JPH0123272B2 - - Google Patents

Description

[Detailed Description of the Invention] The present invention relates to an in-process sizing measurement method,
In particular, by mathematically processing multiple intermittent dimension measurements, the sizing process curve of the workpiece dimensions is determined sequentially over time, and the point at which the workpiece size reaches the set size is determined by extending the sizing process curve. The present invention discloses a novel and progressive in-process sizing measurement method, which is characterized in that it assumes the above point and issues a control signal for the machine tool when the assumed point is reached.

In general, in in-process sizing measurement, when dimensions are measured intermittently rather than continuously, it is more likely to be a coincidence that the measurement operation is carried out at the exact sizing point, and the first measurement after the sizing point is completed. The fixed size was confirmed by the measuring operation. Therefore, even though the dimensions are fixed, they are theoretically forced to vary by an amount corresponding to the common measurement interval, and this variation was considered to be unavoidable.

However, in the mass production of ultra-precision parts, such as mass production grinding of the inner and outer rings of miniature bearings, even if we try to reduce the above-mentioned conventional variations as much as possible, if we shorten the measurement interval and perform frequent measurement operations, the gage As the speed of the traverse mechanism increases, the life of the slide decreases and vibrations occur, which adversely affects machining accuracy and efficiency.

This invention was invented in view of the above circumstances, and as mentioned above, it detects whether the workpiece has reached a fixed size not in terms of dimensions but in terms of time.
In this way, the purpose is to reduce variations in sizing control dimensions and dramatically improve accuracy in mass production.

To explain the method of the present invention using a diagram, the diagram is an example showing the relationship between the cutting cycle and the sizing process curve in grinding, where the horizontal axis is time t, and the vertical axis is the depth of cut s or workpiece size. n is shown, where the straight line indicates rough grinding, the line indicates fine grinding, the line indicates spark out, and the line indicates rapid retreat of the cutting slide table. The sizing process curve just before the end of precision grinding is almost a straight line (n=at+b), and the sizing process curve at the time of sparkout is exponential (N=P
exp(QT)). Therefore, the measured values in each section are n ₁ t ₁ , n ₂ t ₂ …
... or N ₁ T ₁ , N ₂ T ₂ ..., are sequentially memorized, and an approximated curve according to the above formula is sequentially determined from the stored constant number of consecutive measured values using, for example, the least squares method or a Kalman filter. The sizing point is assumed to be one point on the extension of the curve, and the time to reach the sizing point is calculated in real time. Then, each time a measurement operation is performed, such calculations are performed one after another using the measured values, and an updated calculation is performed at the time when the above-mentioned sizing point is reached.

To further explain, for example, if the spark-out start dimension nf is set as one sizing point in advance, and to detect whether this dimension nf has been reached, first, a constant number of the latest consecutive measured values, for example n _{1 t1} ,
An approximate curve (n=at+b) connecting these measured values is determined by n ₂ t ₂ and n ₃ t ₃ . As described above, this approximate curve can be obtained by, for example, calculating using the least squares method or processing using a Kalman filter. Furthermore, by increasing the amount of data used, it is possible to average out variations in individual measurement values and measurement error factors, thereby increasing reliability. However, if the number of data is too large, the responsiveness to changes in the situation (changes in grindstone cutting quality, changes in grinding conditions, etc.) will be poor, and errors will occur in estimating the true dimensional values at that point in time.

Now, the approximate curve obtained as described above is calculated again and updated when the next measurement value n ₄ t ₄ is input. In this way, a more accurate sizing process curve is obtained sequentially in time series, and the time tf to reach the sizing point nf is calculated as one point on the curve. this time tf
can also be determined as the time interval Δt from the latest measurement operation time (for example, t ₄ ). Therefore, when the above time tf or the elapsed time △t from the latest measurement operation is detected, it immediately becomes the sizing point.
nf, and the control signal is generated.

Note that the area FG⌒ of the above-mentioned exponential number is also processed in the same manner.

As explained above, the present invention assumes that the sizing point is a point on the sizing process curve, and by detecting that point in time, it is assumed that the sizing point has been reached. The above-mentioned conventional variations can be eliminated both theoretically and practically, and accuracy in mass production of ultra-precision parts in particular can be dramatically improved.

[Brief explanation of drawings]

The figure is a diagram showing an example of a cutting cycle and a sizing process curve in grinding. Explanation of symbols: Rough grinding, Fine grinding,
BC: Spark out,: Rapid retreat, n ₁ t ₁ ,
n ₂ t ₂ ...: Measured value, N ₁ T ₁ , N ₂ T ₂ ...: Measured value,
nftf: An example of a fixed size point.

Claims (1)

[Claims] 1. An in-process sizing measurement method in which the dimensions of a workpiece being machined are measured intermittently and a control signal is issued to the machine tool when the workpiece reaches a set dimension, and the above measurement values are sequentially stored. , based on at least three or more memorized measurement values, determine the sizing process curve of the workpiece dimensions in a time-series manner, and define the point in time when the workpiece reaches the set dimensions as a point on the extension of the above-mentioned sizing process curve. When the assumed point is reached, the cutting of the workpiece is stopped and spark-out is initiated.Then, the same assumption is made for the dimensions set separately, and spark-out is completed when the assumed point is reached. An in-process sizing measurement method characterized by emitting a control signal that