Research on on-line measurement technology of mach

2022-10-19
  • Detail

Research on machining accuracy measurement technology of ultra precision lathe

1 overview

the goal of machining is to pursue the best combination of machining accuracy, cost and efficiency. In order to achieve this goal, one of the key technologies in urgent need of research and development is machining accuracy measurement technology, especially under the condition of multi variety and small batch production, the research on advanced measurement technology is of great significance, Because measurement is an important part of processing and measurement integration technology, and it is an important means to ensure the quality of parts and improve productivity. Foreign countries have recognized the importance of measurement technology for a long time and have carried out a lot of research, and have been widely used in production practice. The measurement of part machining accuracy can be divided into two cases. One is to directly measure the machined surface of the workpiece in the machining process, and the required accuracy index can be obtained as soon as the machining process is completed [1], which is the most ideal case of measurement; Second, after the processing process, the workpiece is still installed on the machine tool, and the workpiece is measured with a reasonable measuring instrument [2]. In ultra precision machining, the influence of thermal deformation on machining accuracy can not be ignored, so constant temperature oil pouring or cutting fluid cooling is necessary in the machining process. In the case of coolant and high workpiece speed, there is no sensor with a measurement accuracy of 0.01 M. therefore, in ultra precision machining, the detection of part machining accuracy is mainly 4. Pendulum swing radius: r280mm, using the traditional offline measurement method, In many cases, the cost of off-line measurement is equal to or even more than the processing cost of parts. Based on the above reasons, this paper studies the second case to realize the measurement of parts. Its essence is to use the lathe as a coordinate measuring machine. Because the motion accuracy of the moving parts of the submicron ultra precision lathe developed is very high, even higher than that of many measuring instruments and measuring machines, if the machine tool is organically connected with the appropriate measuring instruments, the measurement of the machining accuracy of parts can be realized, so that the machine tool can be used for processing and measurement, which expands the application range of the machine tool and solves the measurement problem of parts [3]. At present, the development trend of machining quality assurance is to make the quality assurance closer to the processing process by replacing off-line measurement and statistical quality control with measurement, and ensure that the parts are qualified when they are unloaded from the processing equipment. Of course, this requires a premise that the efficiency and accuracy of measurement must be guaranteed, so that the comprehensive decision-making and necessary compensation can be achieved within the minimum time delay. Therefore, it is of great practical significance to study the measurement technology of machining accuracy of parts

2 analysis of error sources affecting measurement accuracy

the purpose of measurement is to check whether the accuracy index of the processed parts meets the requirements. If it meets the requirements, remove the workpiece, otherwise carry out necessary compensation processing until the workpiece processing accuracy is qualified. We know that to accurately measure the processing accuracy of parts in recent two years, the accuracy of the measuring equipment must be one order of magnitude higher than the accuracy of the measured parts, that is, the principle of 10 times, In ultra precision machining, there is little difference between the machining environment and the measurement environment. To ensure the measurement accuracy, we can only achieve it through error compensation, that is, measuring the parts without compensation can ensure the measurement accuracy (error compensation can improve the processing accuracy of the parts by an order of magnitude), while measuring the parts with compensation through error compensation cannot meet the principle of 10 times, However, the accuracy of lathe measurement after error compensation is high enough, which is still meaningful. Of course, when the lathe is used as a coordinate measuring machine, the measurement accuracy is also affected by the accuracy of the measurement sensor and the measurement strategy and data processing strategy. In the design and manufacturing process of this lathe (with T-shaped layout), many advanced technologies are adopted to reduce or eliminate the influence of thermal deformation error on the motion accuracy of the lathe. For example, the lathe adopts an Aerostatic Spindle and white dense jade as the material of the spindle and bearing; The lathe slide adopts aerostatic guide rail; The lathe spindle box, slide, bed and guide rail are made of granite; The temperature between machining is controlled at 200.1 ℃, so the error source affecting the measurement accuracy is mainly the geometric error of the machine tool, with a total of 21 items, namely, 6 errors of each moving part and 3 mutual position errors between the three axes. The 21 errors are shown in the table below. Accurate and rapid identification of error sources is the basis for high-precision measurement. Considering that the motion mode of the measurement process and the machining process is very similar (the tool is replaced by a sensor or a measurement probe, of course, the error compensation model of the two is also similar), the influence of error sources in the non error sensitive direction on the measurement accuracy can be ignored, that is, the influence of (x), (x), y (z), (z), y is not considered separately( Ф) And( Ф) Influence on the measurement accuracy, plus XZ does not affect the measurement accuracy (by Z Ф And X Ф Including), during the identification of the machine tool, the rotation accuracy of the spindle was measured, and the measured result was the radial runout error X of the spindle( Ф) And axial displacement error Z( Ф) Both are 0.05 M, which is smaller than the straightness error of the slide (Z (x) 0.18 m/100mm, X (z) 0.20 m/100mm), and the deviation error of the main shaft( Ф) It is also very small, so the influence of spindle rotation error on the measurement accuracy is not considered separately when adopting the cascade tightening mechanism cold technology to measure the machining accuracy of parts. Of course, in order to measure large-size parts (large-size plane mirror) with high accuracy, the rotation error of the spindle (such as( Ф)) Must be considered

