Method implemented for the regulation of a forming machine for molten glass drops

Abstract

The present invention concerns a computer-implemented method (10) for the regulation of a forming machine for molten glass drops of the type comprising: - containment means configured to contain molten glass; - a punch controllable in terms of initial height and in terms of stroke; - cutting means configured to cut the continuous flow of the molten glass in such a way as to form a drop at the end of a punching step. It should be noted that such method (10) comprises at least the following steps in sequence: - a characterization step a1) which provides for selecting a first control variable (u) of the punch chosen between the initial height and the stroke; - a control step b) which provides for calculating, for at least one punching step (Pk) and for the subsequent punching step (Pk+1), respectively a control variation (Δuk) and a subsequent control variation (Δuk+1) of the aforesaid first control variable (u) in such a way as to obtain, with such punching step (Pk), a drop with a desired weight variation (ΔWk) with respect to what is obtainable without the application of the control variation (Δuk) and the subsequent control variation (Δuk+1); more precisely, the control variation (Δuk) and the subsequent control variation (Δuk+1) satisfy the equation with a first coefficient (x1) and a second coefficient (x2) predetermined: ΔW k = x 1Δu k + x 2Δu k+1

Description

[0001] METHOD IMPLEMENTED FOR THE REGULATION OF A FORMING MACHINE FOR MOLTEN GLASS DROPS.

[0002] DESCRIPTION

[0003] The invention concerns a method for the regulation of a forming machine for molten glass drops.

[0004] Today, computer-implemented methods are known for regulating forming machines for molten glass drops, where such regulation allows executing a succession of steps for punching drops with desired weights.

[0005] These methods, although known and appreciated, however, have an important limit.

[0006] In particular, the main limit is linked to the low precision of these regulation processes which therefore involve punching drops with actual weights that deviate also considerably from the corresponding desired weights.

[0007] The task of the present invention is to develop a computer-implemented method for the regulation of a forming machine for molten glass drops capable of overcoming the aforesaid drawback and limit of the prior art.

[0008] In particular, it is an object of the present invention to realize a method for the regulation of a forming machine that allows to obtain molten glass drops with higher precision than similar methods of known type.

[0009] The above task and object are achieved by a method according to claim 1 .

[0010] Further characteristics of the method according to claim 1 are described in the dependent claims.

[0011] The aforesaid task and object, together with the advantages that will be mentioned hereinafter, are highlighted by the description of an embodiment of the invention, which is given by way of non-limiting example with reference to the attached drawings, where:

[0012] - Figure 1 represents the flow chart of the method according to the invention;

[0013] - Figure 2 represents an example of dynamics of the method according to the invention.

[0014] A first embodiment of a computer-implemented method for the regulation of a forming machine for molten glass drops according to the invention is indicated as a whole with the number 10 and is clearly visible in Figure 1 .

[0015] It should be highlighted that such forming machine is of the type comprising:

[0016] - containment means configured to contain molten glass and having an orifice for the extraction of a continuous flow of the molten glass during a punching step of a drop of said molten glass;

[0017] - a punch configured to translate in height in relation to the orifice in such a way as to regulate the aforesaid continuous flow during said punching step; more precisely, the punch is controllable in terms of initial height in relation to the orifice at the beginning of said punching step and is also controllable in terms of stroke defined as the difference between the maximum height reached by the punch in relation to the orifice during the punching step and the aforesaid initial height as represented in Figure 2;

[0018] - cutting means configured to cut the continuous flow in such a way as to form the drop at the end of the punching step.

[0019] The aforesaid method 10 according to such first embodiment of the invention comprises in sequence the following characterization steps: ai ) selecting a first control variable u of the punch chosen between the aforesaid initial height and the aforesaid stroke; in this first embodiment of the invention, the initial height h of the punch was chosen as the first control variable u, but it is not excluded that the stroke c can be chosen as the first control variable u in a different embodiment of the invention; a2) executing a first punching step Pi by controlling the punch with a first predefined initial height hi and a first predefined stroke ci, where such first punching step Pi allows to obtain a drop of a first weight Wi; as) executing a second punching step P2 subsequent to the first punching step Pi by controlling the punch with a second predefined initial height h2 and a second predefined stroke C2, where such second punching step P2 allows to obtain a drop of a second weight W2;

