BACKGROUND ART In industry, there are various kinds of devices for taking dimensional and/or profilometric measurements of workpieces of any type and shape. Mechanical devices, capacitive devices and optical devices in particular are known for high precision measurements.

The mechanical devices include various micrometer comparators, mechanical feelers etc. These devices have disadvantages related to the slowness of the measuring operations (which can not readily be automated) and the usual measurement criteria which require mechanical contact between the measuring instrument and the surface to be measured.

The capacitive devices are generally RLC circuits having a capacitance which varies in dependence on the sensor distance/surface to be measured. These devices have many problems related to their stability in the short and medium term, on the low measurement range and the minimum dimensions of the work surface.

Optical devices include those comprising photocells and those operating on the principle of triangulation. In photocell devices, the light signal received by one or more photocells is interrupted whenever the light is interrupted by the object. These devices are usually very bulky, very limited in their measurements (due to the fact that in an array of photocells, the individual photocells are only able to indicate the presence or absence of incident light, and not the actual variations in intensity of the light) and are very sensitive to external optical disturbances. The devices based on the principle of triangulation can be used on objects having an at least partially reflective surface since the laser beam emitted by a semi-conductor laser source (or a simple LED) is detected and reflected by the (at least partly) reflective surface of the object.

The use of a dimensional measuring method is necessary in many industrial manufacturing processes such as, for example, the manufacture of turbine rotors, to check that the design specifications have been achieved.

A turbine rotor includes several coaxial portions mounted on a single shaft, between which is a circle of blades; during manufacture, these latter must initially be over- dimensioned so that by means of the subsequent removal of material (verified by intermediate check measurements) after mounting on the shaft, they attain the nominal design specification. Before removing the material, the surface profile of each workpiece is detected precisely at several predetermined points of interest and, in particular, at the end points of the workpieces themselves.

Optical devices have been found to the most efficient for obtaining dimensional and profilometric measurements of a turbine rotor. However, these devices and, in particular, those operating on the principle of triangulation, are not usable unconditionally for taking the measurements of interest at any speed of rotation and for all blade thicknesses. Furthermore, fairly complex processing is necessary for the correct interpretation of the signal output to the device, so that it is not possible to take the measurements in real time.

DISCLOSURE OF INVENTION The object of the present invention is to provide an optical method of measurement for taking dimensional measurements, in particular, dimensional measurements of a turbine rotor, which method overcomes the disadvantages of the known techniques described above.

According to the present invention an optical measuring method is provided for taking dimensional measurements of a body, in particular, a turbine rotor, the said body having first and second points the mutual distance along a direction of measurement it is desired to determine, the said first point being located on a surface portion of the said body, the method being characterised in that it includes the stages of : -generating a laser beam along a direction substantially orthogonal to the said direction of measurement ; -locating the said laser beam at an initial known distance, measured along the said direction of measurement, from the said second point ;

-causing relative approach of the said laser beam and the said body along the said direction of measurement in order to bring the said surface portion of the said body into a position of partial interference with the said laser beam; -detecting an optical image generated by the partial interference of the said laser beam by the said surface portion; -generating a measurement signal indicating the intensity of the said optical image; -stopping the said approach movement on reaching a predetermined threshold value of the said measurement signal; -detecting the final distance between the said laser beam and the said second point; and -calculating, on the basis of a comparison between the said initial and final distances, the distance between the said first and second points.

The present invention also concerns a device for taking dimensional measurements, in particular, for taking dimensional measurements of a turbine rotor, which device enables the method described above to be implemented.

According to the present invention, an optical measuring device is provided for taking dimensional measurements of a body, in particular, a turbine rotor, the said body having a surface portion, the distance of which from a reference point is to be determined, characterised in that it includes: -laser emitting means operating to emit a laser beam; -movement means which cause the said body and the said laser beam to perform an approaching movement along a measurement direction substantially