list of 21 error sources of lathe

3 error source identification and modeling

this lathe mainly processes cylindrical surface, end face, conical surface, spherical surface and other parts. Without considering the spindle rotation error, when measuring the shape or cylindricity error of cylindrical surface busbar, only X-direction error compensation is required; Z-direction error compensation is required when measuring the end face; When measuring conical and spherical surfaces, Z-direction and X-direction error compensation need to be carried out at the same time. At this time, the error compensation model must be two-dimensional. Of course, three capacitive sensors can also compensate for the impact of spindle rotation error on measurement accuracy. There are usually two methods to identify the error compensation amount. First, identify each error source offline and obtain the error compensation amount of each point in the machining space of the machine tool through a certain synthesis law (such as homogeneous coordinate transformation); The second is to obtain the error compensation amount by measuring the machined surface of parts. The first method is time-consuming and the assumptions in the modeling process affect the modeling accuracy. The second method can only compensate the specific parts measured, and cannot expand the compensation area to the whole processing area. This paper combines the two cases to identify the error compensation amount, which not only ensures the accuracy but also saves time

3.1 error source identification ① identification of X-direction error compensation: there are 6 error sources that affect the measurement accuracy of X-direction, namely, X (z), Z Ф、 (z) , (z), X (x) and (x), the error compensation amount in X direction is expressed as:

x= x (x) + X (z) + Z Ф. z+ (z). WZ+ (z). Wh+ (x). TZ (1)

where Z is the distance from the origin of the slide; Distance from WZ measuring point to suction cup in Z direction; TZ the distance from the upward sensor to the center of gravity of the moving parts of the X slide; Wh the distance between the measuring point on the workpiece and the center of gravity of the Z slide (the headstock is installed on the Z slide) in the vertical direction; In order to identify the error source, turn the cylindrical surface within the processing range of 100mm in the Z direction (the X slide is stationary), and use the sensor to measure the error of the workpiece bus in the opposite direction of the tool, so as to obtain the error compensation column X in the X direction when measuring the cylindrical surface. As shown in Figure 1, there are four error sources that affect the accuracy of the workpiece bus, namely, X (z), Z Ф、 (z) And (z), the error compensation amount is expressed as:

column x= x (z) + Z Ф. z+ (z). WZ+ (z). Wh (2)

in this way, four error sources can be identified at one time. From equations (1) and (2), it is known that x (x) and (x) in X need to be identified, (x) needs to be identified offline with laser measurement equipment, and X (x) is guaranteed by the lathe full closed-loop system. In this way, the x-direction error compensation can be obtained during measurement, that is:

x=f (x, z) = column x+ (x) TZ (3)

