Subscribe to our newsletter to receive the latest news and events from TWI:

Subscribe >
Skip to content

IR Welding System using Statistically Designed Experiments

   

Assessment of a New, Completely Controlled, IR Welding System using Statistically Designed Experiments

 

Mike Troughton, TWI, Cambridge, UK
Jörg Wermelinger, Georg Fischer Piping Systems Ltd, Schaffhausen, Switzerland

Paper presented at ANTEC 2000, Orlando, USA, 7-11 May 2000

Abstract

Infrared welding has been used successfully for many years for joining plastics pipes in industrial applications. In order to increase the productivity of this technique, a new generation of machines has been developed that can reduce the weld cycle times by up to 45% compared with standard IR machines. This paper describes how a statistically designed experimental programme has been used to assess the performance of these new IR welding machine in the most cost-effective manner. Results have shown that low temperature welding factors achieved are within 4% of the values obtained on standard IR machines. 

Introduction

For high purity liquid distribution systems, such as those found in the semiconductor, pharmaceutical and chemical process industries, IR welding has a number of advantages over traditional butt fusion (hot plate) welding for joining plastics pipes. These include:

  • Non-contact heating, reducing the risk of contamination at the joint
  • Reduction in process variables, such as hydraulic pressure
  • Reduction in size of weld beads, increasing flow and reducing the risk of micro-contamination at the internal bead.

With these advantages, IR welding has been used very successfully for joining industrial piping systems made from polypropylene (PP) and polyvinylidenefluoride (PVDF) since its introduction in the early 1990s [1] . However, as with all current pipe welding techniques, the time to make the joint using the IR technique, although shorter than butt fusion welding, was still an area for improvement. For this reason a new, completely controlled, IR system, has recently been developed which reduces the fusion time by over 50% compared with conventional butt fusion welding. This report details a study, carried out to evaluate the consistency of mechanical properties of joints made using this new technique and compare them to the properties of joints made on standard IR welding equipment.

Principles of completely controlled IR welding technology

In conventional butt fusion and IR welding, the cooling time, which accounts for the majority of the overall cycle time, is fixed for any size of pipe. For the new completely controlled IR welding technology, the cooling time is variable and depends upon the time for the weld to reach a predetermined temperature, the material release temperature (MRT).

The requirements for the MRT are:

  • The weld material must be solid
  • The weld must have a strength which would not allow it to be damaged in the process of removing the pipe from the clamps
  • The MRT must be below the maximum allowable service temperature.

For PP and PVDF, the MRTs have been established as 90°C and 70°C, respectively.

In IR welding, the time to reach the MRT is governed by several factors:

  • The energy absorbed by the material during the heating stage
  • The MRT itself
  • The environment (temperature, air flow)
  • The wall thickness of the component being joined
  • The thermal conductivity of the material
  • The specific heat capacity of the material.

Of the above factors, the energy absorbed can be calculated, the MRT is defined, the wall thickness can be measured, and the thermal conductivity and specific heat capacity are known. So, the only remaining variable that has an influence on the cooling time is the environment. In the completely controlled IR welding technology, the temperature of the environment is monitored, allowing the time to reach the MRT to be calculated.

IR welding machines

Two machines, incorporating the completely controlled IR welding technology, have been developed: a small machine, for welding pipes with nominal outside diameters, dn, ranging from 20mm to 63mm; and a large machine, for welding pipes of dn between 63mm and 225mm.

Small machine

This welding machine is fitted with a sensor for monitoring the temperature of the environment. The time needed to reach the MRT was determined experimentally for each pipe size and material when welding at different ambient temperatures (between 5 and 40°C). This information is programmed into the microprocessor in the welding machine, so that it can then calculate the cooling time according to the measured ambient temperature.