orthogonal to the said laser beam; -photodetector means for receiving the said laser beam, comprising at least first and second adjacent optically sensitive areas, aligned substantially along the said measurement direction and which generate respective first and second signals indicating the intensity of the light incident thereon; -processing means for obtaining, from the said first and second signals, a measurement signal indicating the depth of penetration of the said surface portion of the said body into the said laser beam; BRIEF DESCRIPTION OF THE DRAWINGS The present invention will now be described with reference to the accompanying drawings which illustrate a non-limitative embodiment, in which: -Figure 1 schematically illustrates the device of the invention; -Figure 2 shows the sensitive area of an optical sensor used in the device of Figure 1; -Figure 3 shows the use of the device of Figure 1 for taking dimensional measurements of a turbine motor ; -Figure 4 is a flow diagram showing the principal stages of the method for measuring the dimensions of a turbine rotor; -Figure 5 shows the variation of the characteristics of several signals generated by the device of Figure 1 during a dimensional measurement; -Figure 6 shows schematically several parts of the circuit of the device of Figure 1; -Figure 7a-7d show the variation of signals generated by the circuit parts of Figure 6 during a dimensional measurement;

-Figure 8 shows an operating condition of the device of Figure 1 for taking a dimensional measurement of a turbine rotor ; and -Figure 9 shows a further type of dimensional measurement that can be performed on a turbine rotor by means of the device of the invention.

BEST MODE FOR CARRYING OUT THE INVENTION In Figure 1, the reference numeral 1 generally indicates an optical measuring device that can be used to measure dimensions of objects of any shape and size.

The device 1 essentially includes a laser detection unit 2, referred to below as a laser fork, and a movement device 3 having three axes (shown only partly) which moves the laser fork 2 along three orthogonal axes x, y and z and rotates the laser fork 2 itself about an axis w parallel to the z axis.

The laser fork 2 includes an elongate support element 4, a laser emitter 5 (for example, of the commercial He-Ne type) attached to one end of the support element 4 and which emits a laser beam 6 parallel to the direction defined by the support element 4 itself, and an optical sensor 7 firmly connected to the opposite end of the support element 4 and having a sensitive surface 8 facing the laser erritter 5, to be able to receive the laser beam 6. Alternatively, the optical sensor 7 may be a sensor of the kind having four quadrants (not shown) commonly used in telemetry (and of which, for the application in question, only two of the four quadrants are used), or a sensor of the kind shown in Figure 2 having two contiguous sensitive areas 8a and 8b of equal dimensions. In particular, the sensor 7 of Figure 2 comprises two half-moon shape areas 8a and 8b facing each other and separated by a central line 9. The areas

8a and 8b generate respective electrical voltage signals Va and Vb proportional to the light intensity detected by them. Figure 2 also shows, in broken outline, the"spot"of the laser beam 6 when its path is not interupted.

The output of the sensor 7 is connected to an electronic processing unit 12, this also being located on the support element 4 and operable to receive the voltage signals Va and Vb from the sensor 7 and to generate a digital output signal U, used as described below.

The movement device 3 has movable arms 13 (only one of which is partially shown in Figure 1) which, together with further movement elements that are not shown, supports and moves the laser fork 2 along the x, y and z axes, and a hinge element 20 interposed between the laser fork 2 and one of the movable arms 13, which enables the laser fork 2 to rotate about the w axis; in particular, the laser fork 2 can assume a first angular measurement position (illustrated in Figure 1) in which it faces along the x axis, and second and third angular measurement positions (rotated by 90° and, respectively,-90° with respect to the first angular measurement position) in which it is orientated in the same direction as and, respectively, in the opposite direction to the y axis.

The movement device 3 further includes a control unit 14 which receives an input signal U and which, as a function of the value of the signal U, controls the arms 13 in order to move the laser fork 2 in a controlled and precise manner along the x, y and z axes. The movement device 3 also has a position detecting unit 11 (for example, a unit which reads the position on a graduated scale) which detects the approximate

position of the laser fork 2 along the x, y, z axes.

Finally, the device 1 includes a calculation unit 40 which is connected to the input of the processing unit 12 and the position detecting unit 11 and which provides a precise displacement value of the laser fork 2.

