NL2028215B1 - Method of inspecting hollow glass products of glass product material - Google Patents

Abstract

Method and system for inspecting hollow glass products of glass product material, wherein said glass products are manufactured by: a. heating the glass product material; b. forming the heated glass product material into at least one glass product; c. cooling the formed glass product; wherein inspecting the glass products comprises the following steps: d. successively conveying along a predetermined path the in step b. formed glass products past at least one infrared light-sensitive sensor, wherein an image is made with the at least one sensor per glass product with the at least one sensor of a plurality of the glass products which are successively transported past the at least one sensor, wherein step d. between step b. and c. is carried out; e. processing the in step d. images made to obtain information about a wall thickness of the glass products.

Description

P128760EN00 Title: Method for inspecting hollow glass products from lettuce product material The invention relates to a method for inspecting hollow glass products from glass product material, wherein said glass products are manufactured by: a. heating the glass product material; b. forming the heated glass product material into at least one glass product in a production stream; c. cooling the formed glass product.

The invention further relates to a method for producing and inspecting hollow glass products from lettuce product material, wherein said glass products are manufactured at least by: a. heating the glass product material; b. forming the heated lettuce product material into a glass product in a production stream; c. cooling the formed glass product.

The invention also relates to a system for producing and inspecting glass products from glass product material according to the above-mentioned method for inspecting hollow glass products, the system comprising: a heating device for performing step a.; a product molding device such as a mold for performing step b.; a cooling device for performing step c.; and a conveying device for conveying glass products formed with the product molding device from the product molding device to the cooling device; at least one sensor for scanning the In step b. molded glass products; and a signal processing unit connected to the at least one sensor for processing signals from the at least one sensor.

In addition, the invention relates to a system for inspecting glass products of glass product material according to the above method for inspecting the glass products, the system comprising: at least one sensor for scanning the glass products; and a signal processing unit connected to the at least one sensor for processing signals from the at least one sensor.

Such methods and systems are known per se, for example from WO-2019133504A1. In the known method, images are made of the glass products which have just been manufactured while still warm with the aid of a plurality of sensors. Because the sensors are set up around the glass product, the images cover a complete revolution of the product. The creation of such a group of images is performed repeatedly at different times. The intensity of the infrared radiation is visible on each of the images. By comparing two images of the same part of the product made at different times, a decrease in intensity can be determined. If the intensity decreases relatively slowly, it is determined that the glass material is relatively thick at that location. If the intensity decreases relatively quickly, it is determined that the glass material is relatively thin at that location. In this way a lateral glass thickness distribution of the glass product can be determined. A disadvantage is that this method is relatively inaccurate. For example, the intensity is also a function of the temperature. This must be estimated in order to be able to determine a glass thickness distribution from the intensity.

Furthermore, arranging the sensors around the product is expensive and cumbersome. The said Lateral Glass Thickness Distribution (here: LGV) (English: Lateral Glass Distribution LGD) at a certain height is the collection of the wall thicknesses around the circumference of the product (figure 5). This LGV can be obtained, for example, by measuring glass wall thicknesses at a certain height h and carrying out thickness measurements around the entire circumference of the product. All wall thicknesses of the entire product is the Total Lateral Glass Thickness Distribution of the product. A single wall thickness of the product at a certain height h and angle phi (polar coordinates) is an element of the collection of wall thicknesses: Lateral Glass Thickness Distribution LGV(h,phi). The number of elements in the LGV depends on the freely selectable measuring resolution of the height and the angle.

The LGV is a very important parameter for the quality of the glass product. The strength of the product is mainly determined by the thinnest part of a glass wall. To prevent breakage during normal use of the product, the LGV must meet a manufacturer's specifications. However, with current glass manufacturing technology, the variation of the Lateral Glass Thickness Distribution can vary from 35% to 55% of the average glass wall thickness. In order for the product to be sufficiently strong (minimum failure), this glass thickness variation is compensated for by making the glass wall extra thick. This not only makes the product heavier (more glass), but also more raw materials are used, it costs more energy to produce the product (melting and annealing) and the transport costs of the glass product are higher due to the extra weight. By minimizing the variation of the lateral glass thickness distribution of the product, the design of the product can be adapted with a thinner (more constant) glass wall thickness. The product becomes lighter, production costs decrease proportionally, as do transport costs (and the associated CO2, NOx emissions).

To minimize the variation in the Lateral Glass Thickness Distribution in the industrial glass forming process one needs to have sensors that can determine the LGV in the glass forming process. With the help of these sensors, the root causes of the variation of LGV can be investigated in the production process. It is possible to find out which process settings or process components are responsible for an excessive variation in the LGV. If these causes of the variation of the LGV are known, one can improve the responsible process steps by, for example, optimizing the setting by using the measurement data from the sensor. This can also be done automatically with a feedback system (FeedBack loop) to automate the optimal settings to obtain a minimum variation of the LGV. Improvements in the responsible process steps can also be made to minimize the variation of the LGV.

The object of the invention is to improve the known inspection process and possibly to improve the production process on the basis of the improved inspection process.

The method for inspecting hollow glass products according to the invention is characterized in that inspecting the glass products comprises the following steps: d. successively conveying along a predetermined path the in step b. shaped glass products past at least one glass thickness sensor such as a Chromatic Confocal sensor, wherein at least one glass thickness is measured per glass product at a plurality of the glass products which are successively transported past the at least one sensor using the at least one sensor, step d. between step b. and c. is performed and wherein in step d. of each glass product of the plurality of glass products the rotational position of the glass product about an axial axis of the glass product relative to the at least one sensor is determined, wherein in a step e. the determined glass thicknesses and the associated rotational positions of the plurality of products are processed in combination to obtain information about a lateral glass thickness distribution about an axial axis of a virtual glass product representing the plurality of glass products. The invention is based on the first insight that a production error or production deviation in a glass product will often occur to a comparable extent and in comparable positions in all successively manufactured glass products. When the glass products are formed in step b. they all have the same rotational position. A deviation in the forming process of the products is then always at this same position with the successively made glass products. However, when the glass products are transported along the path, the products have acquired a more or less arbitrary rotational position. This is a consequence of the way in which glass products are transported after performing step b. Thus, when a measurement is performed on a glass product with the at least one sensor, the glass product has a more or less arbitrary rotational position relative to the at least one sensor. This means that products different from each other will have different rotational positions relative to each other and thus to the at least one sensor when a measurement is successively performed on the products with the at least one sensor. Each measurement on the glass product from a more or less arbitrary angle of view relative to the glass product thus relates to a measurement at a specific position of the product. By performing measurements on a plurality of products and analyzing these measurements in combination with information about the rotational position of the products concerned relative to the at least one sensor, an impression can be obtained of the glass thickness at different positions of the average manufactured product. on which the measurements were performed.

Because measurements are generally taken at different positions for different products, when measurements are taken on a sufficient number of products, an impression can be obtained of the glass thickness at different positions in the product manufactured on average in terms of glass thickness. This average product is also referred to here as a virtual product.

A virtual product is thus a product with a lateral glass thickness distribution which is 1s composed of measured glass thicknesses which are obtained by measuring 1s on different products. Each glass thickness at a certain position of the virtual product then corresponds to at least a glass thickness that has actually been measured at the same position on at least one real product (or with an average of glass thicknesses measured at the same position for different products).

With sufficient measurements on different products that have a more or less mutually different rotational position with respect to the at least one sensor, a lateral glass thickness distribution can thus be obtained which extends completely around an axial axis of the virtual glass product. And that while only a single sensor needs to be used. The result is as if a plurality of sensors arranged in the tangential direction of the product around a single product, measure thicknesses at mutually different positions around the axial axis. The difference, however, is that it does not concern a plurality of thickness measurements on a single (same) product and spread around that product, but a plurality of products where per product of the plurality of products at least one thickness measurement at a position of that product is carried out. In this way an impression can be obtained of a lateral glass thickness distribution of an average lettuce product that is formed.

The lateral glass thickness distribution can be expressed in numbers (for example a glass thickness at a certain position of the glass product (absolute glass thickness distribution) or a deviation from an average glass thickness (relative glass thickness distribution)).

Based on the lateral glass thickness distribution, it can be determined whether, for example, a glass thickness distribution is within predetermined limits. If this is not the case, the glass products measured with the sensor can for instance be rejected, but it is also possible to adjust a parameter of the glass production process, such as, for example, the temperature at which the glass product material is heated in step a. forming the heated glass product material into the glass product in step b. For this molding, for instance, use can be made of molds. Adjusting step b. can then be, for example, replacing a mold with a new mold. Also in step b. use is made of gutters through which the glass product material flows in the form of a glass drop towards a mould. Such gutters can, for instance, be lubricated with a lubricant if a deviation from a glass thickness distribution is found. Other adjustments are of course also possible. These adjustments can then be made automatically. However, it is also possible that some adjustments from step b. be performed manually.

Preferably, the glass thickness distribution comprises absolute values of the glass thickness distribution.

The glass thickness distribution can also be shown on a screen by means of a 3D image of a transparency of the virtual glass product. With each of the above-mentioned possibilities, use can be made of the fact that the shape and size of the products that are being measured are known in advance. After all, products are manufactured that are specified in advance. In other words, the specified dimensions (including shape, size and including wall thicknesses) of the product can be copied from the virtual product and corrected on the basis of the measurements on real products where the measurements deviate from the specifications.

For this virtual product, the wall thicknesses can then be drawn in or noted in a table based on the wall thicknesses and associated rotational positions measured for different products.

In particular, it holds that each product is transported in a horizontally oriented plane, with the axial axis of the product oriented vertically.

It further preferably holds that with a sufficiently large number of products from a product flow, measurements are performed with the at least one sensor so that the measuring positions of these measurements at the virtual product have adjacent distances in tangential direction, each of which is smaller than a predetermined value.

It also applies in particular that in step e. the lateral glass thickness distribution is determined in a region of the virtual glass product extending around the axial axis of the virtual glass product. Furthermore, it may in particular hold that step e. is performed repeatedly to obtain a lateral glass thickness distribution in a second region of the virtual glass product extending around the axial axis of the virtual glass product with the first and second regions offset from each other in the axial direction. In particular, it applies here that the at least one sensor is moved in the axial direction for obtaining recordings in the different areas.

