In Measurements under production, the sound velocity or the "sound time of flight" in the pipe is known. For some materials the temperature influence on sound velocity has been researched. It is clear, that the calibration of the wall thickness measurement is essential first.
There are 3 methods known.
At the cooled pipe the wall thickness is to be determined (with mechanical measurements), further a correction factor needs to calculated and the original ultrasonic measurement result to be recalculated. The manual correction has to be done frequently since the pipe temperature has changed and the sound velocity is different.
2. Temperature equation
During the pipe cooling process there is a temperature profile in pipe axis direction. This temperature curve can be, upto the measurement location, determined by mathematics. The mathematics are based on physics models for the cooling procedure. This method is rarely applied, since the mass temperature is not known exactly as a starting point and the cooling model is not always matched with the real cooling devices.
3. Pipe In-Line weight/meter measurements
The safest and simplest method is the calibration with the help of the mass throughput value. A mass throughput measurement unit is a precondition for this function. With the measured "ultrasound time of flight", a relative pipe cross section is determined. The absolute cross section wall thickness is measured by the mass throughput unit. The real sound velocity can be determined and the real wall thickness is calculated.
grafical display explains the Principle.
The preconditions for that calculation are, that the outer diameter and the material density are known. This calibration procedure is automatically by the system done and continuously.
There is one problem that can appears with thick wall pipes: temperature differences at the pipe circumference. These increase during the cooling . But can also rise, depending on extruder die conditions, directly behind the die.
The reason for this are wall thickness and temperature differences in the die, different cooling properties at upper and lower pipe sections and the adjustment of the cooling sprayers in the tank. Materials like PE can create a temperature difference also by "sacking". All these conditions are very complex and cannot be predetermined.
To handle such problems, a few suppliers offer a software function , called "Segmenting Calibration Correction" that is based on manually determined corrections and is entered into the system. Because of the instability of those temperature conditions, the manual procedures have to be done frequently. Through this it lost the purpose of automation and does not convince the users to applying this method.To overcome this problem, others tried to measure close to the die.
Without considering method 2, there is only the choice between either Manual or mass throughput methods. The individual case decides what is the better method. If a mass-throughput unit is not essentially necessary, or not accurate because of too little material throughput, the user often chooses manual calibration. Thus the installed hardware is cheaper and more reliable.
Generally, it should be the user's accuracy demands that assumes the decision. The measurement error is mostly 0.3%/ degree Celsius, thus with constant process and pipe temperature conditions, manual calibration can be recommended. The PC's receipt handling supports the manual method, by recalling the same calibration during a later machine start-up procedure. This has been practiced by a few extruder line suppliers for a number of years.
Users often request only a die centering support as the key function of the system. In that case, an absolute accuracy demand is not necessary.
Also new is a system that compensates with software the influence of cooling water on the calibration. Because of the integrated diameter measurement, the system refers to the water temperature anyway, thus there are no further costs borne.
The accuracy statements of In-Line ultrasonic measurement systems can only be made when facing all the influences mentioned.
The resolution of wall thickness measurement should be 1/100 mm and isconfirmed by most suppliers. The accuracy gained practically is dependant on individual conditions and covers a range from 0,02 mm - 0,10 mm.
Rolf Diederichs 18.Dec. 1995, firstname.lastname@example.org