In order to understand the operation of the device 1, an object of particularly simple shape, such as that indicated 30 in Figure 1, is initially considered. The object 30, supported in a manner not shown, has an elongate shape along an axis r and it is desired to determine the dimensions along the axis r itself. In order to measure this dimension, it is necessary to measure the distance between first and second opposite ends 30a, 30b of the object 30, along the axis r. The measurement of this distance involves the following stages: -positioning the object 30 such that the axis r is orientated orthogonally to the laser beam 6 and with the first end 30a facing the device 1; -positioning the laser fork 2 in an initial position in which the laser beam 6 is at a known distance from the second end 30b of the object 30; -moving the laser fork 2 forward until it reaches a position of partial interference in which the first end 30a of the object 30 interferes partially with the laser beam 6; -processing the voltage signals Va and Vb generated by the areas 8a and 8b to obtain a measurement signal C correlated with the depth of penetration of the first end 30a of the laser beam 6; -stopping the laser fork 2 in a final position in which the measurement signal C has reached a predetermined threshold value THc ;

-detecting, by means of the position detecting unit 11, an approximate value Xapp of the distance moved by the laser fork 2 between the initial position and the final position; -obtaining, by means of the calculation unit 40, from the measurement signal, the depth of penetration of the first end 30a into the laser beam 6 in the aforesaid final position and, based on the approximate value xapp of the distance moved by the laser fork 2 and the depth of penetration of the first end 30a into the laser beam 6 in the aforesaid final position, a precise value for the distance moved by the said laser fork 2; and finally -calculating the difference between the initial distance of the laser beam 6 from the second end 30b of the object 30 and the precise value of the distance moved by the laser fork 2 to obtain the desired dimension.

As is clear, the principal aspect of the method of the invention is the determination of the depth of penetration of the first end 30a into the laser beam 6, which is effected as described below by using the diffraction image produced on the sensor 7 following the movement of the laser beam 6 tangentially to the profile of the first end 30a of the object 30. From this, it is clear that the device of the invention operates quite differently from a photocell device, and that the calculation of the depth of penetration confers a degree of resolution on the measurement that is not otherwise achievable.

It is also clear that the aforesaid measurement can be taken by keeping the device 1 still and moving the object towards the laser beam 6. This arrangement is particularly advantageous in the case of small objects and requires the presence of a translation

device provided with an instrument for detecting precisely the movement, which device supports and moves the object 30 towards the laser fork 2, which is in a fixed position. Similarly to the above case, if the initial distance between the laser beam 6 and the second end 30b of the object 30, the distance moved by the object 30 in order to intercept a predetermined fraction of the laser beam 6 (with the signal C reaching the threshold THc) and the depth of penetration of the first end 30a into the laser beam 6 itself are known, it is possible to obtain the precise value of the distance between the first end 30a and the second end 30b of the object 30.

Figure 5 shows an example of the qualitative variation of the signals Va and Vb generated by the two sensitive areas 8a and 8b of the sensor 7 as a function of the percentage of intrusion of the first end 30a into the laser beam 6. The intrusion is obviously variable between a minimum of 0% when the first end 30a is completely out of the laser beam 6 and a maximum of 100% when the first end 30a completely intercepts the laser beam 6. It should be borne in mind that, for the application in question, a laser beam can be used with a section of variable dimensions within a certain range (laboratory tests have been performed with a laser beam having a diameter of the order of millimetres) and that the dimensions of the areas 8a and 8b are such that the luminous"spot"produced by the laser beam 6 covers a significant fraction of their surfaces, as shown in Figure 2.

As well as the curves of the signals Va and Vb, the curves of the signals Vdifr = Va- Vb and Vso = Va+Vb are shown. As shown by the Figure in question, when the first end 30a is completely out of the laser beam 6, the voltages Va and Vb have the same maximum value Vm. As soon as the first end 30a starts to intercept the laser beam

6, these values become different from one another since the area 8b starts to be obscured. However, as long as the penetration into the laser beam 6 is less than a certain threshold (zone 1 in Figure 1), the total signal Vsom is approximately constant.

This result can be interpreted by envisaging that there is only one movement of energy from an area 8a, 8b to the other, with no significant loss.

In addition, it can be noted that in zone 1, the difference between the two voltages Va and Vb occurs in the opposite direction from that which one would imagine by applying principles of optical geometry: more precisely, it is the end 8b, that is, the area which is initially obscured, which detects an increase of intensity. This behaviour is due to the presence of diffraction phenomena.

Once the limit of the zone 1 has been reached, the voltages Va and Vb return to having the same value (slightly lower than the initial value) and, from here on, as the penetration of the first end 30a in the section of the laser beam 6 increases, the area denominated zone 2 in Figure 5 is entered. In this zone, the total signal Vsom, after an initial decline, decreases approximately linearly. Here, the energy lost both as a result of reflection on the first end 30a of the object 30, and the diffraction which moves the centre of gravity of the luminous image outside the sensitive areas 8a and 8b, becomes significant.

