Introduction to dealing with the measurement of large components

I am Dimitri Favvas. an expert by my own definition in measuring and fabricating large components. I am basing this claim on my long time involvement as a quality control inspector and field representative mainly in the domain of hydraulic turbines.

I have had the opportunity to observe and study the making of the above mentioned components and the reliability of inspection reports, identify problems and propose solutions. Some of my proposals were adopted, I never had a chance to prove some others and more were formed in my mind after my retirement. This is why I am here. To find people interested on my ideas and implement them.

In particular I want to introduce new ways in fabricating with more precision and repairing as well in the field, Francis type runners without the use of templates, new ways in measuring geometric tolerances, spigot diameters and aligning of bores (i.g. bottom ring to head cover). I also want to put an emphasis on setup for machining taking into account the flexibility due to weight of large pieces, a practice that is not new but unfortunately not followed.

More details are provided below, although I am keeping key details off, in case a future user will want to keep them for its exclusive use.

I am making clear however that I am referring to the practices in use by my employer at the time of my retirement.

My firm position in the trade is that while we may use the same tools, the way to use them makes the difference in the first place although new tools should not be excluded. Just think of new ways instead of being conservative.

Interested to see my profile?

The subjects I am referring here in particular are:

1.The tools and the way of measuring
2.The setup on the machines and for measuring
3.How temperature affects large components
4.Geometric tolerances. Problems and solutions
5.Possibility of measuring flatness on the floor and vertical alignment of bores
6.Measuring of a large taper
7.Spigots. What are they and an effective way in measuring them
8.Francis turbine runners. History and progress in fabricating them for the last thirty years
9.How can I contribute to a better and more efficient way in fabrication and repairs at site.

Experience has shown that special care is needed for the measuring of large components as compared to the small and medium size ones. Important factors to consider are the tools and the way of measuring, the setup during machining, temperature and the type of material , the state of the machine itself and the experience of the measuring person and his ability to evaluate each case independently and decide on the best possible way available by minimizing the possibility of errors.

1.The tools and the way of measuring.

The tools illustrated here are mainly used in measuring of small and medium size components. The last picture of a micrometer is for reference only by identifying the micrometer parts since there are similarities with the micrometers used to measure large scale components.

The micrometers for large components

The micrometers for external or internal measurements when measuring large components work on the same principle as on the small pieces except that for the external measurements  instead of the frame shown above, a hollowed rectangular bar is used in order for it  to be light, with two attachments at the two ends, one carrying the anvil and the other the spindle up to the thimble. These attachments are set  in the gauge room to be within the range of the desired size. The cross section of the frame on the ones I have used had a width of 1/2 inch by a height of 1 inch. The ones for the inner measurements resemble the inner micrometer shown in the first picture. The ones I have used are composed of several sections of solid bars of 1″ diameter with adjustments at the ends. These micrometers, because of their weight, are sagging at the center. To compensate for this a support is some times used during their settings in the gauge room, this support must be repeated at the same place and level during measuring. When measuring vertically with this type of micrometer the accuracy may suffer due to deflection.

I have seen in China a micrometer that seems to  compensate or eliminate the sagging problem. It is composed of several sections of tubes, larger in diameter at the center and diminishing in diameter towards the ends. Although I didn’t have the opportunity to examine it fully, I imagine that it is supported internally and vertically with plates or another kind of support.

Another way of measuring large diameters is the Pi (π) tape where π=3.14…measuring on the vernier scale. The disadvantage of this tape is that should be used with extreme care because it gets damaged easily and a kink may result in unreliable measuring and therefore should be re-calibrated quite often. Practice has shown that the accuracy of this tape, especially on very large diameters (6 to 10 meters) is difficult to sustain. The main reason for this is the finishing of the measured surface. The rougher the surface, the larger the unreliability due to the friction on the machined surface.

2.The setup during machining and measuring,

 The ideal way to measure a piece would have been at the same position as working. Since this is almost impossible especially during machining and due to the flexibility of the metals, especially the welded structures, care is taken during the setup that any possible deflection is minimized. As an example I refer here to instructions on setting up a large shaft on a milling machine which I have read at the beginning of my working days but I have never seen it being followed.