According to an advanced embodiment of the method, in step d. the at least one sensor successively determines a plurality of wall thicknesses of a glass product that is conveyed past the at least one sensor, so that the determined glass thicknesses relate to different positions of a wall of the glass product, which positions are in a direction in which the product in step d. being transported are separated from each other. In practice, for example, 300 measurements per second are performed. The transport speed of a glass product past the at least one sensor 1s, for example 0.5 m/sec so that measuring points on a glass product are approximately 0.5/300 = 0.166 cm measured along the outer surface (assuming that a curvature of the outer surface is not too great, otherwise this distance will be slightly greater). If the product has a substantially cylindrical design with a vertical axial axis that is perpendicular to the transport direction and a diameter of 15 cm, the sensor can cover an area that has a maximum dimension in the transport direction of, for example, 5 cm measured along the outer surface. of the product.

This is because to perform a measurement, a line along a measurement direction of the sensor at a position where the line intersects the product first, an angle with a normal of the surface at said position must be smaller than, for example, 5 degrees (at a first possible measurement on a glass product the angle is then 5 degrees with respect to the normal, with a last possible measurement 1s this angle is -5 degrees with respect to the normal and halfway through the number of measurements the angle is approximately 0 degrees). After all, the sensor is usually an active sensor that emits light and receives reflections for determining the glass thickness.

If the said angle is greater than, for example, 5 degrees, a reflection of the light emitted by the sensor on the surface of the product will not reach the sensor, so that no thickness measurement can be performed.

The dimension of 5 cm means that approximately 2* 5/0.166 = 60 thickness measurements can be made on the product (a first measurement at 5 degrees from the normal and a final measurement at -5 degrees from the normal). Incidentally, the measured thicknesses at those places where the said angle is not equal to 0 degrees can be automatically corrected to obtain the thickness locally in a radial direction of the glass product.

This correction can be calculated on the basis of a known geometry of the glass product and the distance over which the product in step d. 1s is transferred between the measurement at 0 degrees and the measurement at another angle, as well as a distance between the sensor and the glass product at the times the sensor is being measured.

However, the correction at the maximum angle of +/- 5 degrees is relatively small.

Furthermore, in that example it holds that the area which has approximately a length of 5 cm in the conveying direction only covers 5/Pi*15 part of the entire circumference of the product.

So this is about 0.106 part.

In order to be able to cover the entire circumference of the virtual product, at least 1/0.106 =9 products are therefore required on which a measurement is performed.

If, for example, the products passing successively through the sensor have an arbitrary rotational position around their axial axis with respect to the sensor, statistically much more than 9 successive products will have to be measured in order for the measurements 1n combination to make the virtual product completely around its axial axis. cover.

Consider, for example, 100 consecutive products.

In a preferred embodiment of the sensor, such as the confocal sensor, it applies that by means of the at least one sensor a first distance is measured between the at least one sensor and the outer surface of the part of the wall and a second distance is measured between an inner surface of the part of the wall in which the glass thickness of the part of the wall is determined from the first distance and the second distance.

In particular, in step d. the product is transported in a horizontal plane with the at least one sensor being moved in a vertical direction for measuring the glass product at different heights.

Preferably, a glass thickness of the product is determined at different heights.

It is further possible that an inclination of the product is determined from the measurement results obtained with the at least one sensor.

This is possible because, in particular with the sensor, a distance from an inner or outer surface of the product to the sensor is also determined.

Doing this at different heights results in a skew.

The glass thickness can be determined by, at a point of a wall of the sensor whose normal includes an angle with a measuring direction of the sensor which is smaller than a predetermined value such as for example 15 degrees, more particularly 5 degrees, the difference at determine that point in distance between the outer wall and the sensor and the difference at that point in distance between the inner wall and the sensor.

According to an advanced embodiment, step d. use is made of at least two sensors facing each other and facing each other, wherein the products are transported between the at least two sensors. In this way a double measurement can be made on a product and thus information about glass thickness of the product can be determined which relates to a surface twice as large as compared to the method in which only one sensor is used. In particular, it applies here that a first glass thickness of a first part of the wall of the glass product is determined with a first sensor of the at least two sensors and that a second glass thickness of a second part is determined with a second sensor of the at least two sensors. of the wall of the glass product is determined, the first portion and the second portion being on opposite sides of the glass product.

Furthermore, it applies in particular here that with the first sensor a first distance is measured between the first sensor and an outer surface of a first part of the wall which is directed towards the first sensor and that with the second sensor a second distance is measured between the second sensor and an outer surface of a second portion of the wall facing the second sensor wherein an outer diameter of the glass product is determined from the first and second distances. In particular, or instead, it may be the case that the first sensor measures a third distance between the first sensor and an inner surface of a first part of the wall facing the first sensor and that of the second sensor. a fourth distance is measured between the second sensor and an inner surface of a second portion of the wall facing the second sensor, the third and fourth distance determining an inner diameter of the glass product.

It is also possible for the first sensor to measure a first distance per glass product between the first sensor and an outer surface of a first part of the wall that is directed towards the first sensor, and for the second sensor to measure a second distance between the second sensor and an outer surface of a second part of the wall that is directed towards the second sensor, wherein the first and second distance determined for different products and associated rotational position determine an outer diameter and/or skew position of the virtual product.

According to a special embodiment, the inspection also comprises a control wherein at least one step of steps a., b. and c. is adjusted on the basis of measurement results obtained with the aid of the at least one sensor and possibly the determined associated rotation positions, more in particular on the basis of the determined lateral glass thickness distribution of the virtual product.

In this way, the production process of the glass products can be greatly improved and optimized.

The process for producing and inspecting hollow glass products of glass product material is characterized in that inspecting the glass products comprises the following steps:

d. successively conveying along a predetermined path the in step b. shaped glass products past at least one sensor for measuring a glass thickness, such as a Chromatic Confocal sensor, wherein]

with the at least one sensor of a plurality of the glass products which are successively transported past the at least one sensor, at least one glass thickness is measured per glass product, step d. between step b. and c. is performed and wherein in step d. of each glass product of the plurality of glass products on which the at least one sensor is measured, the rotational position of the glass product about an axial axis of the glass product relative to the at least one sensor is determined, wherein in a step e. the determined glass thicknesses and the associated rotational positions of the plurality of products are processed in combination to obtain information about a lateral glass thickness distribution about an axial axis of a virtual glass product representing the plurality of glass products.

The system for producing and inspecting hollow glass products is characterized in that the at least one sensor is a sensor for measuring a glass thickness such as a Chromatic Confocal sensor, the system further comprising a rotation mode unit from which output signals are applied to the signal processing unit. supplied for further processing, wherein the at least one sensor is arranged such that the molded glass products supplied to the cooling unit, in use, are sequentially measured with the at least one sensor, wherein the rotary mode unit is arranged such that, in use, the rotational position of the glass product around an axial axis of the glass product relative to the at least one sensor is determined, wherein the signal processing unit is designed to determine a lateral glass thickness per scanned product on the basis of the output signals of the at least one sensor and the rotational position unit distribution around an axial axis v to determine a virtual glass product that represents the plurality of glass products.

The system for inspecting glass products is further characterized in that the system further comprises a rotational mode unit such as a sensor wherein the at least one sensor is a sensor for measuring a glass thickness such as a Chromatic Confocal sensor wherein, in use, the at least one sensor is arranged such that the glass products are successively measured with the at least one sensor wherein the rotational position unit is arranged such that, in use, the rotational position of the glass product about an axial axis of the glass product with respect to the at least one sensor is determined, wherein the signal processing unit is arranged to determine per scanned product on the basis of the output signals of the at least one sensor and the rotational position unit a lateral glass thickness distribution around an axial axis of a virtual glass product representing the plurality of glass products.

In practice, glass products are usually produced parallel to each other in a multitude of moulds.

According to the invention, the glass thickness distribution of at least one virtual product can then be determined per mold for products produced with that mold.

The glass thickness distribution of a virtual product that has been determined on the basis of products that have been produced with a different mold is then determined separately.

If one of the molds shows a deviation that results in a deviation in glass thickness distribution of an associated virtual product, this can be determined separately for that mould.

Also, a deviation in at least one gutter that exclusively each time supplies a glass droplet to one of the moulds, this deviation resulting in a deviation in the glass thickness distribution of at least one virtual product determined on the basis of products 1s that are produced from glass drops that have flowed through the respective at least one gutter, are detected by detecting a deviation in the respective at least one glass thickness distribution.

When deviations in gutters and/or molds are thus detected, they can be corrected, for instance by adjusting a position and/or orientation of a gutter relative to a mold and/or by providing a gutter with a lubricant and/or by replacing a mold and/or by replacing a gutter.

This can be done automatically or manually.

It therefore holds for the method in particular that a plurality of steps b. parallel to each other for producing a plurality of the glass products in a plurality of production flows in parallel, each having a step b. wherein each glass thickness distribution of a virtual product is obtained on the basis of measurements on glass products that are manufactured in the same production flow, more in particular whereby on the basis of at least a certain glass thickness distribution of at least one virtual product resulting from the measurements is obtained from glass products that originate from the same production flow that production flow is controlled (automatically or manually). Controlling a production flow is here understood to mean controlling hardware by means of which the product is manufactured in the production flow. This control may include, for example, setting a position and/or orientation of at least one trough and/or a mold used in the respective production flow, supplying a lubricant to the at least one trough and/or replacing the respective production flow. mold and/or replacing the relevant gutter.

In particular, it applies to the system that production flows that each contain a step b. wherein the signal processing unit is arranged such that, in use, any glass thickness distribution of a virtual product is obtained based on measurements on products manufactured in the same production flow.

More in particular, the system furthermore applies that on the basis of at least a certain glass thickness distribution of at least one virtual glass product obtained from measurements on glass products originating from the same production flow, said production flow is automatically controlled.