The voltage Va increases until it reaches a maximum value Va, max and from here decreases linearly until it disappears. On the other hand, the voltage Vb decreases linearly starting from the boundary between zone 1 and zone 2. Typical values for the dimensions of zone 1 and zone 2, obtained in laboratory measurements (using an He-

Ne laser of 0. 5mW), are approximately 0.3mm for zone 1, and approximately 0.8mm for zone 2. All of the dynamics of the system therefore develop within an interval of the depth of penetration of slightly greater than a millimetre.

The linearity of the characteristics in specific zones provides a test of the possibility of using the device 1 effectively for measurements of the kind described above. By way of experimental laboratory measurements, signals Va and Va have been obtained having characteristics with zones of linear variation (in a common range of approximately 0.2mm) with slopes of approximately 17V/mm and approximately 25 V/mm respectively. This means that variation in depth of 1/100 mm leads to a variation in the voltage Va equal to 0.17 V and a variation in the voltage Vb equal to 0.25 V. Voltages of this order of magnitude can be measured accurately without utilising particularly sophisticated techniques: this also confirms that the instrument in question can be manufactured simply and economically.

The signals Va, Vb described above are used by the processing unit 12 to generate a digital output signal U, the change in level of which coincides with the measurement signal C reaching the threshold value THc. In order to understand how the processing unit 12 generates the signal U from the voltage signals Va and Vb, reference is made to Figure 6 which shows (in schematic manner) the sensor 7 with its sensitive areas 8a and 8b, the laser beam 6, the first end 30a of the object 30 located in a position of partial interference with the laser beam 6, and the processing unit 12.

The processing unit 12 includes a pair of circuits 22a, 22b for adapting the signal, the first of which, indicated 22a, is connected to the area 8a for receiving the signal Va,

and the second of which, indicated 22b, is connected to the area 8b for receiving the signal Vb.

Both of the adaptation circuits 22a, 22b also receive, via a separate input, a threshold voltage value Vt for controlling the offset. Each adaptation circuit 22a, 22b outputs a voltage signal V'which is a function either of the voltage signal V received from the respective area 8a, 8b or of the signal Vt. In particular, the first adaptation circuit 22a outputs a signal V'a equal to G (Va-Vt), while the second adaptation circuit outputs a signal V'b equal to G. (Vb-Vt).

The processing unit 12 also includes a subtractor circuit 23 and a summing circuit 24, each of which is connected as input to either the first or second signal adaptation circuit 22a, 22b to receive either the signal V'a or the signal V'b. The subtractor circuit 23 provides as output the difference signal Vde= V'a-V'b, while the summing circuit 24 provides as output the total signal Vsom = V'a+V'b, the qualitative variations of which are shown in Figure 5.

The processing unit 12 also includes a divider circuit 25 which is connected as input to the subtractor circuit 24 to receive the difference signal Vdiff, and to the summing circuit 23 to receive the sum signal Vsom, and generates a comparison signal C from the ratio between the difference signal Vdjff and the sum signal Vsom: C = Vdia/Vsom.

The signal C provides a measurement of the movement of the centre of gravity of the optical image detected by the sensor 7.

The qualitative variation of the signal C as a function of the depth of penetration of the first end 30a into the laser beam 6 is shown in Figure 7a. As is clear from Figure 7a, it increases linearly until it reaches an intrusion percentage less than 50%, and shows a non-linear decrease for higher intrusion percentages.

The output from the comparison circuit 25 is connected to a first comparator circuit 26 which receives as input the measurement signal C and the threshold value THc, also indicated on the ordinate axis of the graph of Figure 7a. The threshold value THc defines a limit below which the curve representing the signal C certainly shows (in its initial portion) a linear variation and therefore enables it to rise easily to the intrusion percentage of the first end 30a in the laser beam 6. As shown in Figure 7a, two values of the intrusion percentage, Il and 12, correspond to the threshold value THc, of which only the first is relevant to the measurement in question.

The first comparator circuit 26 compares the signal C and the value THc and generates a logic signal F as output, the variation of which as a function of the percentage of intrusion is shown in Figure 7b. As is clear from Figure 7b, the signal F is equal to 1 when the intrusion percentage is between 11 and I2.