“Set the shaft resting on the two ends and place a dial indicator at each end and one on the center, all of them under the shaft. Then start jacking up the center of the shaft and measure the distance traveled on the indicator of the center until the indicators at the end starts moving. Then come down half the travelling distance on the indicator of the center. Support the shaft at this height”

On the same principle we may set any piece on any table for machining.   Care should be taken when jacking a piece laterally  in order to bring it to the desired center position because the piece due to the friction gets distorted laterally.

Setting a piece whose surface is machined, must be evenly supported on the same level either for machining or measuring. Setting a piece on wood for measuring is not recommended  since the wood is compressed  and the part is distorted. We have had a case where we were called on the field because the diameter of a ring type piece was found under the required minimum. The reason was that this piece was set unevenly on the floor and its weight caused it to cave in and measure smaller on one direction.

3.How temperature affects large components: Temperature is a very delicate subject when it comes to large components. Especially in an unstable environment there is not enough time for the temperature to travel and be the same thru the entire component. Here are some examples that I, personally have witnessed:

a) The shaft connecting the runner with the generator is usually made in two parts connected to each other thru spigots. This is a recess (female) and a protrusion (male) called spigots, both of them with very close tolerances so that they will be in the same axis line upon being assembled. One part was made by our shop and the other in Japan. We had used a comparison gauge provided by the Japanese, the same they had used for their part of the shaft. Upon arrival at the site, the male spigot did not fit into the female. I conducted an experiment  by leaving an interior micrometer of 17″ inside the hollowed part of a shaft for 5 days. I was measuring it every morning and every afternoon. There was a constant difference in the morning and afternoon measurements of 0. 0007″ (inches). And this was during the summer time with no severe differences on the ambient temperature.

b) Two identical parts were manufactured for one of our customers who was demanding very close tolerances on all dimensions. One was inspected by me and the other by one of our inspectors and both were found to be within the specified range on a certain diameter, I do not remember the exact size but was around the range of 45”. The customer reported both of them to be under minimum in their diameters. I went to the site to investigate in using the same tools as in our shop. I found them to be under the shop measured size by 0,003” each. The only difference was that the customer had kept the components under a stable temperature for several days.  By investigating further I realized that the table of that particular boring machine was turning hot while the machine was working and arrived to the conclusion that the outer part due to the coolant used while turning had a temperature inferior to the whole body of that component that was affected by the heat of the machine’s table.

c) On a turbine assembly at the site was found that the wicked gate bores of the head cover (20 of them) were on a radius exceeding that of the bottom ring by 0.020”. Upon investigation it was determined that was the difference on temperature. It was winter and they were heating up the place and as the heat rises to the top there was hotter at  that specific height.

About the proper way of measuring I will not say much at this time, I will reserve some comments for later, except that multiple checks in order to arrive to a dimension should be avoided due to the risk of accumulated errors.

I want to mention in this presentation the wards of two men who motivated me thru  my entire working life:  Procura porquê (search the why) were the words of my supervisor Domingo Dias at Mecânica Pesada in Taubaté, Brazil in 1961.

If someone has something to tell you, even if you think he is stupid listen to him, he may know something you don’t know. He was Alec Karoum in Vancouver, Canada in 1969. As an example to this, here is what I have learned from an American at Salt Lake City in 1986. Attach a magnetic base with a dial indicator on the top of a piece of pipe and set it to 0. Then turn the pipe by 180 degrees so that the indicator will be at the bottom of the pipe. No matter how well you tighten it, the reading at this position will be no 0. Up to that time I have aligned  a number of motors to gear boxes or pinions throughout Brazil and Canada all of them with an error.

4.Geometric tolerances. problems and solutions

Concentricity, runout and flatness are the ones I am going to discuss here since I can contribute and perhaps introduce some new ways. Concentricity is the relationship of the center of a specific diameter to a defined datum diameter. If the center of that specific diameter lies within a circle of 0.004” of the datum diameter, we say that it’s concentricity value is 0.004” This means that a diameter may be oval but still concentric. On the other hand the runout is the total indicator reading on the diameter as it turns. In later drawings a runout (total indicator reading) is specified in comparison to the center line of axis of the part. In this case it takes on account the ovality as well. How reliable are the values obtained with a dial indicator is questionable. Since the big machines have considerable weight, moving parts subjected to wear, run on oil etc it is normal that after some time will not perform as expected.

