JPH04115853A - Profile control method - Google Patents

Abstract

PURPOSE:To improve a profile accuracy, by finding the optimum value of a profile motion gain of a VN gain or VT grain, etc., with a fuzzy inference being executed at a fuzzy control part, with the displacement information of each axis as a fuzzy parameter, and performing a profile control by these motion gains. CONSTITUTION:A fuzzy inference part 41 estimates respective membership function from error displacement quantity epsilond, correction displacement quantity epsiloni according to a fuzzy rule. The membership function of VN gain GN and VT gain GT is estimated from this membership function. An interpretation part 42 finds VN gain Gn and VT gain GT with the estimated value of this membership function being fussed by a method of elastic center. A profile control is carried out by the VN gain GN and VT gain GT. Consequently, the optimum value of the VN gain GN and VT gain GT is found always, and the profile control of good accuracy can be performed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a profiling control method for performing processing by tracing a mold, plastic model, etc., and particularly relates to a profiling control method using fuzzy control.

[Conventional technology]

The profiling control device calculates a composite displacement from the displacement of each axis from a sensor such as a stylus, and further determines an error displacement amount which is the difference between the composite displacement and a reference displacement. This error displacement amount is multiplied by a gain based on experience, and the velocity in the normal direction is VN.
and the velocity VT in the tangential direction.

VN gain (GN) is used to calculate the speed in the normal direction, and VT gain (GT) is used to calculate the tangential speed. In other words, the velocity VN in the normal direction and the velocity VT in the tangential direction are determined by VN-εd-GN VT ''' V Cll0-εd-GT.Here, εd is the displacement ε detected by the sensor and This is the difference from the reference displacement amount ε0.

εd=ε−εO The velocity of each axis is as a function of the above normal velocity VN, tangential velocity VT, and displacement generation direction angle θ, as follows: Vα=f (VN, VT, θ) Vβ=g (VN , VT, θ).

Here, α and β are axes that specify the plane on which tracing is performed. For example, in the case of tracing control on the x-2 plane, Vx=VTs inθ−VNcosθVz=−VTc
osθ−VNsinθ.

[Problem to be solved by the invention]

However, in the conventional tracing control method as described above, the VN gain GN and the VT gain GT are always constant, and accurate tracing control cannot necessarily be performed.

For example, it is known from experience that when the error displacement position εd is small, the VN gain GN is large, and when the error displacement amount εd is large, the VN gain GN is made small, which improves the machining accuracy.

The present invention has been made in view of these points, and it is an object of the present invention to provide a tracing control method that improves tracing accuracy by inferring tracing operation gains such as VN gains using fuzzy control.

[Means to solve the problem]

In order to solve the above problems, the present invention applies fuzzy inference using the displacement information of each axis obtained during the profiling operation as a fuzzy variable in a profiling control method that performs machining by tracing a model. A tracing control method is provided, which is characterized in that an optimal parameter of an operating gain parameter is determined, and tracing control is performed using the optimal parameter.

[Effect]

Using the displacement information of each axis as a fuzzy variable, a fuzzy control unit performs fuzzy inference to determine the optimum value of a tracing operation gain such as a VN gain or a VT gain. Tracing control is performed using these operating gains.

〔Example〕

Hereinafter, one embodiment of the present invention will be described based on the drawings.

FIG. 1 is a block diagram showing the configuration of a profiling control device for implementing the present invention. In the figure, a tracer head 20 installed in a profiling machine tool detects displacements εX, εy, and εZ in the X-axis, Y-axis, and Z-axis directions caused by the stylus 21 at its tip coming into contact with the model. Then, the switching circuit 1 and the combining circuit 3 are manually operated. The switching circuit 1 selects the biaxial displacements ε1 and ε2 of the tracing plane commanded via the operation panel 30 from among these displacements and inputs them to the indexing circuit 2 manually.

The indexing circuit 2 uses the displacement amounts ε1 and ε2 to form the following equation, Ec=
Calculate ε1/(ε12+ε22)l/2=cose Es=ε2/(ε12+ε22)I/2=sine to obtain index signals Ec(cosθ) and Es(si
Find nθ).

Further, the synthesis circuit 3 calculates a composite displacement amount ε, ε=(εx2+εy2+εz2)1/2, from each displacement amount εx1εy and ε2, and inputs it to the calculator 4. The calculator 4 calculates and outputs an error displacement amount εd, which is the difference between the composite displacement amount ε and a preset reference displacement amount ε0, εd=ε−60.