A second comparator circuit 27 receives as input the signal V'a and a safety threshold value THs which defines a threshold value of the signal V'a, below which there may be an anomaly and/or a total interruption of the laser beam 6; Figure 7a shows in broken outline a safety zone in which V'a > THS corresponding to an intrusion percentage greater than a value I3; in the example under consideration, the value I3 is between 11 and I2. The second comparator circuit compares the signal V'a and the safety threshold value THs and generates a signal G, the variation of which as a

function of the intrusion percentage is shown in Figure 7c. As is clear from Figure 7c, the signal G is equal to 1 when the intrusion percentage is greater than I3.

Finally, the processing unit 12 includes an OR type logic gate 28. The OR logic gate 28 receives as input the signal F and the signal G, and generates the output signal U, the variation of which as a function of the intrusion percentage is shown in Figure 7d.

In the example considered, the signal U has a logic value 0 for an intrusion percentage of less than Il, and a logic value 1 for an intrusion percentage of greater than 11.

Generally, the signal U will take the logic value 0 when the signal C is less than the threshold THc and when the signal V'a is in the safety zone, while it will take the logic value 1 when one of the two aforesaid conditions is no longer satisfied.

Therefore, in general the signal U will pass from the logic value 0 to the logic value 1 when the laser beam 6 is intercepted with an intrusion percentage such that the signal C moves out of its region of linearity (C > THc), thereby no longer enabling an accurate evaluation of the intrusion percentage itself, or if the signal V'a is too low (Va < THS) which gives rise to an emergency situation in which the first end 30a could collide with the support element 4.

Described below, with reference to Figure 3, is a practical application of the method of the invention, in which the device 1 is used in the process of working a turbine rotor 6 in order to obtain dimensional measurements of the turbine rotor 6 itself. As shown in figure 3, the device 1 cooperates during this working process with a machine tool 15 (in this case, a grinder).

The turbine rotor 16 has a shaft 17 having an axis 18 and also has a ring of radial blades 19 mounted on the shaft 17. Each blade 19 is formed so that, after it has been mounted on the shaft 17, its length exceeds that required for use on the turbine 16, and the process of working the turbine 16 therefore includes a final stage in which, by means of the progressive removal of material from the end of the blades 19 by means of the machine tool 15, the blades 19 themselves are dimensioned according to the design data.

The laser fork 2 is positioned with its support element 4 orientated in a direction substantially orthogonal to the axis 18, and is also disposed in its first angular measurement position, while translation axis x is directly orthogonal to the axis 18, as shown in Figure 1.

By virtue of the possibility of moving along the three axis x, y, z, the laser fork 2 can be moved towards and away from the axis 18 (translation along the x axis), parallel to the axis (translation along the y axis) or parallel to the direction defined by the support element 4 (translation along the z axis).

The sensor 7 shows the line 9 of demarcation of the two areas 8a and 8b orientated orthogonally to the direction x such that when the laser beam 6 is intercepted by the end of one of the blades 19, the"spot"of the laser beam 6 starts to be obscured starting from one of the two areas 8a, 8b; in particular, it is assumed below that the sensor 7 is positioned such that the area 8b is the first area to be obscured.

The machine tool 15 is connected to the device 1 to be able to exchange information therewith during working. This exchange of information principally comprises the processing unit 12 sending the signal U to the machine tool 15 so that the results of the measurements taken by the device 1 can be used in real time in order to control the working operations. These measurements are conventionally taken separately from the working process and the information obtained is used to control the working operations only after all the measurements have been taken.

The operation of the device 1 in the process of working the blades 19 is described below with reference to the flow chart of Figure 4.

In a preliminary stage of the process of working the blades 19 (block 100), the laser fork 2, already positioned along the y axis in a position corresponding to the axial position of the blades 19, is disposed along the x axis in an initial position close to the blades 19 themselves, starting from which the coordinates along the x axis itself will be measured as the laser fork 2 moves.

Before working the blades 19, the coordinate along the x axis of the end of stroke position of the laser fork 2 is established. In this end of stroke position, the distance of the laser beam 6 from the axis 18 is equal to the radial dimension D of the blades 19; in other words, the laser beam 6 in this end of stroke position must define a line which tangentially delimits the ring of blades 19 at the end of the working process on the blades 19 themselves.

In a subsequent stage (block 110), with the turbine 16 rotating and the laser emitter 5 in operation, the movement device 3 slowly moves the laser fork 2 at constant velocity along the x axis towards the turbine 16.