Large machine

Since the pipes to be welded on this machine have thicker walls, and therefore longer weld cycle times, two additional modifications were made to this equipment. Firstly, an infrared sensor was installed to measure the temperature of the weld directly, and secondly, an external cooling fan was added to assist in reaching the MRT faster. A schematic of the equipment is shown in Fig.1.
Fig.1. Schematic of large IR welding machine
Fig.1. Schematic of large IR welding machine

Since the infrared sensor can only measure the temperature of the external bead, the relationship between this measured temperature and the actual temperature at the centre of the weld was calculated theoretically using the relevant material properties, for different ambient temperatures, wall thicknesses and materials. These relationships were verified by carrying out welding trials with a thermocouple positioned at the centre of the weld. The size of the thermocouple was extremely small so as not to interfere with measuring the actual temperature of the surrounding plastic material. A typical relationship between the temperature at the centre of the weld (as measured by the thermocouple) and the temperature of the external bead (as measured by the IR sensor) is shown in Fig.2.

Fig.2. Relationship between external bead temperature and mid-thickness weld temperature during the cooling stage
Fig.2. Relationship between external bead temperature and mid-thickness weld temperature during the cooling stage

This information is programmed into the microprocessor in the welding machine, so that it can calculate the cooling time according to the temperature measured by the IR sensor.

The resulting reductions in weld cycle time using these new, completely controlled, IR welding machines are over 50% compared with conventional butt fusion welding and up to 45% compared with conventional IR welding.

Experimental details

The objective of this study was to prove that the mechanical properties of the welds produced on the new completely controlled IR welding machines are as good, and of the same consistent quality, as those produced on standard IR machines, over the following range of variables:

  • Pipe/fitting material: PP and two grades of PVDF (Standard and HP)
  • Pipe/fitting size: 20-63mm for the small machine, 63-225mm for the large machine
  • Pipe/fitting combination: pipe-to-pipe, pipe-to-fitting, and fitting-to-fitting
  • Ambient temperature: 5 - 40°C
  • Forced cooling: On/off (for the large IR machine only)

In order to do this in the most cost-effective manner, i.e. using the minimum number of experimental trials, a Taguchi experimental design method was used. Since it was not possible to compare forced cooling on both machines, and it was not relevant to directly compare PP with PVDF, the welding trials were divided into four separate orthogonal experimental arrays, as shown in Table 1.


Table 1 Parameters for designed experiments

Array No.DescriptionVariableLevels
A PP pipe on small machine Pipe/fitting diameter, mm 20, 25, 32, 40, 50, 63
Pipe/fitting combination Pipe/pipe, pipe/fitting, fitting/fitting
Ambient temperature,°C 5, 23, 40
B PVDF pipe on small machine Pipe/fitting diameter, mm 20, 40, 63
Pipe/fitting combination PVDF Standard pipe/PVDF Standard pipe,
PVDF Standard pipe/PVDF HP pipe,
PVDF HP pipe/PVDF HP pipe,
PVDF Standard pipe/fitting,
PVDF HP pipe/fitting,
Fitting/fitting
Ambient temperature,°C 5, 23, 40
C PP pipe on large machine Pipe/fitting diameter, mm 63, 90, 125, 160, 200, 225
Pipe/fitting combination Pipe/pipe, pipe/fitting, fitting/fitting
Ambient temperature,°C 5, 23, 40
Forced cooling On, Off
D PVDF pipe on large machine Pipe/fitting diameter, mm 63, 140, 225
Pipe/fitting combination PVDF Standard pipe/PVDF Standard pipe,
PVDF Standard pipe/PVDF HP pipe,
PVDF HP pipe/PVDF HP pipe,
PVDF Standard pipe/fitting,
PVDF HP pipe/fitting,
Fitting/fitting
Ambient temperature,°C 5, 23, 40
Forced cooling On, Off

Each experimental array consisted of 36 (18 x 2) experimental runs. The reason for replicating each experiment was to give a more accurate evaluation of the variability of the process.