[0020] 34) executing a further first punching step Pi-1 by controlling the punch as in the first punching step Pi, where such further first punching step Pi-1 allows to obtain a drop of a further first weight WM; as) executing a further second punching step Pi subsequent to the further first punching step Pi-1 by controlling the punch with a first perturbation Aui at the first control variable u and keeping unchanged the further control variable v with respect to the second punching step P2, where such further second punching step Pi allows to obtain a drop of a further second weight Wi; in this first embodiment of the invention, the first perturbation Aui corresponds to a first initial height perturbation Ahi; ae) determining a first weight variation AW equal to the difference between the further first weight WM relative to the further first punching step Ps-i and the first weight Wi relative to the first punching step Pi; a?) determining a second weight variation AWi equal to the difference between the further second weight Wi relative to the further second punching step Pi and the second weight W2 relative to the second punching step P2; as) calculating a first coefficient xi equal to:

[0021] In this first embodiment of the invention, the first coefficient xi is a first initial height coefficient a equal to: ag) calculating a second coefficient X2 equal to:

[0022] In this first embodiment of the invention, the second coefficient X2 is a second initial height coefficient p equal to: b)

[0023] This control step b) provides for calculating for at least one punching step Pk and for the subsequent punching step Pk+1 respectively a control variation Auk and a subsequent control variation Auk+1 of the first control variable u in such a way as to obtain with such further punching step Pk a drop with a desired weight variation AWk with respect to what is obtainable without the application of said control variation Auk and said subsequent control variation Auk+1, where such first control variation Auk and such second control variation Auk+1 are required to satisfy the equation:

[0024] AVI / fc = Xj^Aiifc + x2Auk+1

[0025] In this first embodiment of the invention, the control variation Auk and the subsequent control variation Auk+i are respectively an initial height variation Ahk and a subsequent initial height variation Ahk+i that are required to satisfy the equation: Wk= a hk+ ?A / ik+1

[0026] It is emphasized for clarity that the index i is a natural number used to identify the i-th punching step implemented by the forming machine for molten glass drops.

[0027] Even more precisely, the present embodiment example of the invention provides that the forming machine executes a succession of N punching steps Pn adapted to be executed successively and in a cyclical manner, where 0 < n <= N, N being a positive integer, and where according to the present embodiment example such succession of N steps is equal to five, as can be seen from Figure 2.

[0028] Therefore, again with reference to the present example, the selected index i - 1 and index i are equal to N+1 and N+2, respectively.

[0029] However, it is not excluded that the index i - 1 as well as the index i are different from what has just been presented as long as they are associated with two subsequent forming steps Ps-i and Pi as previously described.

[0030] It should also be highlighted that the calculation of the first coefficient xi and of the second coefficient X2, as well as the imposition of the equation defined in step b) for the regulation of the aforesaid forming machine, advantageously allow to obtain molten glass drops with higher precision than similar methods of known type.

[0031] The method 10 according to this first embodiment of the invention specifically provides that, given a succession of punching steps Pn adapted to be executed successively in a cyclical manner, the control step b) comprises the definition of a linear system according to which: where each desired weight variation AWI....AWN is preset. Furthermore, such control step b), according to such first embodiment of the invention, comprises the following sub-steps in sequence: bi) calculating for an equation of such a linear system relative to a predetermined punching step Ps, where 0 < s <= N, the corresponding control variation Ausignoring the corresponding subsequent control variation Aus+i, in such a way as to obtain a drop with the corresponding desired weight variation AWs, where such control variation Aussatisfies the equation:

[0032] In this first embodiment of the invention, the predetermined punching step Psis the last punching step PN of the aforesaid succession of punching steps Pn, where therefore the corresponding control variation Ausis the corresponding last initial height control variation Ahu and the corresponding subsequent control variation Aus+i ignored in this step is the first initial height control variation Ahi relative to the first punching step Pi of such succession of punching steps Pn; even more precisely, the corresponding last initial height control variation AhN satisfies the equation: b2) starting from the control variation Auscalculated for the punching step Ps, solving the aforesaid linear system, calculating iteratively for each of the remaining equations of the linear system relative to the remaining punching steps Pm the corresponding control variation Aum, where 0 < m <= N.

