JP3227266B2 - Numerical control unit - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a numerical control device for performing servo control, and more particularly to a numerical control device suitable for a control target such as a machine tool which requires proper lost motion correction.

[0002]

2. Description of the Related Art In a machine tool to which one of numerical control devices is applied, it is always required to improve both machining accuracy and productivity. Since a lost motion is generated in a feed drive mechanism mounted on a machine tool and adversely affects the quality of machined processing, measures for avoiding this have been considered. As shown in FIG. 9, the drive load system of a machine tool or the like to be controlled by the numerical control device includes play and play of the system, loosening of a bearing or a bearing support (not shown), play of a ball screw end, play of a ball screw 8a and a nut. Backlash in a broad sense such as play between the balls, expansion and contraction and torsion of the ball screw 8a depending on the frictional force and rigidity of the system, elastic displacement of the nut in the axial direction, bearing displacement for supporting the ball screw 8a, sliding between the bed 8c and the table 8b. There are so-called lost motion factors such as dynamic resistance. Of these, the elastic deformation of the ball screw 8a and the like changes depending on the position of the table 8b and the weight of the workpiece mounted on the table 8b.
The sliding resistance between b and the friction coefficient of the sliding surface greatly depends on the feed speed or the lubrication state of the sliding surface. FIG. 5 is a diagram showing the relationship between the lost motion amount and the feed speed Vel. As the feed speed increases, the amount of lost motion tends to decrease. Fig. 6 shows the relationship between the length Pos from the motor shaft end to the machine movable part (nut part) and the amount of lost motion.
Shown in It is known that the elastic deformation such as "torsion" or "bending" of the ball screw 8a or the like is a function of the length from the driving portion to the position of the nut portion, and FIG.
It is a graph of the following formula. FIG. 7 shows an example of the relationship between the lubrication state of the sliding surface and the lost motion amount. Oil lubrication is performed on sliding surfaces of machine tools and the like to prevent seizure and stably guide movable parts. For this lubrication, parameters such as the lubrication interval time Tlint, which represents the time interval of lubrication operation, and the lubrication time Tlon, which determines the time during which the shaft lubrication motor is operated and the lubrication oil is actually ejected to the sliding surface, are set in the numerical controller. By doing so, timer operation is performed. As shown in FIG. 7, when the shaft lubrication motor operates and the lubricating oil is output to the sliding surface, the lost motion amount decreases, and after the lubricating oil output (lubrication) stops, the lost motion amount gradually increases. You can see it going on. In the example of a medium-sized NC lathe, FIG.
The maximum value Lmax and the minimum value L of the lost motion amount at
In some cases, the difference Lw from the minimum value may be more than 10 μm. FIG. 7 shows that the lost motion amount depends on the lubrication interval time, the lubrication time, the elapsed time from the stop of lubrication, and the like.
It shows that it changes in a complicated way.

In order to cope with the dynamic change of the lost motion amount as described above, a fuzzy calculation is executed from the measured data of the speed, position and shaft lubrication state which are measured at any time during the operation of the machine, and the lost motion of the shaft is executed. A numerical control device that estimates a motion amount and calculates a lost motion correction amount from the estimated amount has been devised. In fuzzy control, qualitative control knowledge is expressed as if-th
Express by en rule.

(Equation 1) In the ith fuzzy inference process, the input Xi = (xi
, Xir), i = 1, 2,..., N, the estimated value output yi * by fuzzy inference is calculated as shown in Equation 2.

(Equation 2) In Expression 2, μs (Xi), s = 1,... R are membership values of the antecedent part, and are given by Expression 3.

(Equation 3) Here, f1,..., Fr in Expressions 1 and 3 are constants, which is a simple fuzzy inference.