All welds were assessed using the low temperature tensile test, according to the draft CEN standard, [2] except that the thickness of the specimen was taken as the minimum thickness of the specimen either side of the weld, minus any wall offset. This was done to give a more realistic value of the cross-sectional area at the weld, and therefore a better comparison of the tensile strengths of the joints. The test specimen geometry is given in Fig.3. For all specimens, the weld beads were left intact and the test temperature was -40°C.

spmjtmay2000f3.gif
dnLb1b
20 160 10 5
25 160 12 6
32 160 12 6
40 160 12 6
50 160 12 6
63 160 12 6
75 160 14 7
90 160 14 7
110 160 14 7
125 180 16 8
140 180 16 8
160 180 16 8
200 180 18 9
225 180 20 10
(all dimensions in millimetres)

Fig.3. Geometry and dimensions of low temperature tensile test specimen

Two test specimens were cut from each weld, one from the top of the weld and one from the bottom. In addition, a number of specimens were cut from the parent pipes, in order to calculate a low temperature tensile welding factor, f t, defined as:

spmjtmay2000e1.gif

All welding trials and mechanical tests were carried out at Georg Fischer Piping Systems Ltd, Schaffhausen, Switzerland. The analysis of the results was carried out at TWI, using the analysis of variance (ANOVA) technique.

Results and discussion

PP

The results from the ANOVA calculations indicated that for the IR-63 Plus machine, none of the factors, in the ranges studied, had a significant effect on the measured low temperature tensile welding factor, f t. The predicted values of f t, from the ANOVA analysis, of welds made on the IR-63 Plus machine are compared with the measured values from welds made on the standard IR-63 machine in Table 2. This shows that, on average, the values were 4% higher for the IR-63 Plus machine and almost identical to those of the parent pipe material.

Table 2 Comparison of measured values of f t for PP pipes/fittings welded at 23°C on a small standard IR welding machine with predicted values for pipes/fittings welded at 23°C on a small, completely controlled, IR welding machine.

Pipe Diameter, mmAverage low temperature tensile welding factor, f t
StandardCompletely controlled
20 0.96 0.98
25 0.95 0.98
32 0.91 1.00
40 0.92 1.00
50 0.95 0.99
63 0.99 1.02
Average 0.95 ± 0.03 0.99 ± 0.02

For the IR-225 Plus machine, the analysis suggested that both the pipe diameter and forced cooling had a small but real effect on the average low temperature tensile welding factor. The effect of forced cooling was to decrease the average welding factor by around 3%. For the pipe diameter, as can be seen in Table 3, there appeared to be a steady decrease in average welding factor of around 4% with increasing pipe diameter from 63mm to 160mm, and then an increase for the larger pipe sizes. However, the overall values were, on average, the same as for the IR-225 machine, showing properties equal to the parent pipe.

Table 3 Comparison of measured values of f t for PP pipes/fittings welded at 23°C on a large standard IR welding machine with predicted values for pipes/fittings welded at 23°C on a large, completely controlled, IR welding machine.

Pipe Diameter, mmAverage low temperature tensile welding factor, f t
StandardCompletely controlled
63 0.98 1.02
75 1.02 -
90 1.03 1.01
110 1.02 -
125 1.02 0.99
140 1.03 -
160 1.01 0.98
200 0.99 1.04
225 0.98 1.00
Average 1.01 ± 0.02 1.01 ± 0.02

The analysis also showed that the error due to repetitions was always less than the experimental error, which indicates that the consistency of the welds produced on both machines was very good.

PVDF

The results from the ANOVA analysis indicated that, for the IR-63 Plus machine, the pipe diameter had a definite effect on the low temperature tensile welding factor, with the welds in 20mm diameter pipes/fittings having a welding factor around 7% higher than for welds in 40mm or 63mm diameter pipes/fittings (see Table 4). They also showed that, as for the PP pipes, the values for f t were, on average, 4% higher for the IR-63 Plus machine. 

Table 4 Comparison of measured values of f t for PVDF pipes/fittings welded at 23°C on a small standard IR welding machine with predicted values for pipes/fittings welded at 23°C on a small, completely controlled, IR welding machine.