[0033] In particular, the resolution of the linear system according to step b2) provides for initializing m equal to s-1 and executing the following two operations cyclically, as long as m is equal to s:

[0034] - calculating the corresponding control variation Aumin such a way as to obtain a drop with the corresponding desired weight variation AWm by means of the equation:

[0035] - setting m equal to m-1 , and when m is equal to 0, setting m equal to N.

[0036] In this first embodiment of the invention, for each interaction of such cycle, the height variation Ahm is calculated in such a way as to obtain a drop with the corresponding desired weight variation AWm by means of the equation:

[0037] A / im= (Aim- Mim+1) / a bs) having determined the control variation Aus+i relative to the punching step Ps+i calculated in step b2), calculating a first weight error esequal to: es= |X2AUS+1|

[0038] It should be specified that such weight error esis due to the fact that Aus+i was initially ignored for the calculation of the control variation Ausduring step b1 ).

[0039] In this first embodiment of the invention, the first weight error esis linked to the first punching step Pi given the cyclicity of the succession of punching steps Pn and is equal to: b4) when the first weight error esis greater than or equal to a tolerance error etoi, starting a new first processing cycle by calculating again for the predetermined punching step Psthe corresponding control variation Ausconsidering the corresponding subsequent control variation Aus+i calculated in step b2) of the previous processing cycle satisfying the equation:

[0040] AI / I4 = X-LAUS + x2kus+1and repeating then steps b2) to b4).

[0041] In this first embodiment of the invention, the corresponding first initial height control variation AhN is therefore recalculated for the last punching step PN considering the corresponding second control variation Ahi calculated in step b2) of the previous processing cycle satisfying the equation:

[0042] AI / IZ / v = akhN+ / 7A / 1-L bs) ending the current first processing cycle by making available, in particular to a forming machine, the succession of control variations Aunof the succession of punching steps Pn previously calculated; in this first embodiment of the invention, a succession of initial height control variations Ahn of the succession of punching steps Pn is determined.

[0043] The method 10 according to the present embodiment of the invention provides that the forming machine is of the type comprising movement means configured for the movement of said molten glass inside said containment means and controllable in terms of height in relation to the bottom of said containment means.

[0044] In particular, the method 10 still provides for the following control steps in sequence: ci) starting a processing cycle by executing at least once said succession of punching steps Pn using the succession of control variations Aun; in this first embodiment of the invention, the succession of punching steps Pn is executed once using the succession of initial height control variations Ahn;

[0045] C2) determining the average weight Waof the drops obtained with step ci);

[0046] C3) when the average weight Wa is higher than a predefined weight Woof a deviation ew, decreasing the height of the movement means with respect to the previous processing cycle, and then repeating steps ci) to C3), while, when the average weight Wa is less than the predefined weight Woof the deviation ew, increasing the height of the movement means with respect to the previous processing cycle, and then repeating steps ci) to C3).

[0047] Steps ci) to C3) advantageously allow the average of the weights to be shifted without having to modify the regulation parameters determined previously.

[0048] The method 10 according to the present embodiment of the invention comprises the following control steps in sequence: di) executing the succession of punching steps Pn once using the succession of control variations Aun; in this first embodiment of the invention, the succession of punching steps Pn is executed using the succession of initial height control variations Ahn; d2) calculating a succession of weight errors en, each of which is equal to the absolute value of the difference between the corresponding actual weight of the drop obtained with the corresponding punching step and the corresponding desired weight Wn; ds) when the biggest of said succession of weight errors enis greater than or equal to a maximum weight error emax, repeating steps from ai) to ag) calculating a new first coefficient xi* and a new second coefficient X2* starting from a different first perturbation Aui* and using in the remaining steps the arithmetic mean between the previous coefficient xi, X2 and the corresponding new coefficient xi*, X2*; in this first embodiment of the invention, the repetition of steps from ai) to ag) provides for the calculation of a new first initial height coefficient a* and of a new second initial height coefficient p*; in addition, the arithmetic mean between the previous coefficient a, p and the corresponding new coefficient a*, p* is used in the remaining steps according to the following formulas:

[0049] In a method 100 relative to a second embodiment, the invention provides what has been described so far and in addition the following characterization steps in sequence, following the characterization step ag): aio) executing a third punching step P3 by controlling the punch with a third predefined initial height hs and a third predefined stroke C3, where such third punching step P3 allows to obtain a drop of a third weight W3; an) executing a fourth punching step P4 subsequent to the third punching step P3 by controlling the punch with a fourth predefined initial height hu and a fourth predefined stroke C4, where such fourth punching step P4 allows to obtain a drop of a fourth weight W4;

[0050] 312) executing a further third punching step Pj-1 by controlling the punch as in said third punching step P3, wherein such further third punching step Pj-1 allows to obtain a drop of a further third weight Wj-1; aw) executing a further fourth punching step Pj subsequent to said further fourth punching step Pj-1 by controlling the punch with a second perturbation Avj at said further control variable v and keeping unchanged the first control variable u with respect to the fourth punching step P4, where such further fourth punching step Pj allows to obtain a drop of a further fourth weight Wj; in this second embodiment of the invention, the second perturbation Avj corresponds to a second stroke perturbation A ; ai4) determining a third weight variation AWj-1 equal to the difference between the further third weight Wj-1 relative to the further third punching step Pj-1 and the third weight W3 relative to the third punching step P3; aw) determining a fourth weight variation AW equal to the difference between the further fourth weight Wj relative to the further fourth punching step Pj and the fourth weight W4 relative to the fourth punching step P4; aw) calculating a third coefficient yi equal to: In this second embodiment of the invention, the third coefficient yi is a third stroke coefficient 6 equal to: ai?) calculating a fourth coefficient y2 equal to:

[0051] In this second embodiment of the invention, the fourth coefficient y2 is a fourth stroke coefficient 0 equal to:

[0052] The control step b), in accordance with such second embodiment, further comprises the calculation, for at least the punching step Pk and the subsequent punching step Pk+i, respectively of a further control variation Avk and a subsequent further control variation Avk+i of the further control variable v in such a way as to obtain with said punching step Pk a drop with a desired weight variation AWk with respect to what is obtainable without the application of said control variation Auk, of said subsequent control variation Auk+i, of said further control variation Avk and of said subsequent further control variation Avk+i; more precisely, the control variation Auk, the subsequent control variation Auk+i, the further control variation Avk and the subsequent further control variation Avk+i are required to satisfy the equation:

[0053] Aiyk= x1 / \uk+ %2^ / c+i + yi^vk+ y2&vk+1

[0054] In particular, in this second embodiment of the invention, the further control variation Avk and the subsequent further control variation Avk+i are respectively a stroke variation Ack and a subsequent stroke variation Ack+i that are required to satisfy the equation:

[0055] &Wk= ahhk+ / ?A / ik+1+ <5Ack+ 0Ack+1

[0056] Even more in detail, according to this second embodiment of the invention, given the aforesaid succession of punching steps Pn adapted to be executed successively in a cyclical manner, the control step b) provides for defining a linear system according to which: where each desired weight variation LW1... &WNis preset.

[0057] In this second embodiment of the method 100, step bi) further provides for calculating for an equation of the linear system relative to a predetermined punching step Ps, where 0 < s <= N, as well as the control variation Aus(in particular the height variation Ahs), the corresponding further control variation Avs (in particular the stroke variation Acs), ignoring the corresponding subsequent further control variation Avs+i, as well as the subsequent control variation Aus+i, in such a way as to obtain a drop with the corresponding desired weight variation AWs. This control variation Ausand the further control variation Avs having therefore to satisfy the equation:

[0058] AI / I4 = X-LAUS + y^Vg

[0059] In this second embodiment of the invention, the predetermined punching step Ps is always the last punching step PN of said succession of punching steps Pn, where therefore the corresponding further control variation Avsis the corresponding stroke control variation ACN and the corresponding subsequent further control variation Avs+i ignored in this step is the subsequent stroke control variation Aci relative to the first punching step Pi of such succession of punching steps Pn; even more precisely, the corresponding last stroke control variation ACN satisfies the equation:

[0060] AI / IZ / v = a / \hN+ <5Acw

[0061] Step b2), in accordance with this second embodiment of the invention, starting also from said further control variation Avscalculated for the punching step Ps, provides for solving the aforesaid linear system, iteratively calculating for each of the remaining equations of said linear system relative to the remaining punching steps Pm also the corresponding further control variation Avm, where 0 < m <= N.