It therefore applies to a method and system according to the invention that in particular the quality of each production flow is checked separately. A lateral glass thickness distribution is then determined per production flow. If, for example, in a production flow too great a deviation in a glass thickness distribution is determined, this production flow can be intervened automatically or manually,

for example by generating an alarm related to the relevant production flow or by adjusting the relevant production flow or by replacing a mold used in the relevant production flow.

According to the invention, the sensor is not a passive infrared light-sensitive sensor for obtaining an image of the product, such as an infrared sensor.

According to the invention, the at least one sensor for measuring the glass thickness can also operate according to the principle of laser interference.

With laser interference, interference is measured between emitted laser radiation and the laser radiation received by means of reflection on the glass product and/or transmission through the object.

The at least one sensor used according to the invention is thus an active sensor.

An active sensor emits radiation and measures reflections of this radiation to determine a distance.

Thus, the at least one sensor is not a passive sensor that only receives radiation like an IR camera.

An active sensor is here also understood to be an assembly of which a first subsensor emits radiation and of which a second subsensor receives reflections on the glass product of the emitted radiation and/or transmissions through the slash product of the emitted radiation.

The invention will now be further elucidated with reference to the drawing.

In the drawing: Figure 1A shows a possible embodiment of a system according to the invention for carrying out a method according to the invention; Figure 1B shows a possible alternative embodiment of a system according to the invention for carrying out a method according to the invention; Figure 2A shows the use of a part of the system of figure 1A; Figure 2B shows an alternative use of a part of the system of Figure 1A;

Figure 2C shows the use of a part of the system of figure 1B; Figure 3A shows a part of the use according to figure 2A; Figure 3B shows a part of the use according to figure 2B; Figure 3C shows a part of the use according to figure 2C; Figure 4 shows a possible embodiment of the glass product manufactured in the system of figures 1A and 1B; Figure 5 is a schematic 3D view of a glass product showing areas hatched from which an LGV 1s is determined; Figure 6 is a schematic top view of a possible embodiment of the product forming device of figures 1A and 1B.

Figure 7 shows schematically a possible embodiment of the sensor of the system according to figures 1A and 1B; and Figure 8 shows schematically an alternative embodiment of the sensor of the system according to Figures 1A and 1B.

In Fig. 1A, reference numeral 1 denotes a system according to the invention for carrying out a method according to the invention. The system comprises a schematically shown heating device 2 for heating glass product material so that the glass product material comes into a molten state. The molten material is transported to a product forming device 3. Possibly also other materials are supplied to the product forming device 3 if necessary, such as other materials and/or semi-finished products. In this embodiment, the product forming device 3 successively produces one hollow lettuce product 4.1 1=1,2,3,...) in a production flow. The glass product 4.1 is produced after the glass product 4.1-1. In this example, the product forming device 3 comprises one casting mold 104 known per se (shown schematically in Fig. 1A) into which a part of the heated glass product material is introduced and at least one gutter 102.

(shown schematically in Figure 1A) to guide a glass drop to the mold (the glass drop slides through the inclined trough to the respective mold). Also, the product forming device 3 comprises blowing means (not shown) for blowing or pushing the glass product material into the mold to obtain the final shape of the glass product. Although in this example the product-forming device is provided with one mold for manufacturing a glass product in one production flow, it is of course also possible for the product-forming device to be provided with a plurality of molds for each time manufacturing a plurality of molds parallel to each other. products in a multitude of parallel production flows. This example will also be discussed after discussing the variant with one mold on the basis of a variant with six molds.

The successively formed glass products 4.1 are placed on a conveyor belt 6 by means of a placing unit 5 . For different values of 1, a product is therefore produced at different times because each time one product is produced at the same time with one and the same mold in a single production flow.

The glass products 4.1 produced as described above are transported by means of the conveyor belt 6 to a position P where inspection of a glass product can take place as will be explained hereinafter. With the aid of the conveyor belt, the products are then transported further to a cooling device 7 for cooling the glass product. The direction of transport of the conveyor belt is indicated by means of the arrow 8.

The system is provided with at least one sensor 10 for measuring a glass thickness of a glass product 4.1 when the relevant glass product 4.1 is at the position P. In this example, the sensor 10 is a Chromatic Confocal sensor known per se. The principle of operation of the Chromatic Confocal sensor is shown in figure 7. The sensor generates a beam of light with different colors. The focal length F depends on the color of the light.

The beam includes a plurality of colors, in particular a full spectrum of white light.

As an example, in Figure 7, the focal length for blue light with a certain frequency is indicated by F1. The focal length for green light of a certain frequency is indicated by F2, the focal length for yellow light of a certain frequency is indicated by F3, and the focal length for violet light of a certain frequency is indicated by F4. In reality, there are infinitely many focal lengths because the beam includes all colors of visible light.

In figure 7 the light falls on a transparent product with several layers of glass stacked on top of each other.

L.1 (l= 1, 2, 3..5). The light will reflect at the interface G.g (g= 1, 2, 3,4 5, 6,).

Reflections occur on the outer surfaces of the layers L.1, for the sake of simplicity it has been assumed that outer surfaces of different layers and abutting each other coincide perfectly and thus form a single interface.

This is done only to explain the operation of the sensor and is not a limitation of the invention.

At the interface G.1 blue light will reflect with a certain frequency.

The focal length of this light coincides with the interface G.1. The reflected light is passed through the sensor

10 and the sensor outputs a signal representing the color of light being received.

The reception of this color of light can be indicated in a frequency spectrum with a peak P.1. At the interface G.2 green light will reflect with a certain frequency.

The focal length of this light coincides with the interface G.2. The reflected light is received by the sensor 10 and the sensor outputs a signal representing the color of light being received.

The reception of this color of light can be indicated in a frequency spectrum with a peak P.2. In general, at the interface G.g will reflect light at a certain frequency.

The reflected light is received by the sensor 10 and the sensor outputs a signal representing the color of light being received. The reception of this color of light can be indicated in the frequency spectrum with a peak P.g. It will be clear that in the frequency spectrum P.1 corresponds to the distance A.1 between the sensor 10 and the interface G.1 It also applies that in the frequency spectrum P.2 corresponds to the distance A.2 between the sensor 10 and the interface G.2. More generally, in the frequency spectrum, P.g corresponds to the distance A.g between the sensor 10 and the interface G.g. The thickness of layer L.1 then corresponds to the distance in the spectrum between peak P.1 and peak P.2. More generally, the distance between peak P.g and P.g+1 corresponds to the thickness of layer L.g. (g = 1,2,3,4,5,,6). It will be clear that with the sensor 10 a wall thickness of a glass product 4.1 can thus be measured. Other types of sensors for measuring the wall thickness of the glass product can also be used as will be explained hereinafter.

In this example, the system of figure 1A is further provided with a rotational position unit 11 for determining a rotational position Ri of the product 4.1 about an axial axis A of the product 4.1 with respect to the sensor 10. In this example, the rotational position is unit provided with a camera 11 (which is placed on top of the sensor 10 1s) for taking a picture of the glass product 4.1 when the relevant glass product 4.1 is at the position P. In this example it holds that with the sensor 10 a glass thickness of a glass product is measured when an axial axis of the relevant glass product is at least substantially intersected by a line 20 which represents a measuring direction of the relevant sensor. At the same time, a recording of the relevant glass product is also made with the camera 11 . It thus further applies that with the at least one sensor a glass thickness is determined in each case of a part Q of a wall of a glass product, of which an outer surface O is directed towards the at least one sensor.

(see Fig. 3A). It further applies that with the at least one sensor a first distance A1 is measured between the at least one sensor 10 and the outer surface O of the part of the wall and a second distance A2 is measured between an inner surface I of the part of the wall. wherein the glass thickness of the part Q of the wall is determined from the first distance and the second distance.

A signal from the sensor is supplied to a signal processing unit 14 via line 12 which is connected to the sensor 10 . Via line 13, which is connected to the camera 11, a signal from the camera 11, which represents the recording, is supplied to the signal processing unit 14. The signal processing unit 14 is connected via a line 16 to a display 18.

The operation of the system according to the invention is as follows.

In a step a., glass product material is heated with the heating unit 2. Subsequently, in a step b. with a product forming device the heated and molten 'liquid' glass material (which then often has the shape of a drop) formed into a glass product

41. In one step c. the glass product is cooled with a cooling device 7.

The glass product 4.1 formed in the product forming device is placed on the conveyor belt 6 by means of the displacement unit 5 for transport in the direction 8. As stated, glass product 4.1 is produced after glass product 4.i-1. The glass products in this case are bottles as shown in Figs. 4 and 5. The glass product 4.1 is provided with an axial axis A which in this example is oriented vertically. If a product 4.1 formed at the position P arrives, a glass thickness measurement on the product is performed with the sensor 10.

Furthermore, in this example it holds that with the aid of the signal processing unit 14 from the image made with the camera 11 of the product 4.1, the rotational position R.i of the glass product 4.1 on the conveyor belt around its axial axis A relative to the measuring direction in this example 20 of the sensor is determined. This can be done, for example, by detecting in an image made with the camera (using for example pattern recognition techniques) where there is a mark and/or a seam and/or a dot M of the glass product. Leaving aside the difference in height, the measuring direction 20 of the sensor 10 is equal to the direction of the optical axis / viewing direction 21 (see Fig. 4) of the camera 11.

Since the recording with the camera 11 thus already comprises information about the rotational position R.i of the glass product 4.1 that is visible on the recording, it is assumed that the rotational position in step d. is determined. Since the rotational position is recognized by the signal processing unit from a recording, it can also be said that the rotational position in step e. 1s determined. The rotational position Ri of the glass product 4.1 can then be, for example, an angle Ri relative to the measuring direction 20 (see figures 1A and 3A). After all, when the glass product is manufactured, it can be provided beforehand with a marking such as a dot or a seam. When the glass product is placed on a conveyor belt by means of the positioning means 3, knowledge about the rotational position of the product is lost because the product can rotate about its axial axis during placing. It therefore holds that on the conveyor belt each product has a more or less arbitrary direction of rotation around an axial axis of the relevant product. Each product that is placed on the conveyor belt thus has a rotational position that varies per product around an axial axis of the product. It therefore holds that the glass products between step b. and c. is transported along the path on a conveyor belt, in particular each glass product between step b. and c. is placed on the conveyor belt with a rotational position varying per glass product around an axial axis of the glass product. Furthermore, in this example it holds that the glass product between steps b and c. is transported on the conveyor belt from a product forming device such as a mold in which the glass product in step b. is shaped to a cooling device in which the product in step c. is cooled. In this example it therefore also holds that the rotational position of each of the glass products of the plurality of glass products is determined about its axial axis on the conveyor belt. Furthermore, in this example it holds that with the aid of camera 11 the rotational position of a glass product of the plurality of glass products is determined by recognizing a predetermined marking (such as a seam or a dot) in an image of the relevant product camera 11 was created.