Where there are blades 19 which have overall radial dimensions greater than the dimension D, the advance is interrupted every time the laser beam 6 is intercepted by the end 19a of one of these blades, and the blade 19 in question is consequently worked to reduce its dimensions. During this process, the processing unit 12 receives signals Va, Vb generated by the sensor 7, and generates the signals Vdia, Vsom, C, F, G, and U, the development of which, as a function of the intrusion percentage of the end 19a in the laser beam 6, are as already described above with reference to Figures 5 and 7a-7d.

In more detail, each time a blade 19 of radial dimension greater than the predetermined dimension intercepts the laser beam 6 beyond a predetermined threshold, the signals Va and Vb generated by the sensor 7 are such that the processing unit 12 in turn generates a signal U indicating penetration beyond a predetermined threshold, that is, a signal U of logic value 1; once the control unit 14 receives the signal U of logic value 1, it stops the laser fork 2.

At the same time, a device 21 which controls the movement of the turbine 16 (Figure 3), the input of which is connected to the processing unit 12 and the output of which is connected to the turbine 16 itself, receives the signal U of logical value 1 from the processing unit 12 itself and, after having stored the angular position of the blade 19 which intercepted the laser beam 6, causes the turbine 16 to stop in such a position that the particular blade 19 faces the machine tool 15. The machine tool 15, this also

being operated on receiving the signal U of logical value 1, moves towards the blade 19 so as to start removing material from the end of the blade itself (block 120) in order to reduce its radial dimension to a predetermined value D greater than the final envisaged dimension D.

After having repeated the procedure described above for all of the blades 19 having radial dimensions greater than the required radial dimension, and after the laser fork 2 has reached an abscissa x corresponding to a distance of the laser beam 6 from the axis 18 equal to D, the working of the blades 19 continues with a different procedure and, in particular, with no more stopping of the turbine 16 and with the contemporaneous advancement of the laser fork 2 and the machine tool 15 (block 130) towards the axis 18, at a constant velocity and without stopping. This advancement continues until the laser fork 2 reaches its end of stroke position at which the dimension of the blades 19 is equal to the final specified dimension D.

Once this position has been reached, the laser fork 2 is automatically stopped and the machine tool 15 is automatically turned off. In particular, the machine tool is switched off once it receives a signal U of logic value 0 from the processing unit 12, which indicates the fact that the blades have been worked until the percentage of intrusion of the blades 19 themselves into the laser beam 6 has been reduced to a value such that C < THc.

By virtue of this control of the signal C, the predetermined dimension D is achieved with a very high degree of precision. Laboratory tests have shown that the method of the invention enables a radial difference between the longest and shortest blades to be obtained at the end of the working, that is less than 1/100mm (approximately).

The method described above for detecting the radial dimensions of the turbine rotor 16 can be integrated with an additional measurement, this being to establish the maximum radial dimension of the blades 19. This auxiliary measurement can, for example, be taken at the end of the working process described above in order to verify that the final effective dimension Deff of the blades 19 corresponds effectively with the predetermined value D.

The auxiliary measurement is described with reference to Figure 8 and includes, in summary, the following stages: -determining a threshold VTHA which is slightly less than VTHC and still in a linear measurement region, selected depending on the"density"of the blades and the residual errors in the radial measurements thereof ; -moving the laser fork 2 towards the axis 18 in such a way as to locate the laser beam 6 at a distance d from the axis 18 itself (measured along the x axis); -operating the turbine 16 at a constant and known, preferably low, angular velocity m; -detecting, for each blade 19, the time t, in which the blade intercepts a percentage of laser beam such that the signal C reaches the threshold VTHA, and the time t2 in which the blade, on leaving the laser beam, intercepts the same percentage of the laser beam, causing the signal C to reach the threshold VTHA in the opposite sense; and -obtaining, based on the value of the time interval'C = t2-tl, the distance d and the angular velocity m, the value of the effective dimension Deff of the blade itself ; <BR> <BR> <BR> this is possible because, once the distance d and the time interval r is known, the

peripheral velocity v of the end 19a of the blade 19 can be obtained and, from the knowledge of the velocity v and the angular velocity o, it is possible to obtain the dimension Deff of the blade based on the relationship Deff = v/co.

Obviously, this measurement is only possible if the blades 19 are not too close to each other, as otherwise a blade could start to intercept the laser beam 6 before the preceeding blade has left it.