By placing a dial indicator on the same side as the cutting tool it is normal that the runout will be close to 0. If we place a dial indicator at the opposite from the cutting tool side as most of the experienced tradesmen do, what we get is the ovality of the part, simply because the distance between the cutting tool  and the dial indicator which read opposite to the cutting tool is fixed. Then what? My answer is in the following case.

The Beauharnois case

The spigot of a Francis runner, one of the largest we use to manufacture in Quebec, was supposed to be perfectly round. The value of its diameter was the same (within 0.001″) in any different position and the runout (T.I.R) was almost perfect. To bore the coupling bores on the runner’s coupling face we used a jig  which was fitting on the runner’s spigot (male) on the one side and on the mating  shaft’s spigot on the other so that  the bores on the runner to correspond to the same exact bores of the shaft.

When the jig was placed onto the runner’s mounting face there were three equally spaced openings between the runner’s spigot and the mating diameter of the jig. This means that the table of the vertical boring machine was wobbling as it turned, a condition that  could not be detected by the usual means. I thought of a solution to be able to detect this kind of deviation and came up with an idea that at least 4 dial indicators on a fixed position would have detected the error. The same principle may be applied on a horizontal lathe.

The lower bracket in China

In 2004, I was supervising the fabrication of hydraulic components for the Three Gorges project, fabricated in China on North American design. One of the components, the “lower bracket”, a large structure of 296 metric tons,  was noticed with inadequate weld deposit on few vertical ribs around a heavy tube/column between the top and the bottom of this component.

While the fabricator agreed to repair the weld fillets, my concern was that there might be a distortion that could affect the flatness of one of the surfaces with a flatness requirement of .05mm (.002”) and I wanted to put the part on the machine for verification and possible re-machining. The fabricator insisted that he was certain that the amount of weld that was to be deposited on a structure of that size and thickness of plates was impossible to cause any distortion, therefore did not wanted to spare the time and the machine in doing so.  I offered a compromise which he accepted: The inner diameter (the oil well tube) that was welded on the inner part of the surface in question was going to be measured in eight marked spots (4 diameters) before and after the repairs. A deviation on those measurements after the repairs as compared to the ones before, will mean that there was distortion also on the flatness of the machined surface. It turned out to be a deviation of 2.3mm. I do not recall the oil well tube diameter, but I estimate to be around 5 meters.

5.Possibility of measuring flatness on the floor and vertical alignment of bores

Later on I thought of a device that could have verified the flatness requirement on a piece of that size on the floor. This device will be useful as well in setting the bottom ring and the head cover at the turbine assembly on the field and saving time. as compared with the scope that is now presently used.

Another device, still to be tested will be useful in aligning the bores on a vertical sense and if proven right will save a considerable amount of time as compared with the plumb wire that is used up to now.

6.Measuring a large taper

My metrology book instructed  that in order to measure an external taper we place it on the inspection table and measure it’s diameter in two locations, at two different heights separated by gauge blocks. How about a large taper shaft that does not fit on a table?  In the past, in our paper division we were using a precise manufactured angle plate that when strapped onto the taper shaft the top surface of the plate was parallel with the opposite surface of the taper to be measured, meaning that if the measurements at the two ends of the plate were identical, the taper was  as expected.

I have had traveled to Arkansas once to measure the trunnion diameter of a shaft (male) in order to fabricate the mated part (female). The original fabricator of the shaft did not use the same standards as we did and therefore the above method was not applicable. Luckily, that fabricator  was making two groves on the taper at a define distance to each other and I was able to complete the measuring. Then I thought of a method that will allow the measurement of any taper. I welded two balls at the end of a pipe  and the distance between  the centers of the two balls  was measured precisely. Then I placed the bar in a way that the line passing from  the centers of the balls was parallel to the shaft’s axis, By using elementary trigonometry and two measurements from the balls perpendicular to the opposite surface of the taper , one on the large end and the other on the small end, I had everything I needed to figure out the rate of the taper and it’s diameter at any desirable point. By using the same principal I was able to measure the taper on a number of turbine shafts and runner hubs (female) to the extend that in the most cases we had the 95% contact requirement between shaft and runner hub surfaces, saving thus a considerable amount of time, as compared to the time used previously where the measurements were done by using other methods and spent long time afterwards in scraping the hubs until the required contact between surfaces was achieved.