Thus, of each product 4.1 arriving at the position P, at least one glass thickness measurement is carried out with the sensor 10, the information of which is supplied to the signal processing unit 14 via line 12 . Also, of each product 4.1, the recording made with the camera 11 is fed to the signal processing unit 14 via line 13.

The signal processing unit determines the rotational position R.i of the respective product from the information about the image 4.1. Of course, in step d. the rotational position can also be determined in another way, for instance with a different type of sensor 11, the signals of which containing information about the rotational position are supplied to the signal processing unit.

In Fig. 1A, M.i 1 =.1,2,3,..) denotes the marking of the glass product 4.1 (1 = 1,2,3,..) respectively. As can be seen from the markings M.i, the glass products have an arbitrary rotational position with respect to the axis 6' of the conveyor belt 6 and thus an arbitrary rotational position R.i with respect to the measuring direction 20. This is in contrast to the product-forming device 3 in which all products 4.1 as soon as they are formed have the same direction of rotation with respect to the axis 6'. As stated, however, this rotational position can change when a product 4.1 is placed on the conveyor belt.

Because the products 4.1 have an arbitrary rotational position Ri on the conveyor belt 6, the successive measurements on the products 4.1 will represent the products in each case a glass thickness relative to the marking Mi from a different measuring direction. For simplicity, it is assumed that 1 wall thickness measurement is performed per product (an example is given later of a variant in which a plurality of thickness measurements are performed per product at different positions near the product). Assuming that all products are the same, with a sufficient number of measurements, a collection of wall thicknesses that are more or less regularly spread in the tangential direction over a complete revolution around the axial axis of the product is available.

If there is an error or an incorrect setting in, for example, the glass product forming device, this will usually be present in all glass products 4.1 at the same position with respect to the marking M.i. By analyzing sufficient thickness measurements on the successively made products 4.1, an error or deviation can be found in the entire first region. A completely circular wall thickness profile of a virtual product 4' can therefore be constructed from a sufficient number of images of the products 4.1. Use can be made here of the known specifications of dimensions (including shape, size and wall thicknesses) of the products that are being measured. This also makes the shape and size of the virtual product known. With this virtual product, the wall thicknesses measured on real products can then be drawn in or noted in a table on the basis of the wall thicknesses and associated rotational positions measured for various real products. In other words, the specified dimensions (including wall thicknesses) of the product are taken over for the virtual product and corrected on the basis of the measurements, where the measurements deviate from the specifications. It thus holds that with a sufficiently large number of products from a product flow, measurements are performed with the at least one sensor 10 so that these measurements in combination cover the associated virtual product around the axial axis A of the virtual product with mutually adjacent distances in tangential direction. each of which is less than a predetermined value. The direction of the axial axis of a product 4.1 is the same as the axial axis of the virtual product 4'.

It holds here that in step e., the lateral glass thickness distribution is determined in a first region 26.1 (see figure 4) of the virtual glass product which extends around the axial axis of the virtual glass product.

This is further clarified in Figure 2A. In figure 2A, sensor 10.8 indicates the relative virtual position of the sensor 10.8 with respect to the glass product 4.8 when a thickness measurement is performed on the glass product 4.8 at the position P. In figure 2A, the position and orientation of the glass product 4.8 has been chosen such that M.8 coincides with an arbitrarily fixed position M' (in this example, fixedly chosen 'at 12 o'clock' in the drawing). Sensor 10.12 also indicates the relative virtual position of the sensor 10.12 with respect to the glass product 4.12 when a thickness measurement is made on the glass product 4.12 at the position P. In figure 2A the relative position of the product 4.12 is again chosen such that M.12 coincides with the arbitrarily fixed position M'. In figure 2A the rotational position of the glass product 4.8 is thus shown to coincide with the rotational position of the product 4.12, such that M.8 coincides with M.12. If the real position of the sensor 10 with respect to the glass product 4.12 is viewed when the glass product 4.12 is in the position P, the measuring direction 20 of the sensor 10 and the rotational position of the product 4.12 measured with respect to the marking M are closed. 12 into an angle R.12 (see figure 1A). The rotational position of the product 4.12 is indicated in figure 1A by the line L12. The line L12' is parallel to the line L12 and indicates the rotational position of the product 4.12 if it were in the position P. In that position, the rotational position of the product

4.12 relative to the measuring direction 20 of the sensor with the angle R.12. In figure 2A the same rotational position R.12 is indicated. However, because in figure 2A the rotational position of the product 4.12 is shown to coincide with the rotational position of the product 4.8 (after all, M.8 and M.12 coincide in M in figure 2A), the virtual position of the sensor 10.12 will not coincide with the virtual position of the sensor 10.8. In figure 2A, completely analogously for the products 4.7, 4.9, 4.10 and 4.11, the virtual positions of the associated sensors 10.7, 10.9, 10.10 and 10.11 are indicated. The measurements of the virtual sensors 10.7-10.12 in Figure 2A in combination form a measuring range that extends around a virtual product 4'. The position of the sensor 10.1 in Figure 2A can thus be obtained by rotating the product 4.1 and the sensor 10 from Figure 1A such that M.i coincides with M'. In other words, a measurement with the virtual sensor 10.1 on the virtual product 4 in Figure 2A is equivalent to a measurement with the real sensor 10 on the product 4.1 in Figure 1A.

In figure 2A, the position of the virtual sensor 10.12 is therefore chosen such that it has the rotational position R.12 with the virtual product 4 with marking M'. In other words: in figure 2A the product 4.12 and the sensor 10 are rotated around the axial axis of the product 4.12 at the position P such that the marking M.12 of the product 4.12 also coincides with the marking M' of the virtual product. In figure 2A, after rotation, the virtual position of the sensor 10 for product 4.12 is denoted by 10.12 and after rotation the virtual position of the sensor 10 for product 4.8 is denoted by 10.8. This has been done completely analogously in figure A2 for the products 4.7, 4.9, 4.10 and 4.11. For example, completely analogously in figure 1, the rotational position of the product 4.11 is indicated by the line L11. The line L171' is parallel to the line L11 and indicates the rotational position R.11 =0 of the product 4.11 if it were at the position P. In that position, the rotational position of the product 4.11 relative to the measuring direction 20 of the sensor 10 can be indicated by the angle R11. In figure 3A the rotational position R.i for the random product 4.1 is indicated.

In this example, when each time a measurement on six successive products is combined, the positions at which the glass thickness is determined for the virtual product are reasonably distributed around the axial axis of the virtual product, see for example figure 2. The wall thicknesses at these positions give in combination information about a lateral glass thickness distribution in a first area 26.1 (see figure 4 for the areas at the products 4.1) of the virtual product 4' (see figure 5 for the areas at the virtual product 4'). Of course, this is a static process in which the distribution can vary when each time 6 measurements based on 6 real glass products are used to obtain information about a lateral glass thickness distribution. This process can then be repeated for the products 4.13-4.18 to again determine the lateral glass thickness distribution in the region 26.1. In this way a possible course in time (trend) of the lateral glass thickness distribution can be detected. For example, if a trend in a deviation from successively determined lateral glass thickness distributions is determined, it can be deduced that a particular part, such as a mold, IS wearing out.

Of course, it is also possible to combine 8 measurements based on 8 real products in order to obtain information about a lateral glass thickness distribution for a virtual product. It is also conceivable that each time with a sufficiently large number of products from a product flow, measurements are performed with the at least one sensor so that these measurements in combination cover the associated virtual product around the axial axis of the virtual product with mutually adjacent distances (neighboring distance between measurement positions) in tangential direction, each of which is less than a predetermined value.

For the sake of clarity, such a mutual distance between two neighboring measurements is indicated by Z1 in figure 2A. Thus, in Figure 2A, there are G of such distances to be defined, Z2 is a second adjacent distance to be defined, etc.

A glass thickness measurement is then made each time on a new product that is located at position P, until the collection of all measurements that have been made up to that moment is so large that, in combination, the corresponding virtual product around the axial axis of the virtual product with mutually adjacent distances in tangential direction that are each less than a predetermined value.

Z is then smaller than the predetermined value.

If necessary, this can therefore take place in, for example, 20 measurements, so that 20 adjacent distances are formed, each of which is smaller than the predetermined value.

After this, a new series of measurements of successive products is started until this set of these thickness measurements is sufficiently large 1s that the measurements of this set cover the associated virtual product around the axial axis of the virtual product with mutually adjacent distance in tangential direction that are each smaller than a predetermined value.

Because this method is repeated over time, a good impression can also be obtained of the course of a lateral glass thickness distribution over time.

It will be appreciated that performing the plurality of measurements in step d. between steps b. and c. is carried out.

The result of each measurement made with the sensor 10 is fed to the signal processing unit 14 via line 12, respectively.

These signals are added in a step e. together with the associated information about the rotational positions are processed in combination to obtain said information about the lateral glass thickness distribution about an axial axis of the virtual glass product 4' representing the plurality of glass products 4.1 as discussed above.

In particular, the glass thickness distribution indicates relative variations in glass thickness.