The error displacement amount εd is sent to the arithmetic unit 4b and the error integration circuit 4a. The error integrator circuit 4a is a circuit for removing the well-known seat deviation in an automatic control system, and this circuit enables the following operation that satisfies ε-ε0=εd=0. Error integration circuit 4
a integrates the error displacement amount and outputs the corrected displacement amount εl.

The arithmetic unit 4b adds the error displacement amount εd and the corrected displacement amount εi and inputs the result to the tangential velocity signal generation circuit 5 and the normal velocity signal generation circuit 6.

On the other hand, the error displacement amount εd and the corrected displacement amount εi are sent to the fuzzy control section 40. As will be explained in detail later, the fuzzy control unit 40 calculates the error displacement amount εd and the corrected displacement amount εi.
Apply fuzzy inference to the optimal VT gain GT and VN
Find the gain GN.

The VT gain GT and VN gain GN are sent to a tangential velocity signal generation circuit 5 and a normal velocity signal generation circuit 6, respectively.

The tangential velocity signal generation circuit 5 generates the error displacement amount εd, :! :VT
The gain generates a feed rate signal VT perpendicular to the stylus displacement direction (tangential velocity signal). Further, the normal velocity signal generation circuit 6 calculates the error displacement amount εd and V
The N gain GN generates a feed speed signal in the stylus displacement direction, that is, a speed signal VN in the normal direction. That is, as described above, the velocity VN in the normal direction and the velocity VT in the tangential direction are determined by the following equations.

VN=εd-GN VT=VcMo-E a -c'r At this time, the optimal values of the VN gain GN and the VT gain GT are determined by the fuzzy control unit, so that the tracing accuracy can be improved.

The distribution circuit 7 receives index signals cos θ (Ec), sin θ
(Es), the following equation using the tangential velocity signal VT and the normal velocity signal VN, V1=VTX sinθ−VNX cosθV2=−
VTXcosθ−VNXsine is calculated to obtain the velocity components V1 and V2 of the two axes of the tracing plane, and these are sent to the X-axis, Y-axis, and Z-axis velocity signals Vx via the output axis selection circuit 8.
, Vy and V2, and the servo amplifiers 9x, 9
amplified by y and 9z.

Then, the servo motors 22x and 22y are driven by the outputs of the servo amplifiers 9X and 9y, and a table (not shown) is moved. Further, the servo motor 222 is driven by the output of the servo amplifier 9z, and a cutter (not shown) moves in the Z-axis direction together with the tracer head 20.

FIG. 2 is a detailed diagram of the fuzzy control section. The fuzzy control section 40 consists of a fuzzy inference section 41 and an interpretation section 42.

The fuzzy inference unit 41 calculates the error displacement amount εd and the corrected displacement amount εi.
Fuzzy inference is performed from , and the membership functions of VT gain GT and VN gain GN are evaluated. The interpreter 42 defuzzifies the evaluation value from the determined membership function and outputs the actual VT gain GT and VN gain GN.

This fuzzy control section 40 is composed of a processor that controls the profile control device and fuzzy control software.

Moreover, it can also be configured with hardware dedicated to fuzzy control.

Next, we will discuss fuzzy inference. Fuzzy inference consists of an antecedent condition and a consequent part. Here, the error displacement amount εd and the corrected displacement amount εi are used as antecedent conditions, and the magnitude of each is set to three levels: large, medium, and small.

Further, the VN gain GN and the VT gain GT are taken as the consequent part, and the magnitude of each is set to three levels: large, medium, and small.

Nine corresponding fuzzy rules are created from these antecedent conditions and consequent conditions. FIG. 3 is a diagram showing fuzzy rules. Here, the fuzzy rule 43 consists of nine fuzzy rules (a) to (i). For example, rule (a) is ``if the error displacement position εd is small and the corrected displacement amount εi
is small, the VN gain GN is made large and the VT gain G
It means "to make T large". The same applies to other rules, so their explanations will be omitted.