Later I was awarded the U.S. patent 4,777,731 on Oct. 18 1988 and a Canadian patent 1278678 on Jan. 8 1991.

7.Spigots.  What are spigots and a simple and effective way in measuring them

The purpose of the spigots is to connect mating parts e.i adding an extension to a shaft to achieve the length required from the runner end to the turbine end and still precisely matching the axis of the one to the axis of the other so after the assembly will behave as one shaft. Due to the small height on the recess or the protrusion of the spigots it becomes difficult to measure the diameter because of the size of the thimble of the micrometer especially when the piece is still on the horizontal lathe. It is also absolutely necessary that the diameters should be precisely measured in order to preserve the proper alignment.

Sometimes the solution is simple by making a tool based on a small detail. I have this solution and I am ready to disclose it under the proper circumstances.

8.The Francis turbine runners. Progress in fabricating them for the last 30 years

I I remember all types of turbine runners that were made in our shop since 1977. The vast number of them were the Francis type. A number of them  were cast either as a whole piece or in two halves that  were welded together and after some repairs were done where it was visually necessary, the blades and the inflow of the band and crown were splined to engineering instructions. The outflow of the blades was purposely cast longer and was afterwards cut to produce an outflow opening within the specified limits. A part of the back of the blades where most of the wear occurs during operation was cast with a recess where a layer of a minimum thickness of stainless steel was deposited. After a process of heat treatment and pre machining, the  inflow angle was checked in three locations and corrected to be within design limits. The outflow of the blades and the lower distributor line of the blade was checked with a set of templates mounted together and if the template verification and the spline requirements were within the specified limits, the runner was released for the next operation.

On larger runners the blades were casted independently, inspected with a fixture  bent if required and ground to achieve the tolerance specified at specific points on the fixture and inserted into a runner fabrication setup where the crown and the band were already pre-machined and set concentric to each other. The position of the blades at inflow and outflow of the blade were marked on the crown and band and the allowable tolerance marked on the crown and band as well and also controlled with the outflow and lower distributor templates as on the previous paragraph.The accuracy if this setup wasn’t the best, The facts affecting the accuracy were: Inaccuracy in matching the templates from the ones at  the floor first verification to the two at the setup, sagging of the horizontal template on the setup, difficulty in establishing the inflow of the blade lines on the blade itself where was rather a guess than the actual and distortion of the templates during storage. After the blades were fully welded the process will follow as previously.

Then the machined blades made out completely from stainless steel came as the answer to all concerns. A computer generated program will machine a blade to the design. By having all blades almost identical and uniformly placed on the runner will have a more efficient runner with less vibrations and longer life. But the transition to the new technology is to blame, because some of the old habits do not go away at the same time. People tend to be conservative. The outflow and lower distributor templates were still in use at least up to my retirement.

9. How can I contribute to a better and more efficient way in fabrication and repairs at site of Francis runners. I consider that the use of templates is a set-back in the whole process since with an optical coordinate measuring system and the proper manipulation of the coordinates I feel that the results will be far better and the inflow angle check will be no longer necessary. I also feel that during repairs at the site, no templates will be required. Only a robotic arm measuring system is required and a little imagination based in some important simple details.

Summary: Where do I contribute

1)To properly measure geometric tolerances. The method may require a fabricated structure.

2)To measure the spigots efficiently;  A simple but precise tooling is required.

3)Fabrication of a Francis runner with machined blades: An optical coordinate measuring system (theodolites) is preferred and access to the same program as when the blades are machined for comparison. The program I have used in the past measures the actual blade surface in runner coordinates and in the comparison process the results show the material in excess with a + sign or the lack of material with a – sign.

4)Comparing at the site the surface of a blade to the design surface as above or measuring a blade with the purpose of generating a program. A robotic arm is preferred.