It is also possible that the glass thickness distribution comprises absolute values of the glass thickness distribution. Furthermore, it is possible that the later glass thickness distribution in a 3D image of the product 4' is displayed on the display 18. For this purpose, for example, information is stored in the signal processing unit about the dimensions (including shape and wall thickness) of the product to be produced. This is also referred to as the product specifications. Ideally, a manufactured product will have the exact same dimensions as specified. However, the lateral glass thickness distribution in the first region determined by means of the sensor may deviate from this. The determined glass thickness distribution can now be added to an image of the specified product to obtain the 3D image of the virtual product 4'. This virtual product then shows the measured wall thicknesses in the first area. Although the positions on the virtual product to which the measured wall thicknesses relate are separated in the tangential direction, the wall thickness shown in the 3D image may be a wall thickness that progresses gradually and connects the actual measured wall thicknesses (similar to the drawing a smooth line through a number of measurement points in a graph). However, the information obtained about a lateral glass thickness distribution can also be shown in a table on the display.

In the case discussed, the sensor 10 is used to measure in a first area 26.1 (see figure 4) in which the LGV is determined with the signal processing unit (for example the LGV at half height of that area), see figure 5, where the half height of area

26.1 is dotted. In that case, with the aid of a second sensor 10° which is located at a greater height h than the sensor 10 and whose output signals are also supplied to the signal processing unit (not shown in the drawing), a lateral glass thickness distribution in a second area can be obtained 26.2 extending around the product 4' can be determined with the signal processing unit (see Fig. 5). The second region 26.2 is axially offset from the region

26.1 (see figure 5). For example, the LTV in the second region 26.2 is also located at half height of this region (indicated by dotted lines in Figure 5). It is also entirely analogous that the lateral glass thickness determination in an area 26.3 is determined with the aid of a third sensor (not shown in figure 5, but in figure 4). The height of an area 26.1,

26.2 and 26.3 1s, for example, has been chosen such that the glass thickness hardly changes in the vertical direction (= axial direction A). The glass thickness within such a range will then, if it changes, vary in particular in the tangential J0 direction. The areas 26.1, 26.2 and 26.3, for example, connect to each other. It therefore holds for the areas 26.1 and 26.2 that steps d. and e. are applied repeatedly to obtain a lateral glass thickness distribution LGV in the first and second regions of the glass product extending around an axial axis of the product with the first and second regions offset from each other in the axial direction A.

In particular, for regions 26.1 to 26.3, steps d. and e. be performed repeatedly at least three times respectively to obtain lateral glass thickness distributions in respectively at least three mutually different regions each extending around the axial axis and offset from each other in axial direction and preferably in combination, at least substantially cover the entire glass product 4'. It is also possible that the lateral glass thickness distribution in different areas such as the areas 26.1 and 26.2 is determined with one and the same sensor 10 . To this end, the sensor 10 in the system shown in figure 1A must be height-adjustable, so that measurements can be taken per product at different heights, or measurements can be made at a first height for a number of successive products, and at a second height for a different number of successive products. It is also possible that after measuring the wall thickness to obtain a glass thickness distribution in the region 26.1 it is also possible that the sensor 10 is moved upwards in the axial direction to perform a measurement to obtain a lateral glass thickness distribution in the area 26.2. The sensor 10' is then not present and the sensor 10 can then be moved upwards for measuring in the area 26.2 after the sensor 10 has measured in the area 26.1.

After this, the sensor can be moved down to perform a measurement in the area 26.3, etc.

Thus, in the above manner, the lateral glass thickness distribution LGV of the virtual product as a function of h and phi has been determined.

Here h is determined by the height at which the sensor 10 in figure 1A is arranged.

This is explained in more detail with reference to figures 4 and 5. If LGV as a function of certain values of h and phi is known, a continuous function LGV (h, phi) can be determined which for the certain values of h and phi is the measured glass thickness.

The glass thickness distribution LGV (h,phi) can be shown on the display 18 via a line 16, for example by means of numbers, a graph or an image.

LGV is then, for example, a number that indicates the absolute wall thickness at the position (h,phi).

However, it is also possible that LGV indicates a relative wall thickness to a reference wall thickness at the position (h, phi).

Also, the lateral wall thickness can be shown in a 3D image of the glass product shown on the display.

This image then shows, for example, a view of the glass product in transparent.

An operator can monitor the glass thickness distribution of the successively produced products 4.1 in this way.

When the glass thickness distribution starts to deviate, for instance because the wall thickness of the product becomes too large or too small in certain places, the production process can be intervened, i.e. the method step b can be intervened.

Such an operation can be performed manually.

However, it is also possible that such an intervention is carried out automatically with the aid of the signal processing unit, in this case via a feedback control line 30. In the event of a deviation from a glass thickness, it is possible to correct the distribution manually or automatically in the production process of the glass. Products. It is also possible, upon change per unit time in glass thickness, to correct or control distributions successively obtained in time and exceeding a predetermined value, manually or automatically under the control of the signal processing unit via line 30. Under correcting or controlling can for instance be understood to mean adjusting a position and/or orientation of a gutter and/or mould, providing a gutter with a lubricant and/or replacing a mould. All this is schematically indicated in Fig. 1A.

It is further possible that a calibration measurement is performed on a glass product with a known glass thickness, on the basis of which step e. is carried out. After all, if the glass thickness is known during the calibration measurement, the signal processing device 14 can correct the lateral glass thickness distribution of the product 4'. The invention is by no means limited to the embodiments outlined above.

Thus, it is clear that the sensor 10 and the signal processing unit 14 can also be used in production processes for forming glass products other than those described here. In fact, the sensor in combination with the signal processing unit 14 forms an essential part of the invention. According to the invention, the cooling device 7 could also be omitted. After all, even without cooling device 7, the products will eventually cool down by themselves, so that step c. can also be carried out without additional aids. Also, the cooling device may or may not be controlled manually or automatically (for example the temperature of the cooling device) on the basis of the determined LGV.

In this example, a lateral glass thickness distribution at the virtual product 4' is obtained from measurements on the products 4.7-4.12. As stated, hereinafter, for instance on the basis of products 4.13-4.18, a next lateral glass thickness distribution can be obtained for the virtual product 4, completely analogous to that discussed above for products 4.7-4.12. This can be seen as an update in time of the instantaneous lateral glass thickness distribution.

More generally, from m measurements on the products 4.1+k with k=0, 1,2,3,...m-1 the glass thickness distribution of a virtual product can be obtained. Subsequently, a glass thickness distribution of a virtual product can be obtained from m measurements on the products 4.1+m+k with k=0, 1,2,3,...m-1. After this, a glass thickness distribution of a virtual product can be obtained from m measurements on the products 4.1+2m+k with k=0, 1,2,3,...m-1 and from m measurements on the products 4.1+ 3m+k with k=0, 1,2,3,...m-1 a glass thickness distribution of a virtual product can be obtained. This can be continued so that after every m measurements a glass thickness distribution of a virtual product is determined. This implies that changes of the glass thickness distribution over time can be observed. This makes it possible to detect changes in the glass thickness distribution of successively determined virtual products due to e.g. wear in gutters and the mold or changes in the glass thickness distribution of successive virtual products due to varying settings of the gutter and/or mold. When such a change exceeds a limit value, the production flow can be controlled automatically or manually (such as setting a position and/or orientation of a trough and/or mould, supplying a lubricant to a trough, replacing a mould). , etc).

It is also possible that m is variable. In that case, in general according to the invention (apart from the embodiments), the signal processing unit is arranged to select a sufficient number of successive measurements so that the virtual sensors associated with these recordings (Figure 2A) detect the associated virtual product around the axial axis of the virtual product with adjacent distances measured at angles in the tangential direction, each less than a predetermined angle value. This distance can be, for example, 1/20th of the circumference so that an accurate course of the LGV is obtained.

The signal processing unit then automatically decides which and how many images are to be processed in combination for calculating a glass thickness distribution of a virtual product.

In particular, in step d. the at least one sensor 10 successively determines a plurality of wall thicknesses in a glass product that is transported past the at least one sensor, so that the determined glass thicknesses relate to different positions of a wall of the glass product which positions in a direction in which the product step b. being transported are separated from each other. This has the advantage that glass thicknesses are determined with the at least one sensor 10 at a plurality of positions which are separated from the product in the tangential direction. Suppose this involves three measurements, as shown in figures 2B and 3B. In figures 2A and 2B corresponding parts are provided with the same reference numerals. Also in figures 3A and 3B corresponding parts are provided with the same reference numerals. In figure 3B this means that the wall thicknesses are measured at the positions S1.1, S2.1 and S3.i of product 4.1. This is because, while the product 4.1 is being transported in the direction 8, glass thickness measurements are successively performed with the sensor 10 at the positions S1.1, S2.1 and S3.i, respectively. In figure 3B, the positions S1.1 for all values of 1 (i.e. for all glass products 4.1 that are measured at) coincide. Also in figure 3B the positions S2.1 for all values of i coincide. Furthermore, in figure 3B the positions S3.1 for all values of 1 coincide. In figure 2B the positions S1.1 for all values of 1 (here 1 = 7,8,9,10,11,12) just do not coincide. This also applies to the mutual positions S21 and the mutual positions S3.2) ‚In figure 2B this would mean that with product 4.7 not a single measurement was made with the virtual sensor 10.7 as in figure 2A, but three measurements. Also in the virtual product 4 the positions of the three measurements S1.7, S2.7, S3.7 on the product 4.7 in the tangential direction of the glass product 4.1 are separated from each other with the same mutual distance as in figure 3B. The processing of the thickness measurements S1.7 and S3.7 is therefore equal to the processing of the thickness measurement S2.7 as has already been described above with reference to Figs. 2A and 3A. Entirely analogously, in this example, each combination of three measurements for an arbitrary product 4.1 yields three measurements for the virtual product 4. As an example, the measurements on product

4.9 shown in Figure 2B with S1.9, S2.9 and S3.9. Incidentally, the wall thickness Dm measured at position S1.1 (see figure 3B) is equal to the wall thickness measured in measuring direction 20. This thickness is slightly larger than the wall thickness Dr which would be measured at position S1.1 in a radial direction R (= the direction of the normal of the outer surface of the glass product at the position S1.1) at the position S1.1. In practice this concerns only a small deviation when the said radial direction includes a small angle B (for instance smaller than 5 degrees) with the measuring direction. Of course, the wall thickness Dm measured with the sensor 10 can also be automatically corrected by the signal processing unit (converted into Dr) when the said angle B is known. In particular, the system can automatically determine or estimate this angle. For example, the angle B can be estimated from a measured distance X2 between the sensor 10 and an outer surface of the glass product at position S2.i and the measured distance X1 between the sensor 10 and the outer surface of the sensor at position S1.1. In figure 3B the measuring positions S1.1 and S2.1 are connected with a straight line as an approximation for the course of the outer surface. This straight line intersects an angle y with a line perpendicular to the measuring direction. This line also includes an angle 90-y with the measuring direction. The angle B between the measuring direction and the line R is part of an isosceles triangle comprising two equal angles 90-y. It follows that B= 2y. The arctangent of X1-X2/Y (where Y is the distance over which the glass product has been transported in the transport direction 8 between the measurement at position S1.1 and the measurement at position S2.1) is then equal to y where y is an estimate is for 8/2 (see also Fig. 3B). Since X1, X2 and Y are known, the angle B can be estimated from this. The thickness Dr in the radial direction at position S2.1 is then equal to the measured thickness Dm times the cosine of the relevant estimated angle for Ba. This applies completely analogously to the measurement at S3.1. The measurement can also be corrected here, in this example also with the cosine of the estimated angle for B. The thickness measured at the position S2.1 needs no correction, because here the angle B is equal to zero. Of course there are also other possibilities for determining and/or estimating the angle B. Furthermore, the angle B or an estimate for it can be determined once if the positions S1.1, S2.1 and S3.1 are fixed and are always the same for each glass product 4.1 that is measured against.