Pipe Diameter, mmAverage low temperature tensile welding factor, f t
StandardCompletely controlled
20 0.91 0.98
25 0.88 -
32 0.91 -
40 0.87 0.91
50 0.90 -
63 0.90 0.92
Average 0.90 ± 0.02 0.94 ± 0.04

For the IR-225 Plus machine, the analysis suggested that both the pipe/fitting combination and the ambient temperature had a definite effect on the low temperature tensile welding factor. The effect of ambient temperature was to decrease the welding factor by around 5% on going from 5°C to 40°C. The effect of pipe/fitting combination is given in Table 5 and suggested that the values of f t for the pipe-to-fitting combinations were around 5% lower than for pipe-to-pipe or fitting-to-fitting combinations.

Table 5 Values of average low temperature tensile welding factor for various pipe/fitting combinations for PVDF pipes/fittings welded on a large, completely controlled, IR welding machine.

Pipe/fitting combinationAverage low temperature tensile welding factor, f t
PVDF Standard pipe/PVDF Standard pipe 0.89
PVDF Standard pipe/PVDF HP pipe 0.86
PVDF HP pipe/PVDF HP pipe 0.85
PVDF Standard pipe/fitting 0.82
PVDF HP pipe/fitting 0.82
Fitting/fitting 0.85

The predicted values of f t from the ANOVA analysis of welds made in PVDF pipes on the IR-225 Plus machine are compared with the measured values from welds made on the standard IR-225 machine in Table 6. This shows that, on average, the values were 4% lower for welds made on the IR-Plus machine.

Table 6 Comparison of measured values of f t for PVDF pipes/fittings welded at 23°C on a large standard IR welding machine with predicted values for pipes/fittings welded at 23°C on a large, completely controlled, IR welding machine.

Pipe Diameter, mmAverage low temperature tensile welding factor, f t
StandardCompletely controlled
63 0.90 0.84
90 0.89 -
110 0.85 -
125 0.90 -
140 0.90 0.85
160 0.88 -
200 0.88 -
225 0.88 0.86
Average 0.89 ± 0.02 0.85 ± 0.01

The results also suggested that welding 63mm pipes/fittings on the IR-63 Plus machine would be expected to produce joints with welding factors around 6% higher than if they were welded on the IR-225 Plus machine. In addition, as for the welds made in PP pipes, the analysis showed that the error due to repetitions was always less than the experimental error, indicating good consistency in the quality of the welds.

Conclusions

The new IR Plus TM machines, developed by Georg Fischer, can reduce the overall welding times by up to 45% compared with the standard IR welding machines, and by over 50% compared with conventional butt fusion welding.

A statistically designed set of experiments has been used to assess the performance of these new machines in the most cost-effective manner, and has shown that the low temperature welding factors achieved are within 4% of the values obtained on standard IR welding equipment. In addition, the analysis has shown that the consistency of the welds is well within the experimental error.

References

  1. Taylor N S, Klaiber F and Wermelinger J: 'IR welding afforded promising future', TWI Bulletin 4 July/August 1993 86-89.
  2. PrEN 12814-6: 'Testing of welded joints of thermoplastics semi-finished products - Part 6: Low temperature tensile test', 1999.

Mike Troughton graduated in Physics at the University of Leeds, and continued there to obtain a PhD. He worked in industry for seven years before joining TWI in 1993. He works in the Advanced Materials and Processes Department, where he is responsible for the development of techniques for the welding of plastics and also for determining the structural integrity of plastics welds.

Jörg Wermelinger studied mechanical and plastics engineering and graduated as a Mechanical Engineer in 1984. After joining Georg Fischer Piping Systems, Schaffhausen, Switzerland, he was in charge of developments in industrial pressure piping applications. Since 1990 he has been Manager of Jointing Technology, R&D, responsible for Bead and Crevice Free (BCF) and infrared welding machines.

For more information please email:


contactus@twi.co.uk