Quality Control (QC) activities in a glass container plant – Overview.

In the glass container manufacturing process everything starts with the reception and storing of the different raw materials.

As a first step, in order to make it possible to control the quality of the incoming raw materials, it is necessary that the glass manufacturer establishes a technical specification for each raw material that is used to produce glass. Cullet – recycled glass from external or internal origin – should be treated like any other raw material; therefore a technical specification should be developed as well for cullet.

Typically the specification defines optimum and normal reject limits for key parameters specific of each raw material: chemical composition and grain size.

Percentages for the concentration of main oxides, desired an undesired (contaminations) are established.

The goal it is not having a detailed chemical composition description of the raw material but rather focus on the oxides that are considered to be important. That is, those that can affect the outcome of the melting process, in the end the glass quality.

Grain size is another key parameter to control in a raw material. Economical and quality concerns are associated with it.

For cullet are established limits for contaminants: foreign materials (ceramic, organic) and colours (especially important if producing in flint glass colour). If the cullet used in the plant has an external source – bought from a supplier, opposed to an internal source, coming from internal glass rejections – it is of special importance the control of its quality. The uncertainty is greater.

Usually this is done by operating a formal compliance certificate system.

Ideally these technical specifications should be part of the commercial contract between glass manufacturer and the raw material supplier. As a minimum they should be of proven knowledge by the supplier.

Once established technical specifications for the raw materials, physical and chemical checking of incoming raw materials on receipt can be undertaken.

These controls are undertaken either in a local laboratory in the glass plant or subcontracted to an outside accredited lab. The latter has been of preference – as long as the response time is adequate – to decrease costs.

Nowadays these checking’s have largely been replaced by the provision of supplier certificates of conformance, by the access of the glass manufacturer to the raw material supplier’s process control information complemented by audits of the raw material supply locations. The purpose of such audits is to ensure familiarity with the process control methods and standards used by the supplier and to confirm their ability to adequately control the raw material within the glass manufacturer technical specification. These audits are part of the organization supplier audit plan.

Depending of the type of system that is in place to assure the quality of the raw materials received, the incoming check can vary from a confirmation of the nature of the material, the quantity being delivered and a validation of the certificate of conformance.

If a chemical and/or physical checking is required small samples can be taken and retained for laboratory examination. This may be done either on a random basis or regularly in accordance with an inspection plan.

A typical check that is done locally at the plant at the moment of raw materials receiving is the moisture content of the sand. It is necessary to control this to tight limits and if these are exceeded then corrections have to be made to the batch composition to compensate at the time of mixing. Also there are obvious economic implications in receiving a truck load of sand with excess of moisture.

Other materials that are used – or aid – in the container glass manufacturing process are also submitted to some kind of quality check. Among these we can refer: moulds, packaging material and coating materials.

In what concerns moulds, typically compliance certificates are required together with data from the checks carried out from the supplier. Alternatively, or sometimes additionally, the glass manufacturer samples the incoming mould items and carries out his measurements against a defined plan and the mould specification. Again, to decrease costs the latter option has becoming less used by the glass manufacturers.

The typical approach concerning other materials is the monitoring of the suppliers through certificates of conformance systems and supplier audits whenever appropriate and integrated in the supplier audit plan of the organization.

As in other cases the observed tendency is not to exert the control at plant reception using plant resources but rather to have tools that allow supplier monitoring continuously or at prescribed occasions.

The result of the batching, mixing and melting of the raw materials is glass. Independently of its end usage – that is, if it is going to be moulded into a bottle or jar – there are a couple of controls that should be performed to assess the quality of the molten glass.

The density test is a quick and easy way to indirectly detected unexpected changes in the glass composition due to mistakes in the raw materials batch and mixing process. The test involves the determination of the density of a glass specimen and comparison with previous results, monitoring significant deviations.

