, F.: PV 201 1 - 651 (201 1 ).

26. Cooper, A.R.: Collected Papers of the XI V International Congress on Glass ( 1986), pp.1-8.

27. Consortium on development of innovative glass melting BRIG 2020: The International Conference of Glass Experts, Brig, March 26-29

(Switzerland).

28. Nemec, L., Jebava, M.: Eur. J. Glass Sci. Technol. A, 47 (3), 68-77 (2006).

29. Nemec, L., Jebava, M., Cincibusova, P.: Ceramics-Silikaty, 50 (3), 140- 152 (2006).

30. Nemec, L., Cincibusova, P.: Ceramics-Silikaty, 52 (4), 240-249 (2008).

31. Nemec, L., Cincibusova, P.: Ceramics-Silikaty, 53 (3), 145-155 (2009).

32. Polak, M., Nemec, L.: Ceramics-Silikaty, 54 (3), 212-218 (2010).

33. Polak, M., Nemec, L.: Journal of Non-Crystalline Solids 357 ( 16-17), 3 108 - 31 16 (201 1).

34. Polak, M., Nemec, L.: Journal of Non-Crystalline Solids 358 (2012).

35. Cincibusova, P., Nemec, L.: Accepted in: Eur. J. Glass Sci. Technol. A, 53 (2012).

36. Polak, M.: Ph.D. Thesis. ICT Prague (2012).