FIGS. 8 to 14 show an embodiment of a conventional numerical controller having a lost motion correction function using simplified fuzzy inference. FIG. 9 shows a conventional lost motion correction function using simplified fuzzy inference. FIG. 2 is a block diagram illustrating a configuration of a numerical control device having the same. The operation of the prior art will be described below with reference to FIG. In the figure, an NC not shown
On the basis of the NC program 1 read and interpreted by the program reading unit and the NC program interpreting unit,
The function generator 2 calculates the movement information (target position, speed, etc.) of the axis and transfers a position command to the axis controller 3. The axis control unit 3 includes a main control unit 31, a control program storage unit 32, an output unit 33, an input unit 34, a lost motion correction amount calculation unit 35, a control rule storage unit 36, a membership function storage unit 37, and a consequent unit constant storage. It comprises a part 38. The main control unit 31 includes the position command, a servo control program stored in the control program storage unit 32,
Based on the position detection value of the motor position detector 7 which detects the rotor position of the servo motor 5 input through the input unit 34, the control loop of the desired position, speed and current is calculated, and finally Is for transferring a current command value to the power amplifier 4 via the output unit 33. Power amplifier 4
Then, the current command value (P
WM command), performs PWM processing by a known technique,
Each phase voltage to be applied to the servo motor 5 is generated. By applying this voltage, a driving torque is generated in the servo motor 5, and the driving load system including the ball screw 8a, the table 8b, and the bed 8c is relayed through the coupling 6 to a desired position.
Drive at speed.

In the following description, each variable and mathematical expression will be described without the suffix i indicating that it is the i-th fuzzy inference process. In the embodiment according to the prior art shown here, the following five variables are adopted as input variables X of the fuzzy control rule. That is, X = (x1, x2, x3, x4, x5). Here, x1: feed rate Vel [mm / min] x2: distance from motor shaft end to machine movable part Pos [m
m] x3: Lubrication interval time indicating the time interval of lubrication operation
Tlint [minutes] x4: Lubrication time that determines the time for actually ejecting the lubricating oil to the sliding surface Tlon [minutes] x5: Elapsed time after stopping lubrication Tidle [minutes] The control rule storage unit 36 stores R1 to R1 in FIG. A control rule for lost motion correction processing from the viewpoint of the feed speed as shown by R3, a control rule for lost motion correction processing from the viewpoint of the position of the machine movable part from the motor shaft end as shown by R4 to R6 in FIG. Control rules for the lost motion correction process from the viewpoint of the shaft lubrication state as indicated by R7 to Rr in FIG. 8 are stored. These rules are determined empirically, and the antecedent is a qualitative expression. As described above, there are five input variables for the entire control rule, but there are three or less for each rule. That is, the input variable in the rules R1 to R3 is one variable of X = (x1 = Vel), the input variable in the rules R4 to R6 is one variable of X = (x2 = Pos), and the rule R
The input variables in 7 to Rr are X = (x3 = Tlint, x
4 = Tlon, x5 = Tidle). Accordingly, the antecedent fuzzy set F in each rule
sj, s = 1,... r, j = 1,... 5 and antecedent part membership value μs (Xi), s = 1,.
r is specifically given by Equation 4.

(Equation 4) The nonlinear equation f in the consequent part of each rule in FIG.
s, s = 1,... r is a constant f in the simplified fuzzy inference
s, s = 1,... r, which are stored in the consequent part constant storage unit 38 as a constant table.

FIG. 10 shows the membership function storage unit 37.
14 are stored. FIG. 10 shows the membership functions PB (NB), PM (NM), and PS (NS) for the feed speed Vel. The membership function P shown in FIG.
B (NB) is a high-speed membership function, FIG.
The membership function PM (NM) shown in FIG. 10B is a medium-speed membership function, and the membership function PS (NS) shown in FIG. 10C is a low-speed membership function. Next, FIG. 11 shows the membership functions LG, MD, and ST for the position of the machine movable section from the motor shaft end (the distance from the motor shaft end to the machine movable section) Pos, and FIG. The membership function LG shown is a membership function at a position far from the motor shaft end, and FIG.
The membership function MD shown in FIG. 11B is a membership function at a position midway from the motor shaft end, and the membership function ST shown in FIG. 11C is a membership function at a position near the motor shaft end. FIG. 12A shows membership functions SI, MI, and BI for a lubrication interval time Tlint representing a lubrication operation time interval. The membership function SI is a membership function for a short lubrication interval time. The membership function MI is a membership function for a medium lubrication interval, and the membership function BI is a membership function for a long lubrication interval. FIG.
3 (a) shows the membership functions SO, MO, and BO for the lubrication time Tlon, which determines the time during which the shaft lubrication motor is operated to actually eject the lubricating oil to the sliding surface. SO indicates a membership function with a short lubrication time, a membership function MO indicates a membership function with a medium lubrication time, and a membership function BO indicates a membership function with a long lubrication time. FIG. 14 (a) shows the membership functions SD, MD,
The membership function SD is a membership function having a short elapsed time, the membership function MD is a membership function having a middle elapsed time, and the membership function BD is a membership function having a long elapsed time. Indicates a membership function.