Homogeneity test indicates the presence of any inhomogeneous glass which shows up as coloured streaks under polarized light. The degree of inhomogeneity – permanent induced stress - can be quantified if necessary by examination under a petrological microscope. These inhomogeneities are known as cords.

Bubble and seed – gaseous inclusions in the glass, potentially stress concentrators – count, give an indication regarding the quality of the melting process. The result of the count is compared against the glass manufacturer established rejection limits.

Stones are solid un-molten material embedded in the glass that acts as stress concentrator - induces permanent stress in the glass surface. The total amount of stone contamination – stone count - is compared with established rejection limits. In parallel the origin of the stone is determined by visual comparison using reference literature.

If producing coloured glass it is fundamental to control the glass colour. Glass colour can be measured using a spectrophotometer to determine: dominant wavelength, purity and brightness. Again, the result achieved is compared against the established colour standard.

Manual QC starts at the Hot End. The Hot End comprises basically glass conditioning and container forming.

Immediately after the glass containers leave the forming machine. The Hot End operator should perform a visual and dimensional verification.

Samples from each individual mould cavity on the forming machine – a stratified sample -are taken off at regular predetermined intervals and allowed to cool for inspection in a light box. They are inspected visually and gauged either with fixed go-no go gauges or by a conventional S.P.C. measurement system where this method of control is preferred or more appropriate.

S.P.C. system is preferred since it is of preventive and not of reactive nature.

The glass weight is another parameter which is regularly monitored at the Hot End.

In fact – although sometimes undervalued – the detection of defects at this stage it is very important. Here it is possible to act immediately on the machine for correction. There is no significant delay involved between the moment of defect detection and the action for correction.

One of the idiosyncrasies of this manufacturing process is that – simply putted – the latter part of the process is almost entirely dedicated to QC activities. We call this area the Cold End.

In the Cold End, immediately after the containers leave the annealing lehr and after the application of the Cold End Coating, they are submitted to a second QC, now performed by a Cold End (Quality) operator. Again, we have an attributes inspection for stratified sampling. This inspection is performed at regular intervals in accordance with an inspection plan.

The controls are similar to those performed on the Hot End but with the advantage that at this stage the container is annealed and coated (hot end and cold end coated). That is, the container is finished.

The aim of this control is to detect and eliminate defects and give important information to the hot end for the correction of defects. In fact one of the key aspects for the success of a Quality Control System in a glass container plant is precisely the communication between Hot End and Cold End operators.

The results of these controls are recorded either electronically or on paper documents (each defect has a specific code which facilitates the communication and recording).

After this manual inspection the containers are placed in single lines and passed by various automatic inspection machines. This type of inspection should be the most effective. Every single container produced is inspected as opposed to the other inspections that are based on sampling plans.

Associated with the inspection machines we have a crucial tool used for defect elimination. That is the mould number code engraved in the containers as a dot or alphanumeric code. This code allows the automatic – and reliable – rejection of specific mould numbers by the inspection machines.

Important as well is to assess regularly the effectiveness of the inspection machines. The monitoring of such condition is done with the use of challenging samples. These are samples of containers with specific defects used to verify if the machines are performing as expected. However as simple as this may seems, rules and procedures must be followed in order to assure that only the best and most adequate samples are used. Otherwise it will be a wasteful and ineffective operation.

Inspection machines can be grouped – generically - in 3 big groups depending of the type of inspection they perform.

Sidewall inspection machines, inspect for defects – visual and some dimensional – in the body of the glass container.

Base and finish inspection machines, inspect – as it says – the base and finish areas of the containers.

These two groups of inspection machines use cameras for performing their inspection.

A third type of inspection machine - multi inspection machine - uses light reflection principles to detect and automatically reject defects. These machines mechanical in nature – several inspection stations, containers indexed by a star wheel – also check selected dimensions such as the bore of a bottle or the waviness of the sealing surface of a jar and glass thickness, amongst other possible detections.