[0062] In particular, such resolution of the linear system provides for initializing m equal to s-1 and executing the following two operations cyclically, as long as m is equal to s:

[0063] - calculating also the corresponding further control variation Avmin such a way as to obtain a drop with the corresponding desired weight variation AWm, where such corresponding control variation Aum, such corresponding subsequent control variation Aum+i, such corresponding further control variation Avmand such corresponding subsequent further control variation Avm+i satisfy the equation:

[0064] - setting m equal to m-1 , and when m is equal to 0, setting m equal to N.

[0065] In this second embodiment of the invention, therefore, the stroke control variations Acmand Acm+i are also considered that are required to satisfy the following equation:

[0066] AWm= aAhm+ / ?A / im+1+ <5Acm+ 0Acm+1

[0067] Given also the subsequent further control variation Avm=s+i relative to the punching step Pm=s+i calculated in step b2), according to step bs) of the method, the first weight error esis equal to: es= lx2Aus+1+ y2Avs+1|

[0068] In this second embodiment of the invention, the first weight error esis always linked to the first punching step Pi given the cyclicity of the succession of punching steps Pn and is equal to: eN= I / 7A / 1-L + 0Acx|

[0069] Step b4), when the first weight error esis greater than or equal to the aforesaid tolerance error etoi, provides for recalculating for the predetermined punching step Psalso the corresponding further control variation Avsconsidering the corresponding subsequent further control variation Avs+i calculated in step b2) of the previous processing cycle, satisfying the equation:

[0070] AIVS= xrkus+ x2Aus+1+ yiAvs+ y2Avs+1 and furthermore such step b4) provides then for repeating steps b2) to b4).

[0071] In this second embodiment of the invention, the corresponding stroke control variation ACN is therefore recalculated for the last punching step PN considering the corresponding subsequent control variation Aci calculated in step b2) of the previous processing cycle, satisfying the equation:

[0072] AI / IZ / v = akhN+ / 7A / 1-L + <5Acw+ 9Ac1

[0073] Step bs) provides for ending the current first processing cycle by making available also the succession of further control variations Avnof the succession of punching steps Pn previously calculated; in this second embodiment of the invention, in particular, a succession of stroke control variations Acnof the succession of punching steps Pn is also determined.

[0074] Step ci), according to such second embodiment of the invention, comprises the execution, at least once, of the succession of punching steps Pn using the succession of control variations Aunand the succession of further control variations Avn.

[0075] In particular, in this second embodiment of the invention, the succession of punching steps Pn is executed once using the succession of initial height control variations Ahn and the succession of stroke control variations Acn.

[0076] Step di), according to such second embodiment of the invention, comprises the execution, once, of the succession of punching steps Pn using the succession of control variations Aunand the succession of further control variations Avn.

[0077] In particular, in this second embodiment of the invention, the succession of punching steps Pn is executed once using the succession of initial height control variations Ahn and the succession of stroke control variations Acn.

[0078] Step ds) then comprises repeating steps aw) to ais) by calculating a new third coefficient yi* and a new fourth coefficient y2* starting from a different second perturbation Avi* and using in the remaining steps the arithmetic mean between the previous coefficient yi, y2 and the corresponding new coefficient yi*, y2*.

[0079] In this second embodiment of the invention, the repetition of steps from aw) to ais) provides for the calculation of a new third stroke coefficient 6* and of a new fourth stroke coefficient 0*; furthermore, the arithmetic mean between the previous coefficient and the corresponding new coefficient is used in the remaining steps according to the following formulas: In this second embodiment of the method 100, given a predetermined priority coefficient A, for each of the successions of punching steps Pk, the desired weight variation AWk comprises a first component AW’ linked to the first control variable u and a second component AW” linked to the further control variable v according to the following equations: W" = (1 - A)AlVk

[0080] In this second embodiment of the invention, the first component AWUis linked to the initial height h and the second component AWVis linked to the stroke c: Wh= AAWk

[0081] AWC= (1 - A)AlVk

[0082] In this embodiment of the invention, a priority coefficient A equal to a fixed value of 0.9 is selected.

[0083] It is not excluded, however, that such coefficient is different from such fixed value.