All this applies in each case to the measurements for the products 4.7-4.12, so that with the virtual product of figure 2B in fact 3*6=18 glass thickness measurements are available which are separated from each other in the tangential direction of the virtual product 4' and which are used to be able to determine the lateral glass thickness distribution more accurately than with the method according to figures 2A and 3A. The chance that with these 18 glass thickness measurements the distance in the tangential direction between neighboring measurements is smaller than the aforementioned predetermined value increases as a result. In practice, for example, 30 successive glass thicknesses can be determined for a single product, as has already been discussed above for 3 successive glass thicknesses.

As already stated, it applies in particular that in step d. the product is transported in a horizontal plane with the at least one sensor being moved in a vertical direction for measuring the glass product at different heights. This makes it possible to determine at least one glass thickness of the product at different heights.

If a glass thickness is determined at different heights, it is also possible that an inclination of the product 4.1 is determined from the measurement results obtained with the at least one sensor. Figure 4 shows an example in which a first sensor 10 and a second sensor are used. sensor 10'. With the first sensor 10, the distance Al is measured between the sensor 10 and an outer surface of the product 4.1. With the second sensor 10', a distance Al' is measured between the outer surface of the product 4.1 and the second sensor 10'. The height difference between sensors 10 and 10' is D. From the difference between the measured distances A1-A1' and the associated difference in heights D, the skew of the product 4.1 can be determined. It is also possible that the distance A1' is measured with the first sensor 10 by moving it vertically to the position of the second sensor in figure 4. The second sensor 10' is then omitted.

It is further possible that in step d. use is made of at least two sensors facing each other and facing each other, wherein the products are transported between the at least two sensors. The same measurements are then performed per sensor 10, 10° as discussed above in the context of figure 1A. In figure 1B an example is given in which the sensors 10 and 10' are used which are located opposite each other. In figures 1A and 1B corresponding parts are provided with the same reference number. The associated positions of the virtual sensors 10.1, 101 relative to the virtual product are shown in Figure 2C. Here, 10.1 are the virtual sensors associated with the sensor 10 of Figure 1B with respect to the glass product 4.1 and 10.1 are the virtual sensors associated with the sensor 10' of Figure 1B with regard to the glass product 4.1. For example, for product 4.8, the virtual sensor 10.8 relative to the virtual product 4' (which corresponds to the sensor 10 relative to product 4.8) is in the position as indicated. The position of M has been chosen the same as in figure 2A. For example, for the product 4.8 it further holds that the virtual sensor 10°.8 with respect to the virtual product 4 (which corresponds to the sensor 10' with respect to the product 4.8) is in the position as indicated. In general, in figure 2C the positions of the virtual sensors 10.7-10.12 correspond to those of figure 2A and that the position of the virtual sensor 10'.i is opposite the position of the virtual sensor 10.1 for 1=7-12 . As a result, by chance, the position of the virtual sensor 10.11 coincides with the position of the virtual sensor

10.7. Completely analogous, the positions of the virtual sensors happen by accident

10.10 and 10°.9, 10.7 and 10°.11, 10.9 and 10°.10 together respectively. At the positions where the positions of virtual sensors coincide, there are thus two measured values for the glass thickness. These can be averaged. At positions where the positions of the virtual sensors do not coincide, there is an additional measured value relative to the system of Figure 1A at a position that is new relative to the system of Figure 1. This thus gives two additional measured values for the determination of a lateral glass thickness, wherein the mutual distance between each additional measured value and its neighboring measured values is smaller than the mutual distance between the respective neighboring measured values. In other words, an average spacing between neighboring measurements decreases due to additional measurements that do not coincide with the original measurements. This gives a better insight into a lateral glass thickness distribution.

All measured values can be processed in combination as discussed for figure 1. It thus holds that with the sensor 10 of the at least two sensors a first glass thickness of a first part of the wall of the glass product is determined and with the second sensor 10° of the at least two sensors a second glass thickness of a second part of the wall of the glass product is defined wherein the first portion and the second portion lie on opposite sides of the glass product. Furthermore, it applies in particular (see figure 3C) that with the first sensor 10 a first distance A1 is measured between the first sensor and an outer surface O of a first part of the wall which is directed towards the first sensor and that with the second sensor 10' a second distance ATT is measured between the second sensor 10' and an outer surface O of a second portion of the wall facing the second sensor, from the first and second distance through the signal processing unit 14 an outer diameter of the glass product is determined. This outer diameter is equal to E-A1-A1'' where E is the distance between the sensors 10 and 10'. Furthermore, in this example it holds that with the first sensor 10 a third distance A2 is measured between the first sensor 10 and an inner surface I a first part of the wall that is directed towards the first sensor 10 and that with the second sensor 10° a fourth distance A2' is measured between the second sensor and an inner surface I of a second part of the wall facing the second sensor 10', whereby an inner diameter of the glass product is determined by the signal processing unit 14 from the third and fourth distances. This inner diameter is equal to E-A2-A2' where E is the distance between the sensors 10 and 10'. The determined inner and/or outer diameter can be displayed per product on the display 18. Alternatively, also, for example, a running average of the last x produced products can be calculated by the signal processing unit and displayed on the display 18 where x is an integer greater than 1.

In particular, it further applies that, for each product, a first distance A1 is measured between the first sensor and the outer surface O of a first part of the wall which is directed towards the first sensor 10 and that is measured with the second sensor. 10' a second distance

ATT is measured between the second sensor 10' and an outer surface O of a second part of the wall that is directed towards the second sensor 10', wherein from the first and second distances A1 and A2 determined for different products and associated rotational position an outer diameter and /or skew of the virtual product is determined.

For example, an outer diameter B.i can be determined of the virtual product from the distance Alien Ala associated with the sensors 10.1 and 10'.i from figure 2C.

This outer diameter is E-A1.1-A1'.1. In this way, 6 outer diameters are determined, which are respectively obtained on the basis of the products 4.1 (i=

7.8,..12). This means that the outer diameter of the virtual product can vary in the tangential direction.

In this example, in the virtual product, the position of the butene diameters B.7 and B.11, B.10 and B.9, B.8 and B.12 coincide.

These pairs can be averaged for a better estimate of the outer diameter of the virtual product at the position in question.

It is therefore possible that the outer diameter is determined depending on a tangential position phi at the virtual product.

After all, the outer diameter for different tangential positions on the outer surface of the virtual product corresponds to the outer diameter determined for real products whose rotational position corresponds to a tangential position at the virtual product.

Thus the outer diameter at the tangential position R8 in figure 2C corresponds to the outer diameter of the real product 4.8. The same applies to the outer diameter at other positions R.i (1= 7,9,10,11,12). The determined outer diameter for the virtual product 4' can also be an average of the six outer diameters B.i (1= 7,8,..12) of the virtual product 4'.

What has been explained above applies completely analogously to the inner diameter of the virtual product. For example, an inner diameter C.1 can be determined of the virtual product from the distance A2.i and A2'i associated with the sensors 10.1 and 10.1 from figure 2C.

This inner diameter is

EB-A21-A2'.1. In this way, 6 inner diameters are determined, which are respectively obtained on the basis of the products 4.1 (1=7,8,..12). This means that the inner diameter of the virtual product can vary 4' in the tangential direction. In this example, in the virtual product, the position of the inner diameters C.7 and C.11, C.10 and C.9, C.8 and C.12 coincide. These pairs can be averaged for a better estimate of the inner diameter of the virtual product at the position in question. It is therefore possible that the inner diameter is determined depending on a tangential position phi at the virtual product. After all, the inner diameter for different tangential positions on the inner surface of the virtual product corresponds to the inner diameter determined for real products whose rotational position corresponds to a tangential position at the virtual product. Thus the inner diameter at the tangential position R8 in figure 2C corresponds to the inner diameter of the real product 4.8. The same applies to the inner diameter at other positions R.i (1= 7,9,10,11,12).

Preferably, the inspection also comprises a control wherein at least one step of steps a., b. and c. is adjusted on the basis of measurement results obtained with the aid of the at least one sensor (such as the distance A1, A2, A1', A2', the above-mentioned diameters and/or, skew positions) and possibly the determined rotational positions, more in particular on the basis of the determined lateral glass thickness distribution of the virtual product. This control can then be determined and carried out by the signal processing unit 14 which supplies control signals via the line 30 for this purpose.

According to a practical embodiment, it holds that a plurality of the steps b. are carried in parallel to each other to produce a plurality of the products in parallel to each other in a plurality of production flows each comprising a step b. wherein each lateral glass thickness distribution of a virtual product is obtained from measurements with the at least one sensor on products manufactured in the same production flow.