In addition to enabling the measurement of the radial profile of the turbine 16, the device 1 can be used to measure the axial distance (along the axis 18) of axially separated parts of the turbine 16.

In order to understand how this measurement can be taken, reference is made to Figure 9 which shows a partial perspective view of the turbine rotor 16. For example, it is desired to measure the distance, along a direction parallel to the axis 18, between two substantially flat surfaces 32,33 (real or theoretical) which define, for a given direction of advance along the axis 18, respective discoid portions 34,35 mounted on the shaft 17 of the turbine 16 and coaxial with the axis 18.

In order to take this kind of measurement, the laser fork 2 is disposed in the second or third angular measurement position (rotated by +90° or, respectively, by-90° with respect to the first angular measurement position), depending on the envisaged direction of advance for the measurement. For example, it is supposed that, to take the measurement of interest, it is necessary to dispose the laser fork 2, as shown in Figure 9, in the second angular measurement position and that the orientation with

respect to the x, y and z axes will therefore be as shown in the drawing.

The measurement method involves the following operations: -locating the laser fork 2 in a starting position (indicated PO and represented in broken outline) facing the surface 32; -moving the laser fork 2 towards the surface 32 by moving the laser fork 2 itself in the direction y until it reaches a position (indicated PI) in which the laser beam 6 is partially intercepted by the surface 32 and the signal C reaches the threshold THc (portion Y1); -stopping the laser fork 2 in the position P 1; -detecting, according to the method described above (and, therefore, with a resolution of the order of hundredths of a millimetre), the distance yPl travelled by the laser fork 2 from the starting position; -moving the laser fork 2 away from the axis 18 by moving it in the direction x until the laser fork 2 is at a distance from the axis 18 that is greater than the maximum radial dimension of the turbine 17 (portion xl); -moving the laser fork 2 in the direction y in order to bring it closer to the discoid portion 35 (portion y2); -moving the laser fork 2 towards the axis 18 in the direction x until the laser fork 2 is facing the surface 33 (portion x2); -moving the laser fork 2 towards the surface 33 by moving the laser fork 2 itself in the direction y until it reaches a position (indicated P2) in which the laser beam 6 is partially intercepted by the surface 33 and the signal C reaches the threshold THc (portion y3); -stopping the laser fork 2 in the position P2;

-detecting, according to method described above, the precise distance yp2 travelled by the laser fork 2 from the starting position; -subtracting the value yPl from the value yp2 to obtain the value of the distance of the surface 32 from the surface 33.

The advantages of the method and the device are clear from the above description.

As already discussed above, the measurement method of the invention is applicable to any kind of object. It is clear that the use of the measurement method is advantageous in all industrial applications in which it is necessary to take diagnostic and/or dimensional measurements on work in progress. Furthermore, as described above, the method of the invention can also be applied to manufacturing processes in which verification measurements on workpieces are necessary during the process itself, and has the advantage that it can be performed automatically and synchronised with the working operations.

Experimental measurements of the radial dimensions of the blades 19 of a turbine 17 conducted during the working of the turbine itself have confirmed that it is possible to finish the working process with a significant time saving and at a lower cost with respect to conventional working processes, in which the measurement and working stages are performed separately; it is, in fact, considered that in the conventional measuring techniques, the measurement of the radial dimensions of the blades 19 must be done several times following every reduction in the dimensions in order to make sure that the final specified dimensions are not exceeded.

The hinged structure of the movement device 3 of the laser fork 2 also makes the device 1 extremely versatile and able to take dimensional measurements of a single object along various directions.

Furthermore, the method of the invention provides a precision of measurement and, therefore, of working, which is not otherwise achievable with known techniques unless an extremely large number of measurement/working steps is used.

Finally, the device of the invention is particularly simple and can be manufactured, at least as regards the optical part, with commercially available components.

It should also be remembered that the precision of the measurement is conferred by the particular configuration of the sensor 7 and the method of processing the signals Va, Vb generated by the sensitive areas 8a, 8b of the sensor 7 itself ; in particular, the use of the diffraction images and the variation in position in the optical centre of gravity of the sensor 7 represent a particularly innovative aspect with respect to the previously known measuring techniques.

Finally, it is clear that modifications and variations can be introduced to the method and the device described and illustrated here without by this departing from the ambit of the present invention.