[0084] It is not excluded, yet, that such priority coefficient A is equal to 1 if |<5| < |0|, while it is equal to the aforesaid fixed value, otherwise.

[0085] In both embodiments of the invention, step b4) is executed for a predefined maximum number of iterations.

[0086] It is not excluded, however, that such step b4) is executed for an indefinite number of iterations as long as the aforesaid equation relative to the value of AWs is not satisfied.

[0087] It should also be highlighted that, in both embodiment examples of the invention, whenever it is envisaged recalculating the new first initial height coefficient a* and the new second initial height coefficient p*, the step corresponding to the biggest error of the succession of errors endetermined during the previous step d2) is used as the new predetermined punching step Psof step bi).

[0088] Practically, it has been established that the invention achieves the intended task and objects.

[0089] In particular, with the invention, a regulation method has been developed that allows to obtain molten glass drops with higher precision than similar methods of known type.

Claims

CLAIMS1 ) Computer-implemented method (10) for the regulation of a forming machine for molten glass drops of the type comprising:- containment means configured to contain molten glass and having an orifice for the extraction of a continuous flow of said molten glass during a punching step of a drop of said molten glass;- a punch configured to translate in height in relation to said orifice in such a way as to regulate said continuous flow during said punching step, said punch being controllable in terms of initial height in relation to said orifice at the beginning of said punching step, said punch being controllable in terms of stroke defined as the difference between the maximum height reached by said punch in relation to said orifice during said punching step and said initial height;- cutting means configured to cut said continuous flow in such a way as to form said drop at the end of said punching step; where said method (10) comprises the following steps in sequence:- the characterization step ai) which provides for selecting a first control variable (u) of said punch chosen between said initial height and said stroke;- the control step b) which provides for calculating, for at least one punching step (Pk) and for the subsequent punching step (Pk+i), respectively a control variation (Auk) and a subsequent control variation (Auk+i) of said first control variable (u) in such a way as to obtain, with said punching step (Pk), a drop with a desired weight variation (AWk) with respect to what is obtainable without the application of said control variation (Auk) and said subsequent control variation (Auk+i), said control variation (Auk) and said subsequent control variation (Auk+i) by satisfying the equation with a first coefficient (xi) and a second coefficient (X2) predetermined:characterized in that it comprises the following characterization steps in sequence between said characterization step ai) and said control step b): a2) executing a first punching step (Pi) by controlling said punch with a first predefined initial height (hi) and a first predefined stroke (ci), said first punching step (Pi) allowing to obtain a drop of a first weight (Wi); as) executing a second punching step (P2) subsequent to said first punchingstep (Pi) by controlling said punch with a second predefined initial height (h2) and a second predefined stroke (C2), said second punching step (P2) allowing to obtain a drop of a second weight (W2);34) executing a further first punching step (Pi-1 ) by controlling said punch as in said first punching step (Pi ), said further first punching step (P ) allowing to obtain a drop of a further first weight (WM ); as) executing a further second punching step (Pi) subsequent to said further first punching step (P ) by controlling said punch with a first perturbation (Aui) at said first control variable (u) and keeping unchanged the further control variable (v) with respect to said second punching step (P2), said further second punching step (Pi) allowing to obtain a drop of a further second weight (Wi); as) determining a first weight variation (AWM ) equal to the difference between said further first weight (WM ) relative to said further first punching step (PM ) and said first weight (W1) relative to said first punching step (Pi); a?) determining a second weight variation (AWi) equal to the difference between said further second weight (Wi) relative to said further second punching step (Pi) and said second weight (W2) relative to said second punching step (P2); as) calculating said first coefficient (xi) equal to:ag) calculating said second coefficient (X2) equal to:AtVi-i X2 = -Alii2) Method (10) according to claim 1 , characterized in that, given a succession of punching steps (Pn) adapted to be executed successively in a cyclical manner, where 0 < n <= N, N being a positive integer, said control step b) provides for defining a linear system according to which:wherein each desired weight variation AW-L... AI / Vnis preset, said step b) further comprising the following control sub-steps in sequence: bi) calculating for an equation of said linear system