In such a practical embodiment it is for instance possible that 6 products are formed parallel to each other in 6 parallel production flows. To this end, the system, in this example the product-forming device 3, is provided with 6 production flow paths 106.) each comprising, for example, at least a gutter and a mould. This is shown in Figure 6 in which a glass droplet formed in the heating device 2 via a switch 100 to a production flow path selected with the switch.

106.j (=1,2,3,..6) with each production flow path 106.) comprising at least one trough 102.j and a mold 104.j. The glass droplets which are successively transported along the production flow path 106j together form a production flow 107.J. By rotating the switch in the direction of arrows 110 about an axis 108, other production flow paths 106.j can be selected with the switch. The glass drop flows in this example via the selected at least one gutter

102.j to the mold 104. In this example, the switch 100 successively selects the value 1,2,3,4,5,6 for j. So in this example there are 6 parallel production flows.

Because there are 6 production flows, 6 products 4.1, 4.2, 4.3, ...4.6 are successively formed, which are placed in a row of 6 products on the conveyor belt. Here product 4.1 is formed from a glass drop that has been transported via production flow path 106.1 in production flow 107.1, product 4.2 is formed from a glass drop that is transported via production flow path 106.2 in production flow 107.2, product 4.3 is formed from a glass drop that is via production flow path 106.3 in production flow 107.3 transported, etc. More generally, product 4.j is formed from a glass droplet transported via production flow path 106.j into production flow 107.j for j = 1,2,3,4,5,6. When 6 products have been produced in this way, this process repeats itself.

Here, product 4.7 is formed from a glass drop that has been transported via production flow path 106.1 in production flow 107.1, product 4.8 formed from a glass drop that is transported via production flow path 106.2 in production flow 107.2, product 4.9 is formed from a glass drop that is transported via production flow path 106.3 in production flow 107.3, etc. More generally, product 4.J+6 is formed from a glass droplet transported via production flow path 106.) into production flow 107. for j = 1,2,3,4,5,6. When 6 products have been produced in this way, this process repeats itself. In general, therefore, it holds that product 4.j+n.6 with n=0,1,2,3,....formed from a glass drop that has been transported via production flow path 106.) into production flow 107.) for j = 1 ,2,3,4,5,6.

The products 4.j, 4.746, 4.j+12, ….4.j+n.6 are then made in the same production flow j, where j= 1,2,3,4,5,6 and n= 0 ,1,2,3,4, According to this variant, a number of successive measurements of products

4.j+n.6 for different values of n and the same value of j (and thus belonging to the same production stream j) are processed in combination with the signal processing unit to obtain a glass thickness distribution of a virtual glass product. This number can be, for example, 8. Thus, for a value of j, measurements on the products for n=0,1,2,..7 can be combined to obtain a glass thickness distribution of a virtual product 4'. Then for the same value of j measurements on the products for n=8, 9, 10, 15 can be processed in combination with the signal processing unit to obtain a glass thickness distribution of another virtual product 4', etc. These measurements are distributed in obtained over time and obtained from products produced using the production flow path 106.) in the production flow 107.). This allows changes in the glass thickness distribution of successive virtual products due to, for example, wear in the at least one trough and/or mold of production flow path 106.) or changes in the glass thickness distribution of successive virtual products due to varying settings in production flow path.

106.) to detect. When such a change exceeds a threshold value, settings of production flow path 106.j can be changed automatically or manually (such as a position and/or orientation of a trough and/or mold of the respective production flow path), a trough of production flow path 106.j can be changed. lubricated, a mold and/or trough of production flow path 106.j are replaced, etc. It is thus possible to correct by hand or automatically upon a change in glass thickness distributions obtained successively over time and exceeding a predetermined value. under the control of the signal processing unit 14 via line 30. It is also possible to correct a deviation in a single glass thickness distribution exceeding a predetermined value manually or automatically under the control of the signal processing unit via line 30. means adjusting a position and/or orientation of a gutter and/or mould, providing a gutter with a lubricant and/or replacing a mould.

All this can also be done for all other possible values of j so that glass thickness distributions are obtained for different virtual glass products relating to different production flow paths 106.j and therefore to different production flows 107). If there is a deviation in a glass thickness distribution of a virtual product 4' obtained from a production flow j, for example, the position of a trough and a mold belonging to the production flow path j can be corrected with respect to each other, or a trough belongs to the production flow path j are supplied with a lubricant. Also, a mold of the production flow path j can be replaced. All this can be done manually or automatically via line 30 under the control of the signal processing unit. Also, changes in glass thickness distributions of virtual products obtained sequentially using a same production flow path 106.j can be detected by the signal processing unit. If these changes exceed a predetermined value, settings of the respective production flow path 106.) can again be adjusted as discussed above, a trough of the relevant production flow path 106.) can be lubricated and/or a mold of the relevant production flow path replaced, by hand. or automatically.

In this example, each time 8 measurements on products obtained from the same production flow were processed in combination to obtain information about the glass thickness distribution of a virtual product. Instead of using a=8 (or another fixed number) of images each time, the signal processing unit can also be arranged to select a variable, for instance such that a sufficient number of successive measurements is selected so that the associated virtual product in combination with these measurements around the axial axis of the virtual product with a tangential spacing between adjacent measurements (more precisely positions of measurements) that is less than a predetermined value (as discussed above in the context of a single production flow of Figure 1A).

The signal processing unit then automatically decides which and how many measurements are each time processed in combination to calculate a glass thickness distribution of a virtual product associated with a particular production flow path j. It therefore holds in particular according to the invention that a plurality of steps b. parallel to each other for producing a plurality of the products 1n parallel to each other a plurality of production flows each having a step b. wherein each glass thickness distribution of a virtual product is obtained on the basis of recordings of products manufactured in the same production flow, in particular wherein based on at least a certain glass thickness distribution of a virtual product obtained from measurements at products that originate from the same production flow automatically control the production flow, such as adjusting settings of that production flow.

Furthermore, it therefore applies in particular according to the invention that the system is arranged such that, in use, a plurality of the steps b. parallel to each other for producing a plurality of the products in a plurality of production flows in parallel, each having a step b. wherein the signal processing unit is arranged such that, in use, any glass thickness distribution of a virtual product is obtained on the basis of measurements on products manufactured in the same production flow, in particular whereby on the basis of at least a certain glass thickness distribution of at least one virtual product obtained from measurements of products originating from the same production flow, the production flow is automatically controlled, such as adjusting settings of that production flow.

Instead of a chromatic confocal sensor as described above, other sensors can also be used to measure a glass thickness in the systems and methods described above. For example, the sensor according to figure 7 can be adapted so that it emits only one color of light and thus has only one focal point (see figure 8). By moving the sensor of figure 8 back and forth in its measuring direction 20, the sensor can detect when its focal point F coincides with one of the interfaces G.g (g = 1,2,3,...7) because the light then becomes strong. reflected and detected by the sensor. In figure 8 it has been chosen as an example that the focal point F coincides with the interface G.4. Then the sensor is moved towards the interface G.4 and can be detected with the sensor when the focal point F coincides with the interface G.5. Even then, the intensity of the received light will have a peak. This situation is indicated by dotted lines in figure 8. The sensor is moved over a distance D for this purpose. This distance then corresponds to the thickness D of the layer L.4. By repeating this for all interfaces, the thickness of each layer can be determined from the measured differences in sensor positions. Incidentally, it will be clear that with a product 4.1 the wall thickness at a certain position of the product can be determined from the difference in positions of the sensor at which a peak in reflections is received on the inner wall and the outer wall of the product, respectively. In the embodiments outlined above, sensor 10 of Fig. 1A must then be arranged to be movable in its measuring direction. It will also be clear that in this way a skew of a product can be measured by comparing the position of the sensor at which reflection at the outer surface of the product occurs with the position of the same or another sensor at a different height at which reflection at the outer surface occurs. The difference in position measured in the measuring direction then corresponds to a possible skew. Also, when the sensors 10 and 10' as discussed with reference to Fig. 4 are replaced by sensors which emit light of a single frequency and which are arranged movably in their viewing direction, the positions of the sensors at which reflection occurs on the inner wall respectively and the outer wall an inner and outer diameter of the product can be determined from the associated positions of the sensors. Other sensors that can be used for performing a thickness measurement are, for example, laser interference sensors.

Finally, it is noted that automatic control loops through conduit 30 may include adjusting based on a particular glass thickness distribution of: - Feeder temperature and temperature distribution - Droplet temperature distribution - Droplet forming process

- Drop loading process - Mold cooling and residence time in the mold - Design of the “parison” and the preform molds - Blow-out process (B&B process) - Plunger press process (NNPB, PB process) - Plunger cooling process - Front blow-out process - Shape and design of the blowpipe - Optimal lubrication method and lubricant molds - Optimal tool life determination of the molds - Optimal timing setting of the 1.S. machine (glass molding machine) - Optimal temperature settings of the molds and automatic control thereof - Optimal air pressure regulation of the blowing of products - Optimal air pressure regulation for making the parison - Active reheating for an optimal temperature distribution of the glass droplets for an optimal LGV Such variants each fall within the scope of the invention.