FIG. 4 is a diagram showing the membership function of the degree of error displacement of the antecedent condition. The horizontal axis in Figure 4 is the error displacement position ε
d, and the vertical axis is the degree (grade). here,
Line S1 is a membership function indicating that the error displacement amount εd is small, line M1 is a membership function indicating that the error displacement amount εd is medium, and line B1 is the error displacement The quantity εd is large (b i
g) is a membership function representing that.

FIG. 5 is a diagram showing the membership function of the degree of correction displacement amount of the antecedent condition. The horizontal axis in Figure 5 is the corrected displacement amount ε
i, and the vertical axis is the degree (grade). here,
Line S2 is a membership function indicating that the corrected displacement amount εi is small, and line M2 is a membership function indicating that the corrected displacement amount ε1 is medium, line B2. The corrected displacement amount εi is large (b
i g).

FIG. 6 is a diagram showing the membership function of the degree of VN gain of the consequent part. The horizontal axis in FIG. 6 is the VN gain GN, and the vertical axis is the grade. Here, line S
3 is a membership function representing that the VN gain GN is small (sma l 1), line M3 is a membership function representing that the VN gain GN is medium (m e di um), and line B3 is a membership function representing that the VN gain GN is medium (medium). has a large VN gain GN (
b i g).

FIG. 7 is a diagram showing the membership function of the degree of VT gain of the consequent part. The horizontal axis in FIG. 7 is the VT gain GT, and the vertical axis is the grade. Here, line S
4 is a membership function indicating that the VT gain GT is small, and the line M4 is the VT gain GT.
The membership function, line B4, indicates that the VT gain GT is medium (medium).
) is a membership function representing that

The fuzzy inference unit 41 evaluates each membership function from the error displacement amount εd1 and the corrected displacement amount εl according to the fuzzy rule shown in FIG. (Figures 4 and 5
figure). From this membership function, the sixth part, which is the consequent,
The membership functions of the VN gain GN in the figure and the VT gain GT in FIG. 7 are evaluated. The interpreter 42 defuzzifies the evaluation value of this membership function using the centroid method to obtain the VN gain GN and VT gain GT.

Tracing control is executed using the VN gain GN and VT gain GT.

Therefore, the optimum values of the VN gain GN and the VT gain GT are always determined, and highly accurate tracking control can be performed.

Note that the above membership function is a rough example, and in reality, a more specific membership function can be determined to perform tracing control. For example, in the above example, each membership function is represented as small (S), medium (M), and large (B).
However, this can also be expressed in 5 or 7 stages.

Furthermore, the membership function can also be trapezoidal or bell-shaped.

Further, in the above description, the error displacement amount and the corrected displacement amount are used as the antecedent condition, but only one of the two can also be used.

Furthermore, as the consequent, VN gain GN and VT gain G
Although T has been determined, it may be useful in some cases to determine only one of them by fuzzy inference and fix the other.

〔Effect of the invention〕

As explained above, in the present invention, the VN gain GN and VT
Since the profiling operation gain such as the gain GT is determined, an optimum gain can be obtained, accurate profiling control is possible, and profiling processing accuracy is improved.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing the configuration of a convention tracing control device implementing the present invention, Fig. 2 is a detailed view of the fuzzy control section, Fig. 3 is a diagram showing fuzzy rules, and Fig. 4 is an antecedent section. Figure 5 is a diagram representing the membership function of the degree of error displacement of the condition, Figure 5 is a diagram representing the membership function of the degree of correction displacement of the antecedent condition, and Figure 6 is the figure of the degree of VN gain of the consequent. Figure 7 shows the membership function of the degree of VT gain of the consequent. Fuzzy reasoning section interpretation section

Claims (6)

[Claims] (1) In a profiling control method that performs machining by following a model, fuzzy inference is performed using the displacement information of each axis obtained during the profiling operation as a fuzzy variable, and the optimal parameter of the profiling operation gain parameter is determined by the fuzzy inference. , A tracing control method characterized in that tracing control is performed using the optimal parameters. (2) The profiling control method according to claim 1, wherein the profiling information includes at least one of an error displacement amount and a corrected displacement amount. (3) The profile control method according to claim 1, wherein the operational gain parameter includes at least one of a normal direction gain and a tangential direction gain. (4) The tracing control method according to claim 1, wherein the sensor is a stylus. (5) The profile control method according to claim 1, wherein the sensor is a non-contact sensor. (6) The tracing control method according to claim 1, wherein the fuzzy control is processed by a processor that performs tracing control.