Next, the operation of the lost-motion correction processing of the apparatus configured as described above will be described. Now, it is assumed that a movement command requiring a lost motion correction process has been transferred from the function generator 2 to the main controller 31 in the axis controller 3. At this time, the main control unit 31 obtains the current moving direction and feed speed from the previous movement command and the current movement command, and transfers these information to the lost motion correction amount calculation unit 35. Further, the current position data of the movable part by the motor position detector 7 is input via the input part 34 to the setting part / not shown in the numerical control device (not shown) for the lubrication interval time / lubrication time and the elapsed time from the stop of lubricating oil supply. The data sequentially updated by the timer is transferred to the lost motion correction amount calculation unit 35. Upon receiving these data, the lost motion correction amount calculation unit 35 first obtains a membership value from each membership function corresponding to each control rule antecedent in FIG. That is, the control rules R1 to R3 in FIG.
Membership values μ1, μ2, μ3 corresponding to the antecedent part of
From the membership functions shown in FIGS. 10A, 10B, and 10C, the membership values μ4, μ5, and μ6 corresponding to the antecedents of the control rules R4 to R6 in FIG.
It is obtained from the membership functions shown in (a), (b) and (c) according to each input variable. Also, the membership values μ7 to μr corresponding to the antecedents of the control rules R7 to Rr in FIG. 8 are μs = μs3 × μs4 × μs5, s = 7,.
, R, the respective membership values μs3, μs4, and μs5 must be determined from the respective membership functions. For example, the membership value μ7 corresponding to the antecedent part of the control rule R7 in FIG. 8 is obtained by calculating the membership value μ73 for the lubrication interval time from the membership function shown in FIG. Membership value μ74 for lubrication time from ship function
From the membership function shown in FIG. 14 (b), a membership value μ75 for the elapsed time from the stoppage of lubricating oil supply, and the product of these are calculated. When each membership value is obtained in this way, the consequent constant fs s = of each control rule shown in FIG.
,..., R are read from the consequent part constant storage unit 38, and a lost motion amount estimated value yi * in the i-th process is obtained based on Expression 2. When the estimated lost motion amount yi * is obtained as described above, the lost motion correction amount calculation unit 35 determines the final lost motion correction amount Lca by multiplying by a correction coefficient as needed. The lost motion correction amount calculation unit 35 calculates the correction amount data Lca
Is transferred to the main control unit 31, and the lost motion correction amount Lca is added to the position command data at a predetermined timing in the position control loop in the main control unit 31 to correct the movable portion movement amount.

[0008]