These three groups of machines are installed sequentially in the production line and sometimes subdivided in equal parallel branches on the Cold End.

After the inspection machines in the cold end usually we have a human visual inspection as the containers pass in front of an on line light box (light screen).

The cold end operator inspects at regular intervals – during a short period of time – the containers that are passing in the line, in front of the light box screen.

The objective of this inspection is often source of much misunderstanding.

Here we want to assess the visual quality of the containers and therefore check the effectiveness of the inspection machines regarding this particular aspect. It is also another point to collect information regarding the overall visual quality of the containers being produced and to forward this information to the hot end.

Only in very particular situations – if the inspection machines cannot properly visually inspect the container due to its geometry or engravings (limitations to inspection) – this inspection can be used 100% of the time as a backup visual inspection but with obvious limitations.

Care should be taken to frequently rotate the operators that perform this inspection due to visual fatigue.

A final statistical check, or audit, is done as the containers are being, or when they have been, assembled on pallets for shrink-wrapping. Random samples are made accordingly with a sampling plan and the containers are visually inspected.

For the purpose of these checks, imperfections are often classified into three main groups: critical, major and minor defects. AQL levels are established in the organizations for each type of defect and the result of this inspection is compared against that standard.

This final inspection is used to formally approve the batch in production. We should be able to validate all the controls that are upstream.

Once the containers have been assembled onto pallets, each pallet is assigned a sequence coded label. The information in the label can be cross-referenced to all control inspection records. The batch sequence code number on the pallet label therefore provides traceability to the rest of the manufacturing chain and the other on-line control points.

Where traceability of individual containers is required this is achieved by marking each individual bottle at the time of production usually with an ink jet mark. Although very often a customer requirement, this marking is very useful for the producer as well. Especially when there is a need to investigate causes for specific failure or defect.

In parallel in the Quality Laboratory a series of parameters are tested in order to determine the container resistance and compliance with specifications.

Typical tests encompass: thermal shock, internal pressure, impact test, vertical load, annealing level, hot end coating level and cold end coating level.

Also, quantified dimensional measurements – in the production line the assessments are qualitative: go, no-go! – are performed and calculated simple averages and ranges. These results are handled statistically to detect trends and trigger corrective actions at the earliest possible moment.

The frequency of testing depends on several factors. These include the method of test, the magnitude of the control values, and how quickly a feature could go to out of control.

Resources are scarce and it is the task of the Quality Control Manager to determine were efforts must be made and where they can be relieved. The outcome of this management is documented in the Quality Control Plan for the specific container.

Samples from each mould are checked when making a routine control test. Apart from the annealing and cold end coating test, where the samples are selected based on their position in the annealing tunnel.

A failure during a routine process control test initiates an immediate retest of a larger sample (usually 3 to 6 containers from the mould cavity concerned) to establish whether or not there is a downward quality trend.

When sub-standard ware is detected, all the containers from the suspect mould cavities are rejected until there is a successful test. In addition, all the ware packed as good from the suspect mould cavities since the last successful test is regarded as suspect and reinvestigated so that all substandard ware can be rejected.

In the end – regardless of how cliché this might sound – people make the difference between an effective and an ineffective quality control system.

Key aspects encompass their attitude when performing the prescribed tasks and decision making. There is a significant connection between knowledge, skill and good decision making. Correct decisions are best made by the person performing the task.

The operator’s key role is to follow standard procedures relative to process control and document problems and suggestions.

Understanding customer needs and interacting with the customer via plant visits and cross functional teams is absolutely critical. Observe their filling lines and get direct feedback on the quality of the glass containers.

Every opportunity to interact with customers should be taken on a regular basis.

Stress analysis simulation to determine glass bottles failure

That the glass thickness and distributions are paramount for a glass container is a statement that comes without question. But most probably we’ve found situations where a container is within the specifications and still breakages occur.