② identification of Z-direction error compensation: there are 6 error components that affect the measurement accuracy in Z direction, namely Z (x), (z), X Ф、 (x) , (x) and Z (z), and the error compensation amount is expressed as:

z= Z (x) + X Ф. x+ (x). Th+ (x). TX+ (z). Wx+ Z (z) (4)

where x is the distance from the origin in the X direction (end face center); The distance between the measuring point on the Wx workpiece and the center of the suction cup in the X direction; The distance that the th sensor is vertically away from the center of gravity of the moving parts of the X slide; TX the distance from the X upward sensor to the center of gravity of the moving parts of the X chute

when measuring the end face of the part (the Z slide is stationary), there are four error sources that affect the measurement accuracy of the end face, namely X Ф、 Z (x), (x) and (x), expressed by the formula:

face x= (x) + X Ф. x+ (x). Th+ (x). TX (5)

in order to identify these four error sources, first turn an end face with a diameter of 200mm (the diameter can be appropriately increased), assuming that the end face is symmetrical along the radial direction [3], install the prism behind the tool holder at the same height as the tool and keep it collinear with the tool along the Z direction [4], at the same time, replace the tool with an inductive sensor, first measure the straightness of the X slide several times with a laser, and use the least square method to fit the included angle between the laser and the X slide movement direction, Take the workpiece center as the starting point (based on this point) and measure along the feeding direction:

= transmission direct

type intermediate transmission sensor reading; Straight x chute linear motion error. Fitting a straight line to a sequence yields 2 x Ф, At the same time, the shape error line (x) of the radial bus of the end face is obtained, so the error compensation volume from the center of the workpiece to the outer measuring end face is obtained. Plane z= line (x) + X. X Ф, In order to overcome the influence of random error, multiple turning and multiple measurement are required. It is known from equations (4) and (5): there are still (z) and Z (z) in the z-direction error source that are not identified, (z) offline measurement and identification with laser measurement equipment, and Z (z) is guaranteed by the lathe full closed-loop system. When measuring the end surface, the z-direction error compensation is:

z=f (x, z) = surface z+ (z) TX (6)

3.2 error source modeling considering that as long as the number of hidden layer nodes of BP neural network is enough, any nonlinear function can be fitted with any accuracy, and the interpolation accuracy is very high, this paper uses neural network to establish error compensation model. John Egert first applied this method, but failed to solve the problem of training samples [5]. This paper uses matlab5.1 neural network toolbox to establish a measurement error compensation model, which is very convenient. BP network based on Levenberg Marquardt optimization algorithm is used in all network topology and compensation structures. Tansig function is used in hidden layer and purelin function is used in output layer. ① Modeling of X-direction error compensation: take the position Z of the Z slide as the input of the network, take the column x as the output, and use the neural network a (structure as shown in Figure 2) to fit the function 1x= column x=f (z); Take the position X of the X slide as the input of the network, and take (z) TZ is used as the output, and the neural network B is used to fit the function 2x= (z) Tz=f (x), where 1x= column x=f (z) is the error compensation model when measuring part cylindricity and bus straightness

② modeling of Z-direction error compensation: take the position X of the X slide as the network input and the face Z as the output, and directly use the neural network C to fit the function 1z= face z=f (x); Take the position Z of the Z slide as the input of the network, and take (z) TX is used as the output, and the neural network D is used to fit the function 2z= (z) Tx=f (z), 1z= face z=f (x) is the error compensation model when measuring the flatness of the end face. In actual processing, the frequently used area is 100mmm 100mm. Because the pitch of the driving lead screw is 5mm, and the straightness error obviously has the component of taking the pitch as the cycle. In order to accurately model, the processing interval is divided into 5151 points (to ensure that the sampling points have sufficient density), that is, 2601 training samples are required. The calculation method of training samples is as follows. Calculation of training samples of X-direction error compensation:

Copyright © 2011 JIN SHI