The fuzzy control has a problem of tuning. Tuning refers to the work of changing and reconfiguring fuzzy inference rules so that the difference between the inference value and output data is minimized after the production rules are constructed. Specifically, this is an operation of redefining the shape of the membership function in the control rule, the non-linear equation of the consequent part (constant in this embodiment), and the like according to the control target. The absolute value of the lost motion amount of a drive control target of a machine tool or the like largely and complicatedly depends on various factors such as a physical size of a machine drive unit, a guide system, a feed speed, and a guide surface lubrication state. Of course, if the specifications of the machine to be drive-controlled change, even if the machines with the same specifications change, or if the installation environment and driving conditions change even with the same machine, the same machine Even if the installation environment and the driving conditions are the same on the table, the amount of lost motion changes in a complicated manner due to aging. Specifically, the amount of lost motion varies, for example, in a range of about 15 to 35 μm for a medium lathe, and about 40 to 100 μm for a slightly large horizontal machining center. Therefore, at the stage of shipping a machine equipped with a numerical controller according to the prior art, the center value and width of the membership functions shown in FIGS. 10 to 14 are specifically determined based on the actually measured data, and the membership function storage unit It must be set to 37. Similarly, the consequent part constant in FIG. 8 also needs to be specifically determined based on the actually measured data and set in the consequent part constant storage unit 38. The analysis of these measured data and the determination of each fuzzy parameter based on them are performed manually by the designer, and there is a problem in that considerable man-hours must be generated each time the machine specifications change and must be dealt with. . Furthermore, as described above, even if the machine has the same specification, if the machine is different, the amount of lost motion is different, and even if the end user intends to change the fuzzy parameter in order to cope with aging, a means is provided. There is a problem that the lost motion correction processing may not work effectively because there is no such function. It is virtually impossible for a manufacturer's designer to substitute for this, because the lost motion measurement data used to determine the fuzzy parameters of the control target machine is not stored.

[0009]

The present invention SUMMARY OF] has been made to solve the problems described above, estimated using fuzzy inference lost motion amount of the shaft from the data from time to time measured when the machine operation However, in a numerical control device that performs a desired position control by calculating a lost motion correction amount from the lost motion amount estimated value, a semi-cloth immediately after inversion of the current position data, which is measured at any time during machine operation , is performed.
Current loop position and full closed loop current position
Is used as lost motion data, a difference between the lost motion data and the estimated lost motion amount is obtained, and an error back propagation learning model is used based on the difference .
There are those characterized by comprising a fuzzy parameter learning unit for learning the fuzzy parameters in fuzzy control rules.

As described above, in the present invention , the fuzzy parameters are learned based on the lost motion measurement data and are automatically updated. Therefore, it is not necessary to generate the fuzzy parameters for each machine specification each time, and the optimum fuzzy parameters can be quickly obtained. Can be determined.

[0010]

FIG. 1 to FIG. 4 show an embodiment of a numerical controller according to the present invention. The functions and processing contents of the components denoted by the same reference numerals as those of FIG. Since they are the same, a description thereof will be omitted, and only essential parts of the present invention will be described below. In the present invention, the above-described lost motion correction processing operation in the related art is interpreted as a forward processing operation of a neural network as follows. That is, as shown in FIG. 3, the lost motion correction processing function of the prior art has one input layer (the number of neurons = 5), one intermediate layer (the number of neurons = r), and one output layer (the number of neurons = 1). Is regarded as a neural network having the following configuration. Each neuron in the input layer has an input variable vector X = (x1, x2, x3, x4, x5;
x1 = feed speed Vel [mm / min], x2 = distance Pos [mm] from motor shaft end to machine movable part, x3 = lubrication interval time Tlint [min] representing time interval of lubrication operation,
x4 = Lubrication time Tlon [minutes] to determine the time for actually ejecting the lubricating oil to the sliding surface, x5 = Elapsed time T after lubrication stop
Each element variable of idle [minutes] is input, and the fuzzy set matching degree of each fuzzy control rule antecedent in FIG. 8 for the input variable is transferred as an output to each neuron in the hidden layer. For each neuron in the hidden layer, the membership value μs related to the fuzzy control rule antecedent in FIG.
(Xi), s = 1,..., R are calculated according to Equation 4, and the output layer neurons are output to the output layer neuron together with the fuzzy control rule consequent constants fs, s = 1,. Transfer to neuron. In the output layer neuron,
By substituting the membership value and consequent part constant for each antecedent part of each fuzzy control rule in FIG. 8 transferred from each neuron in the hidden layer into Equation 2, the lost motion estimation value yi * in the i-th fuzzy inference processing is obtained. Is calculated. In other words, the lost motion correction processing using fuzzy inference is the forward processing of the neural network shown in FIG. 3, and the coupling coefficient between the input layer and the intermediate layer is based on the input variable and the fuzzy control rule antecedent fuzzy part. The degree of agreement with the set, or membership function,
The coupling coefficient between the hidden layer and the output layer can be considered as a consequent constant. Therefore, in order to execute the lost motion amount estimating process so as to be adapted to the mechanical specifications, operating conditions, aging, etc. of the machine and to produce an effect, the input layer, the intermediate layer, the intermediate layer and the output in FIG. What is necessary is just to prepare the means which can tune the coupling coefficient between layers according to the actual situation of the control object machine.