Claims (35)

CONCLUSIONS A method for inspecting hollow glass products of lettuce product material, wherein said glass products are manufactured by: a. heating the glass product material; b. forming the heated glass product material into at least one glass product in a production stream; c. cooling the formed glass product; characterized in that inspecting the glass products comprises the following steps: d. successively conveying along a predetermined path the in step b. shaped glass products along at least one sensor, such as a Chromatic Confocal sensor, for measuring a glass thickness, wherein, using the at least one sensor, a plurality of the glass products which are successively transported past the at least one sensor per glass product have at least one glass thickness is measured, step d. between step b. and c. is performed and wherein in step d. of each glass product of the plurality of glass products the rotational position of the glass product about an axial axis of the glass product relative to the at least one sensor is determined, wherein in a step e. the determined glass thicknesses and the associated rotational positions of the plurality of products are processed in combination to obtain information about a lateral glass thickness distribution about an axial axis of a virtual glass product representing the plurality of glass products. Method according to claim 1, characterized in that in step d. the product is transported in a horizontally oriented plane, with the axial axis of the product oriented vertically. Method according to claim 1 or 2, characterized in that the glass thickness distribution indicates relative variations in glass thickness. Method according to claim 1, 2 or 3, characterized in that the glass thickness distribution comprises absolute values of the glass thickness distribution. Method as claimed in any of the foregoing claims, characterized in that with a sufficiently large number of products from a product flow, measurements are performed with the at least one sensor so that these measurements, in combination, form the associated virtual product around the axial axis of the virtual product in such a way. cover that any adjacent distance between the positions of the measurements at the virtual product in the tangential direction is less than a predetermined value. A method according to any one of the preceding claims, characterized in that in step e. the lateral glass thickness distribution is determined in a region of the virtual glass product extending around the axial axis of the virtual glass product. A method according to any one of the preceding claims, characterized in that step e. is performed repeatedly to obtain a lateral glass thickness distribution in a second region of the virtual glass product extending around the axial axis of the virtual glass product with the first and second regions offset from each other in the axial direction. Method according to claim 7, characterized in that step e. is performed repeatedly at least three times respectively to obtain lateral glass thickness distributions in respectively at least three different regions each extending around the axial axis of the virtual glass product and offset from each other in axial direction and which at preferably in combination, cover at least substantially the entire virtual glass product. Method according to at least claims 7 or 8, characterized in that in step d. the sensor is moved in the axial direction to obtain recordings in the different areas. A method according to any one of the preceding claims, characterized in that a plurality of steps b. parallel to each other for producing a plurality of the products in a plurality of production flows in parallel, each having a step b. wherein each lateral glass thickness distribution of a virtual product is obtained from measurements with the at least one sensor on products manufactured in the same production flow. A method according to any one of the preceding claims, characterized in that in step d. wall thicknesses of a glass product that is transported past the at least one sensor are successively determined with the at least one sensor, so that the determined glass thicknesses relate to different positions of a wall of the glass product, which positions are in a direction in which the product in step b. being transported are separated from each other. 12. Method as claimed in any of the foregoing claims, characterized in that the at least one sensor in each case determines a glass thickness of a part of a wall of a glass product, the outer surface of which is directed towards the at least one sensor. A method according to claim 12, characterized in that with the at least one sensor a first distance is measured between the at least one sensor and the outer surface of the part of the wall and a second distance is measured between an inner surface of the part of the wall, the glass thickness of the part of the wall being determined from the first distance and the second distance. A method according to any one of the preceding claims, characterized in that 1n step d. the product is transported in a horizontal plane with the at least one sensor being moved in a vertical direction for measuring the glass product at different heights. A method according to claim 14, characterized in that a glass thickness of the product is determined at different heights. A method according to claim 14 or 15, characterized in that an inclination of the product is determined from the measurement results obtained with the at least one sensor. A method according to any one of the preceding claims, characterized in that in step d. use is made of at least two of the sensors, wherein the at least two sensors are directed opposite and towards each other and wherein the products are transported between the at least two sensors. Method according to claim 17, characterized in that a first glass thickness of a first part of the wall of the glass product is determined with a first sensor of the at least two sensors and that a second sensor of the at least two sensors is used to determine a second glass thickness of a second part of the wall of the glass product is determined wherein the first part and the second part lie on opposite sides of the glass product. A method according to claim 17 or 18, characterized in that with the first sensor a first distance is measured between the first sensor and an outer surface of a first part of the wall which is directed towards the first sensor and that with the second sensor a second distance is measured between the second sensor and an outer surface of a second portion of the wall facing the second sensor, an outer diameter of the glass product being determined from the first and second distances. A method according to claim 17, 18 or 19, characterized in that a third distance is measured with the first sensor between the first sensor and an inner surface of a first part of the wall which is directed towards the first sensor and that with the second sensor measures a fourth distance between the second sensor and an inner surface of a second portion of the wall facing the second sensor, determining an inner diameter of the glass product from the third and fourth distances. 21. A method according to any one of claims 17-20, characterized in that a first distance is measured per product with the first sensor between the first sensor and an outer surface of a first part of the wall that is directed towards the first sensor, and that with the second sensor a second distance is measured between the second sensor and an outer surface of a second part of the wall that is directed towards the second sensor, wherein the first and second distance determined for different products and the associated rotational position have an outer diameter and/ whether skew of the virtual product is determined. A method according to any one of the preceding claims, characterized in that the glass products between step b. and c. are transported along the path on a conveyor belt, in particular each product between step b. and c. is placed on the conveyor belt with a rotational position that varies per product around an axial axis of the product. Method according to claim 22, characterized in that the glass product between step b. and c. is conveyed on the conveyor belt from a product forming device such as a mold in which the glass product 1n step b. 1s formed into a cooling device in which the product in step c. is cooled. A method according to claim 22 or 23, characterized in that the rotational position of each of the glass products of the plurality of glass products is determined about its axial axis on the conveyor belt. A method according to any one of the preceding claims, characterized in that the rotational position of a glass product of the plurality of glass products is determined with the aid of at least one rotational position measuring unit such as a camera. A method according to claim 25, characterized in that the rotational position measuring unit makes an image of the glass product, wherein in the image a predetermined marking on the glass product, such as a seam and/or a dot, is detected for determining the the rotation position. A method according to any one of the preceding claims, characterized in that the inspection also comprises a control wherein at least ah is a step of steps a., b. and c. is adjusted on the basis of measurement results obtained with the aid of at least one sensor and possibly the determined rotational positions, more in particular on the basis of the determined lateral glass thickness distribution of the virtual product. A method for producing and inspecting hollow lettuce products of glass product material, wherein said glass products are manufactured at least by: a. heating the glass product material; b. forming the heated lettuce product material into a glass product in a production stream; c. cooling the formed glass product; characterized in that inspecting the glass products comprises the following steps: d. successively conveying along a predetermined path the in step b. shaped glass products along at least one sensor, such as a Chromatic Confocal sensor, for measuring a glass thickness, wherein at least one glass thickness is measured per glass product from a plurality of the salad products that are successively transported past the at least one sensor , where step d. between step b. and c. is performed and wherein in step d. of each glass product of the plurality of glass products a rotational position of the glass product about an axial axis of the glass product relative to the at least one sensor is determined, wherein in a step e. the determined glass thicknesses and the associated rotational positions of the plurality of products are processed in combination to obtain information about a lateral glass thickness distribution about an axial axis of a virtual glass product representing the plurality of glass products. The method of claim 28, further characterized by any one of claims 2-27. The system for producing and inspecting glass products from glass product material according to the method of the preamble of claim 26, wherein the system comprises: a heating device for performing step a.; a product molding device such as a mold for performing step b.; a cooling device for performing step c.; and a conveying device for conveying glass products formed with the product molding device from the product molding device to the cooling device; at least one sensor for scanning the in step b. molded glass products; and a signal processing unit connected to the at least one sensor for processing signals from the at least one sensor, characterized in that the at least one sensor comprises a sensor, such as a Chromatic Confocal sensor, for measuring a glass thickness wherein the system is further provided with a rotational position measuring unit for obtaining information about a rotational position of a glass product against which measurements are taken with the at least one sensor, wherein, in use, the rotational position about an axial axis of the relevant glass product relative to the at least one sensor is determined, wherein the at least one sensor is arranged such that the formed glass products supplied to the cooling unit, in use, are successively measured with the at least one sensor, wherein the signal processing unit is arranged to measure per measured product on the basis of measurement results from the at least one sensor and information from the rotation mode unit e to determine a lateral glass thickness distribution about an axial axis of a virtual glass product representing the plurality of glass products (so that the system is further adapted to perform the method according to the characteristic of claim 27), wherein in particular the signal processing unit is adapted to act on on the basis of at least a certain glass thickness distribution of at least one virtual product obtained from measurements of products originating from the same production flow to automatically control the production flow, such as adjusting settings of that production flow. 31. A system according to claim 30, characterized in that the system is further designed for performing the method according to any one of claims 2-27. A system according to any one of the preceding claims 30 or 31, characterized in that the system is arranged such that, in use, a plurality of steps b. parallel to each other for producing a plurality of the glass products in a plurality of production flows in parallel, each having a step b. wherein the signal processing unit is arranged such that, in use, the determined glass thicknesses and the associated rotational positions of the plurality of glass products originating from the same production flow are processed in combination to obtain information about a lateral glass thickness distribution about an axial axis of a virtual glass product that represents the multitude of glass products from the same production flow. The system for inspecting glass products of glass product material according to the method according to the preamble of claim 1, wherein the system comprises: at least one sensor for scanning the glass products; and a signal processing unit connected to the at least one sensor for processing signals from the at least one sensor, characterized in that the at least one sensor is a sensor, such as a Chromatic Confocal sensor, for measuring a glass thickness wherein the system further comprises a rotational position unit such as a camera for obtaining information representing a rotational position with respect to the at least one sensor and about an axial axis of a glass product measured with the at least one sensor, wherein, in In use, the at least one sensor is arranged in such a way that the glass products are successively measured with the at least one sensor, wherein the signal processing unit is arranged for per scanned product on the basis of measurement results from the at least one sensor and the associated information from the rotational position. unit a lateral glass thickness distribution around an axial axis of a virtual glass product containing the v a plurality of glass products so that the system is further adapted to perform the method according to the feature of claim 1. A system according to claim 33, characterized in that the system is further designed for performing the method according to any one of claims 2-27. A system according to any one of the preceding claims 33 or 34, characterized in that the system is adapted to be used when a plurality of steps b. parallel to each other for producing a plurality of the glass products in a plurality of production flows in parallel, each having a step b. in use, wherein, in use, the produced glass products are successively measured with the at least one sensor, whereby information about the rotational position of the produced glass products relative to the at least one sensor is obtained per glass product with the rotational position unit, wherein the signal processing unit so day is arranged that, in use, the measurement results of the at least one sensor and the information about the determined glass thicknesses and the associated rotational positions of the plurality of products from the same production flow are processed in combination to obtain information about a lateral glass thickness distribution around an axial axis of a virtual glass product that represents the multiplicity of glass products from the same production flow.