This article will explore a real case situation, where it was requested to Empakglass to evaluate the design of a 750ml bottle, which has a consistent glass thickness and distribution and nevertheless, breakages did occur when internal pressure load was applied.

For the simulations presented on this article, the Empakglass Forming Software was used (Empaktor Suite).

The method applied by Empakglass consisted on the following:

• The client provided the current bottles in order to make cuts to determine the glass distribution;

• The client provided the glass composition and gob temperature;

• The current client’s parison design was simulated using Empakglass’s Forming Software and based on the physical properties of the used glass and the client’s IS machine timing;

• A comparison between the real current glass thickness distribution and the one achieved by the simulation was performed. Although there was still a deviation on the settle wave line position It was verified that the achieved simulation profile was quite similar to the actual bottles,

• Using the achieved wall thickness profile, a stress analysis for internal pressure load was performed on the model to assess the resistance of the current bottle design.

Following the principle that failure only occurs when generated stresses exceeds the glass strength, all the maximum surface stress values achieved on the above shown Internal pressure simulations are below the glass strength limits, meaningthis design is theoretically acceptable for beverage usage with a Carbonatation of 4Vol-CO and for pasteurization at 70°C.

Nevertheless and although the simulation results theoretically pass the bottle design, a pattern between the breakage origin and the simulation can be determined.

The sidewall area of a bottle is one of the identified critical areas of a container in what regards its resistance to internal pressure. The type of breakage shown on the above pictures has undoubtedly a sidewall origin (settle wave area).

On this particular bottle the settle wave line (associated with lower wall thickness values) is placed on the contact area of the bottle.

Due to the production process limitations (B&B), to have an absolute control on the thickness values achieved on the settle wave area during production is statistically low and therefore will generate bottles that will fail under the tests performed by QC.

The parison design is without any doubt one of the main tools where a glass producer can try to compensate the lack of glass on the settle wave area.

In order to prove the relation between the wall thickness values and the internal pressure resistance, a new parison was developed and again simulated. This parison used the same glass weight. As well, the new parison was also shorter in order to compensate the fast rundown seen on the forming simulations

The forming simulation results show that on the new developed parison, a higher wall thickness on the settle wave area has allowed to shift the weakest area on the container from a glass contact area to a non-contact area on the shoulder.

By increasing 0.55mm on the settle wave area (see figure 4), there was an increase of extra 15% resistance to the internal pressure load even by keeping the same glass weight.

Also, complementing the proper mould design and having in mind that it is critical to protect the surface areas against surface abuse, especially in the contact points, where the probability is higher of having any kind of abuse.

By achieving this is by having good Hot End Coating (HEC) and Cold End Coating (CEC) application. The combination of these two coatings protects the glass surface against friction damage. This damage typically occurs when a blunt hard object slid across a glass surface – for example – two bottles slid against each other in manufacturing line and filling line. Resistance to scratching must be achieved so as to keep the inherent bottle strength high.

For HEC – for carbonated bottles – the level should between 30 to 40 CTU and 25 CTU should be the absolute minimum. It should be guaranteed an even distribution of the coating throughout the glass surface.

For CEC the recommend values are between 9 to 12 degrees of slip angle (determined with AGR Tilt Table). Again, it should be guaranteed an even distribution of the coating throughout the glass surface.

So, summoning this article:

• We can have a bottle with a reasonable glass thickness and distribution;

• The glass distribution is within Quality AQL’s for that particular container;

• Nevertheless, the container fails under internal pressure loads;

• By developing a new parison and keeping the same glass weight on this container, it is possible to shift the weakest area on the container from the labeling area (glass contact area) to the shoulder (non contact area).

When less glass wall thickness becomes a solution

It isn’t the most common issue to find in a glass plant. However for niche markets - where the same bottle or jar is produced in different colors, this becomes something to have in consideration technical wise. The end results will be significantly different and commercial wise has implications when budgeting the costs of a production.