FIG. 1 is a block diagram showing the configuration of a numerical controller according to the present invention. In FIG. 1, a component different from the prior art is a fuzzy parameter learning unit 40. It learns the coupling coefficient between the intermediate layer and the output layer. As parameters for specifically determining the triangular membership function of the antecedent part of the fuzzy control rule, attention is paid to the center value asj and its width bsj of the triangular membership function shown in FIG. In the fuzzy parameter learning unit 40, the center value asj and the width b
sj, s = 1,..., r, j = 1,..., 5 and the consequent constants fs, s = 1,.
r is adjusted by an error back propagation learning model using the steepest descent method. However, this adjustment is based on the estimated value yi * of the fuzzy inference when the input data Xi in the i-th processing is input.
And the corresponding lost motion measurement data yi, that is, the estimation error E is calculated as shown in Expression 5, and the calculation is performed so as to minimize this.

(Equation 5) That is, the teacher data at the time of update operation in the neural network of FIG. 3 is the lost motion measurement data yi. Here, lost motion data y to be measured from time to time
The main control unit 31 adds the lost motion correction amount Lca obtained based on the i-th lost motion estimated value yi * to the position command data at a predetermined timing in the position control loop. The difference data between the semi-closed-loop current position and the fully-closed-loop current position at the timing when the semi-closed-loop current position data is first inverted after the input becomes valid. The vector indicating the direction in which the estimation error E decreases most is represented by Expression 6.

(Equation 6) FIG. 4 shows a procedure for learning and updating fuzzy parameters in the fuzzy parameter learning unit 40. Hereinafter, the procedure of learning and updating fuzzy parameters in the fuzzy parameter learning unit 40 will be described with reference to FIG. First, i
Input data Xi = (xx) necessary to calculate the lost motion amount estimated value yi * in the th fuzzy inference process
1, x2, x3, x4, x5; x1 = feed speed Vel [m
m / min], x2 = distance from motor shaft end to machine movable part
Pos [mm], x3 = Lubrication interval time Tlint [min], representing the time interval of the lubrication operation, x4 = Lubrication time Tlon, which determines the time during which lubricating oil is actually jetted to the sliding surface
[Minutes], x5 = elapsed time since refueling stop (Tidle [minutes])
Are taken in (step S1). Next, the i-th fuzzy inference process is performed based on the information stored in the control rule storage unit 36, the membership function storage unit 37, and the consequent part constant storage unit 38 using the input variables taken in step S1. Lost motion estimation value y
i * is calculated (step S2). Next, the consequent constant fs of each fuzzy control rule is calculated by using the steepest descent method by using Equation 7.
, S = 1,..., R, and the consequent part constant storage part 3
Set to 8.

(Equation 7) Gf in Equation 7 is a learning coefficient, and the superscript t is
This indicates that the data is obtained by the second update (step S3). Further, the central value asj of each fuzzy control rule antecedent triangular membership function and its width bsj, s = 1,.
= 1,..., 5 are updated and set in the membership function storage unit 37.

(Equation 8) Ga and Gb in Expression 8 are learning coefficients, and a superscript t
Indicates that the data is obtained by the t-th update (step S4). When the learning update of the fuzzy parameters and the resetting to the respective storage units are executed by the fuzzy parameter learning unit 40 as described above, the lost motion of the actual control target machine is performed in the subsequent lost motion amount estimation processing. Estimation processing is performed using fuzzy parameters updated to match the occurrence situation. If the time factor for determining the lost motion amount of the control target machine is, for example, related to the lubrication state of the guide surface, the time factor changes in the order of minutes to hours as shown in FIG. Since the change of the dynamic factor is slow, the learning update of the fuzzy parameter and the resetting process to each storage unit may be executed in a period of several tens of seconds to several tens of minutes, and the CPU processing load per unit time is large. Does not affect.