relative to a predetermined punching step (Ps), where 0 < s <= N, the corresponding control variation (Aus) ignoring the corresponding subsequent control variation (Aus+i), in such a way as to obtain a drop with the corresponding desired weight variation (AWs), said corresponding control variation (Aus) thus satisfying the equation:AI / I4 = X-LAUS b2) starting from said control variation (Aus) calculated for the punching step (Ps), solving the aforesaid linear system, by iteratively calculating for each of the remaining equations of said linear system relative to the remaining punching steps (Pm) the corresponding control variation (Aum), where 0 < m <= N, said resolution of said linear system providing to initialize m equal to s-1 and to execute cyclically the following two operations, as long as m is equal to s:- calculating the corresponding control variation (Aum) in such a way as to obtain a drop with the corresponding desired weight variation (AWm)- setting m equal to m-1 , and when m is equal to 0, setting m equal to N; bs) having calculated the control variation (Aum=s+i) relative to the punching step (Pm=s+i) calculated in step b2), calculating a first weight error (es) equal to: es= |X2AUS+1| b4) when said first weight error (es) is greater than or equal to a tolerance error (etoi), starting a new first processing cycle by calculating again for said predetermined punching step (Ps) the corresponding control variation (Aus) considering the corresponding subsequent control variation (Aus+i) calculated in step b2) of the previous processing cycle, satisfying the equation:AI / 14 = xj±us+ x2&us+rand repeating then steps b2) to b4); bs) ending the current first processing cycle by making available the succession of control variations (Aun) of said succession of punching steps (Pn) previously calculated.3) Method (10) according to claim 2, characterized in that said forming machine is of the type comprising movement means configured for the movement of said molten glass inside said containment means and controllable in terms of height in relation to the bottom of said containment means, said method (10) comprising the following control steps in sequence: ci) starting a processing cycle by executing at least once said succession of punching steps (Pn) using said succession of control variations (Aun);C2) determining the average weight (Wa) of the drops obtained with said step ci);C3) when said average weight (Wa) is higher than a predefined weight (Wo) of a deviation (ew), decreasing said height of said movement means with respect to the previous processing cycle, and then repeating steps ci) to cs), while, when said average weight (Wa) is less than said predefined weight (Wo) of said deviation (ew), increasing said height of said movement means with respect to the previous processing cycle, and then repeating steps ci) to C3).4) Method (10) according to claim 2 or 3, characterized in that it comprises the following control steps in sequence: di) executing once said succession of punching steps (Pn) using said succession of control variations (Aun); d2) calculating a succession of weight errors (en) each of which is equal to the absolute value of the difference between the corresponding actual weight of the drop obtained with the corresponding punching step and the corresponding desired weight (Wn); ds) when the biggest error of said succession of weight errors (en) is greater than or equal to a maximum weight error (emax), repeating steps ai) to ag) calculating a new first coefficient (xi*) and a new second coefficient (X2*) starting from a different first perturbation (Aui*) and using in the remaining steps the arithmetic mean between the previous coefficient (xi, X2) and the corresponding new coefficient (xi*, X2*).5) Method (10) according to any one of the preceding claims,characterized in that it comprises in sequence the further characterization steps, following said characterization step ag): aio) executing a third punching step (P3) by controlling said punch with a third predefined initial height (hs) and a third predefined stroke (C3), said third punching step (P3) allowing to obtain a drop of a third weight (W3); an) executing a fourth punching step (P4) subsequent to said third punching step (P3) by controlling said punch with a fourth predefined initial height (h4) and a fourth predefined stroke (C4), said fourth punching step (P4) allowing to obtain a drop of a fourth weight (W4);312) executing a further third punching step (Pj-1 ) by controlling said punch as in said third punching step (P3), said further third punching step (Pj-1) allowing to obtain a drop of a further third weight (Wj-1); aw) executing a further fourth punching step (Pj) subsequent to said further fourth punching step (Pj-1) by controlling said punch with a second perturbation (Avj) at said further control variable (v) and keeping unchanged said first control variable (u) with respect to said fourth punching step (P4), said further fourth punching step (Pj) allowing to obtain a drop of a further fourth weight (Wj); ai4) determining a third weight variation (AWj-1) equal to the difference between said further third weight (Wj-1) relative to said further third punching step (Pj-1 ) and said third weight (W3) relative to said third punching step (P3); aw) determining a fourth weight variation (AWj) equal to the difference between said further fourth weight (Wj) relative to said further fourth punching step (Pj) and said fourth weight (W4) relative to said fourth punching step (P4); aw) calculating a third coefficient (yi) equal to:ai?) calculating a fourth coefficient (y2) equal to:said step b) comprising the calculation, for at least said punching step (Pk) and said subsequent punching step (Pk+1 ), respectively of a further control variation (Avk) and a subsequent further control variation (Avk+1) of said further controlvariable (v) in such a way as to obtain with said punching step (Pk) a drop with a desired weight variation (A Wk) with respect to what is obtainable without the application of said control variation (Auk), of said subsequent control variation (Auk+i), of said further control variation (Avk) and of said subsequent further control variation (Avk+i), where said control variation (Auk), said subsequent control variation (Auk+i), said further control variation (Avk) and said subsequent further control variation (Avk+i) satisfy the equation:AlVk= x1 / \uk+ x2Auk+1+ yi^Vk + y2^vk+16) Method (10) according to claim 5 in combination with claim 2, characterized in that:- given said succession of punching steps (Pn) adapted to be executed successively in a cyclical manner, said control step b) provides for defining a linear system according to which:where each desired weight variation LW1... &WNis preset;- said step bi) also provides for calculating for said one equation of said linear system relative to said one predetermined punching step (Ps), where 0 < s <= N, the corresponding further control variation (Avs) ignoring the corresponding subsequent further control variation (Avs+i) in such a way as to obtain a drop with the corresponding desired weight variation (AWs), said corresponding further control variation (Avs) thus satisfying the equation:AI / I4 = XjAits + y^Vg- said step b2), starting also from said further control variation (Avs) calculated for the punching step (Ps), provides for solving the aforesaid linear system, iteratively calculating for each of the remaining equations of said linear system relative to the remaining punching steps (Pm) also the corresponding further control variation (Avm), where 0 < m <= N, said resolution of said linear system providing for initializing m equal to s-1 and cyclically executing the following two operations, as long as m is equal to s:- calculating also the corresponding further control variation (Avm) in such a way as to obtain a drop with the corresponding desired weight variation (AWm), said corresponding control variation (Aum), said corresponding subsequent control variation (Aum+i), said corresponding further control variation (Avm) and said corresponding subsequent further control variation (Avm+i) satisfying the equation:- setting m equal to m-1 , and when m is equal to 0, setting m equal to N;- said step bs), having also determined the further control variation (Avm=s+i) relative to the punching step (Pm=s+i) calculated in step b2), provides for calculating said first weight error (es) equal to: es= |x2Aus+1+ y2Avs+1|- said step b4), when said first weight error (es) is greater than or equal to said tolerance error (etoi), provides for recalculating for said predetermined punching step (Ps) also said further control variation (Avs) considering the corresponding subsequent further control variation (Avs+i) calculated in step b2) of the previous processing cycle, satisfying the equation:AIVS= x Aus+ x2Aus+1+ y1Avs+ y2Avs+1and repeating then steps b2) to b4);- said step bs) provides for ending the current first processing cycle by also making available the succession of further control variations (Avn) of said succession of punching steps (Pn) previously calculated.7) Method (10) according to claim 6 in combination with claim 3, characterized in that said control step ci) comprises executing, at least once, said succession of punching steps (Pn) using said succession of control variations (Aun) and said succession of further control variations (Avn).8) Method (10) according to claim 6 or 7 in combination with claim 4, characterized in that said control step d1 ) comprises the execution, once, of said succession of punching steps (Pn) using said succession of control variations (Aun) and said succession of further control variations (Avn), said control step ds) comprising the repetition of the characterization steps from aio) to ais) calculating a new third coefficient (yi*) and a new fourth coefficient (y2*)starting from a different second perturbation (Avi*) and using in the remaining steps the arithmetic mean between the previous coefficient (yi, y2) and the corresponding new coefficient (yr, y2*).9) Method (10) according to any one of claims 5 to 8, characterized in that, given a predetermined priority coefficient (A), for each of said successions of punching steps (Pk) said desired weight variation (AWk) comprises a first component (AW’) linked to said first control variable (u) and a second component (AW”) linked to said further control variable (v) according to the following equations:AW' = AAWkAW" = (1 - A)AWk

Images