On this article we will show a practical case, where for the same mold design, very different results were achieved in terms of glass distribution. Glass distribution is, without debate, the most important contributor for the integrity of any glass container. The simulations here shown were made using Empakglass’s Forming Software.

Glass color in terms of physical properties directly influences the thermal transmission. This is common sense: we know how sun light crosses more or less along a lighter or darker window.

On a IS machine, the reheating time, which is part of the production process is directly influenced by the thermal conductivity of glass.

The reheating is the period of time between the end of the parison transfer and the start of the final blow. During this time, the parison tends for temperature equalization (reheat) and gravity stretches it. Excessive reheating on the blank mold side allows the parison to sag and on the blow side to run, and the two effects have to be counterbalanced. Stretching and cooling of the parison can be helped by the use of overhead cooling over the blow mold.

The speed of invert affects the glass distribution of the finished bottle: if it is too slow, the parison will sag backward due to gravity; if too fast, the parison is thrown forward by centrifugal force. The speed must be varied to suit the weight, viscosity and shape of parison.

Coming back into the glass physical properties. When comparing a darker glass versus a lighter color glass, the time to equalize the parison temperature will take longer on the first case. Therefore the sag/run effects will be less for the same actual time cycle comes to place.

For this particular case study example, the blank molds were developed for a Empakglass client flint glass. The consistency of the Forming Simulation results was confirmed in the actual bottle production.

(images from left to right: Glass thickness simulation ;3D Rendering; Actual bottle)

When the client decided to use the same mold set with “black” color glass, a visual defect appeared around the heel area. A “wave” shape, with a “cold” appearance and the heel region wasn’t completely formed.

(images from left to right: Wave shape defect and heel region not completely formedly formed)

Another defect that was detected were “dropped bottoms”. These did not appear in the flint bottles produced with the same mold set.

This defect is consistent with temperatures higher than the softening point (log 7.65).

Curiously, although this might have two different reasons (cold appearance and dropped bottoms), the root cause is just one – the thermal conductivity of black glass.

By using Empakglass’s Forming Software and simulating in black and comparing with flint, :

1. the glass outside surface temperature on the moment where the final blow starts has an average of 40 degrees lower in black color when compared with flint => cold appearance on the out surface and "wave" look;

2. The isotherm areas above 1080ºC in black glass are bigger than in flint, mainly around the shoulder and mostly on the bottom, where the glass thickness is higher => longer to equalize temperature and therefore the occurrence of dropped bottoms already on the annealing lehr;

This is already showing that the black glass reheat time is too slow when compared with the flint (more difficult to have thermal transmission).

(image above shows client's Flint Glass thermal profile)

(image above shows client's Black Glass thermal profile - forming simulation)

In practice, this means that in “black” when the final blow is applied, the glass is still cold on the outer surface and will not form completely on the blow side.

This confirms as well the practice from Production Staff when it states that with darker colors and the higher the thickness the worse it will be in terms of reheat time.

The excessive glass thickness on the bottom together with the reheat time difference between flint and black is too big (almost double). In black glass we can see a deformation on the bottle heel, that doesn’t appear in flint, already after the bottles entered the annealing lehr.

(image on the left shows the temperature/glass thickness profile using black glass physical properties and image on the right the actual bottle)

In black, due to the thermal conductivity, when comparing with flint glass it is required to reduce the bottle bottom thickness to get a faster reheat time on the glass. This can be achieved by developing a different parison design, with higher overcapacity.

By doing so, both the “wave” look appearance on the heel and also the dropped bottoms effect can be solved.

(image above shows the comparison between the glass distribution for flint (in blue), with a blank mould with 25% overcapacity and a 40% overcapacity blank mould for black glass (in black) keeping the same weight)

Therefore this is one of the actual practical cases, where by keeping the same weight and increasing the overcapacity, a lower thickness on the bottom is achieved, in order to solve the issues presented above.