The present invention is not limited to the embodiment of the present invention shown in FIGS. 1 to 4 described above, and various variations should be considered without departing from the gist of the present invention. For example, the fuzzy parameter learning unit 40 does not always perform the fuzzy parameter learning update and the resetting process to each storage unit, but always executes the t-th time before the execution of step S1 in FIG. It is checked whether or not the difference between the value of Equation 5 of Equation 5 and the value of Equation 5 in the previous t-1 time is within a certain value ε as shown in Equation 9 below. If Equation 9 is satisfied, A numerical control device that does not perform the learning update of the fuzzy parameter and the resetting process to each storage unit can be considered.

(Equation 9)

[0013]

As described above, in the numerical controller according to the first and second aspects, the analysis of the actually measured data such as the determination of the center value and the width of the membership function or the determination of the consequent constant of the fuzzy control rule. In addition, it is possible to solve the problem that considerable man-hours must be generated each time the machine specifications are changed to cope with the determination of each fuzzy parameter based thereon. Furthermore, since the fuzzy parameters are learned based on the lost motion measurement data that is the source of the fuzzy parameter determination of the control target machine and automatically updated as necessary,
Even if machines of the same specification are used, there is an excellent effect that it is possible to quickly respond to the problem that the amount of lost motion is different if the machines are different, and to avoid time-consuming human adjustment.

[Brief description of the drawings]

FIG. 1 is a configuration block diagram of a numerical control device according to an embodiment of the present invention.

FIG. 2 is a diagram showing a triangular membership function.

FIG. 3 is a diagram illustrating a lost motion amount estimation calculation and a fuzzy parameter learning calculation by a neural network.

FIG. 4 is a diagram showing a procedure for learning and updating the consequent constant of the fuzzy control rule and the center value and width of the triangular membership function by the steepest descent method.

FIG. 5 is a diagram showing a relationship between a feed speed and a lost motion amount.

FIG. 6 is a diagram showing a relationship between a position of a machine movable portion from a motor shaft end and a lost motion amount.

FIG. 7 is a diagram showing a relationship between a lubrication state and a lost motion amount.

FIG. 8 is an explanatory diagram showing a fuzzy control rule.

FIG. 9 is a configuration block diagram of an embodiment of a numerical control device according to the related art.

FIG. 10 is a diagram showing each membership function for a feed speed.

FIG. 11 is a diagram showing each membership function with respect to the position of the machine movable section from the motor shaft end.

FIG. 12 is a diagram showing each membership function for a lubrication interval time.

FIG. 13 is a diagram showing each membership function for lubrication time.

FIG. 14 is a diagram showing each membership function with respect to the elapsed time from the stop of refueling.

[Explanation of symbols]

 Reference Signs List 1 NC program 2 Function generator 3 Axis controller 4 Power amplifier 5 Servo motor 6 Coupling 7 Motor position detector 8a Ball screw 8b Table 8c Bed 31 Main control unit 32 Control program storage unit 33 Output unit 34 Input unit 35 Lost motion correction Amount calculation unit 36 Control rule storage unit 37 Membership function storage unit 38 Consequence part constant storage unit 40 Fuzzy parameter learning unit

Continuation of the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G05B 19/404 G05B 13/02

Claims (1)

(57) [Claims] 1. A desired position control is performed by estimating a lost motion amount of a corresponding axis from data measured at any time during machine operation using fuzzy inference, and calculating a lost motion correction amount from the estimated lost motion amount. In the numerical control device that performs the operation immediately after reversing the current position data, which is measured as needed during machine operation,
Miclosed loop current position and full closed loop current
The difference data from the current position is used as lost motion data.
Between the lost motion data and the estimated lost motion amount, and based on the difference, the error back propagation learning mode is calculated.
Numerical control apparatus characterized by comprising a fuzzy parameter learning unit for learning the fuzzy parameters in fuzzy control rules using Dell.