WO2004106895A1 - Fatigue tester - Google Patents

Abstract

There is provided a fatigue tester having an improved accuracy. The fatigue tester (10) includes: a main body (11) having a test piece mounting section (20) and an actuator (40); and control means (60) for performing feedback control for feeding back a detection signal in accordance with a load applied to a test piece (1). In the fatigue tester (10), amplification ratio correction/zero-point correction calculation means (78) compares a waveform of a detected signal to a waveform of a target signal and performs calculations of the amplification ratio and the zero-point correction according to the comparison result. In this calculation, according to the convergence degree of the detected signal to the target value signal, the algorithm of calculation concerning the amplification ratio correction is stepwise switched. Moreover, in addition to the amplification ratio, the algorithm of calculation concerning the zero-point correction may also be switched stepwise.

Description

 Specification

 Fatigue testing machine

 Technical field

 The present invention relates to a fatigue tester, and can be used, for example, for a hydraulic servo fatigue tester for acquiring data in a very high cycle range.

 Background art

[0002] Fatigue accounts for more than 70-80% of the damage to mechanical structures, and countermeasures against fatigue are one of the most pressing issues for mechanical engineers. Generally, in the SN curve of a steel material, a horizontal portion, that is, a fatigue limit appears at a repetition rate of about 10 ^{6 to} 10 ⁷ . Unless it is assumed that repairs and parts replacement for airplanes and nuclear equipment over a certain period of time, general machines are designed based on the fatigue limit.

[0003] However, in recent years, there has been reported a phenomenon in which a horizontal portion does not appear in the SN curve even in a long life region of 10 ^{7 to} 10 ⁸ or more, and a fatigue limit is not recognized in high-strength steel and surface hardened steel. It has become so. This phenomenon is called ultra-high cycle fatigue or ultra-long life fatigue, and active research is currently being conducted to elucidate its mechanism.

 [0004] Test equipment for such ultra-high cycle fatigue is required to have higher performance than conventional fatigue test equipment in repetition rate, load accuracy, load stability, removal of uneven load, and the like. You. Accordingly, the applicant of the present application has used a DSP (Digital Signal Processor) controller to perform digital control and operate the hydraulic servo mechanism to improve load accuracy, load stability, and operability. Has been developed (see Patent Document 1). This axial load fatigue tester secures high responsiveness by increasing the natural frequency of the system by integrating the components of the main body of the tester to increase rigidity, and secures data of ultra-high cycle range. Many acquisitions have been achieved. In addition, a spherical bearing (spherical bearing) is incorporated in the test piece mounting part, and self-alignment is performed when a tensile load is applied to the test piece. In addition, a computer is connected to the DSP controller, and it is possible to display measured values during testing and input control parameters on the screen of this computer.

[0005] In addition, as a hydraulic fatigue tester that performs digital control, for example, a control target output There is a low-cycle fatigue tester or the like in which data sampling is performed every time the value of the force reaches each of a plurality of predetermined levels (see Patent Document 2 and the like). In this low-cycle fatigue tester, the number of data points per cycle of the repetitive change of the controlled object output amount can be made closer to an appropriate number without being affected by the length of the cycle of the repetitive change. : JP 2002-243606 A (FIGS. 1 and 3, paragraphs [0061] [0063], [006 8], [0074] [0076])

 Patent Document 2: JP-A-2000-131203 (FIGS. 1, 7, paragraphs [0007] [0011], [002 6], [0040])

 Disclosure of the invention

 Problems the invention is trying to solve

[0006] Incidentally, since the fatigue tester is affected by various disturbances, the load signal may not follow the target value signal when the amplification factor is set to a constant value. In such a case, the maximum value and the minimum value of the load signal are monitored in real time, and the optimum amplification factor correction amount and zero point correction amount are calculated based on the difference between the monitoring result and the target value signal. Based on the above, a correction operation for disturbance that corrects the amplification factor and the zero point is performed. Such proportional control is sometimes referred to as adaptive proportional control because it changes the amplification factor and the zero point every moment in response to disturbances and the like, and is described in Patent Document 1 described above. This adaptive proportional control is also performed by a fatigue tester.

 [0007] However, even when such adaptive proportional control is performed, if the load signal converges to the target value signal to some extent, the load signal may not converge any more. It is desired to improve the test accuracy by increasing the degree.

 [0008] Further, Patent Document 2 mentioned above discloses an AGC (Auto Gain) for automatically controlling an amplification factor.

The (Control) method is a method in which the magnitude of the amplitude of the target value signal used in the next cycle is corrected based on the amplitude of the control target output amount actually generated in a certain cycle during control execution. It is described that high-precision control cannot be performed from one cycle, and it is difficult to apply it to a low-cycle fatigue test in which the area of the hysteresis curve of the stress-strain curve of the test piece is relatively large. Therefore, as described in Patent Document 2 mentioned above, The low cycle fatigue tester described above does not control the AGC method in the first place, so it does not solve the above problems.

[0009] Furthermore, in the fatigue tester described in Patent Document 1 described above, a measured value during a test (for example, a load applied to a test piece, a load applied to a test piece, or the like) is displayed on a computer screen connected to a DSP controller. The maximum value, the minimum value, the value of the signal of the function generator, etc.) are displayed, but these displays are displayed as load values and voltage values, and the stress display required by the experimenter is Absent. Therefore, the experimenter had to calculate and understand the stress based on the specimen diameter. In addition, control parameters could not be input as stress values, but it was necessary to calculate the load and voltage values to be controlled based on the specimen diameter and input them as control parameters. Furthermore, erroneous recognition or erroneous input hinders improvement of test accuracy. For this reason, it is desirable to create an environment in which the operability and usability of the tester are improved, the labor of the experimenter and the labor of judgment are reduced, and the test accuracy can be more reliably improved.

 [0010] By the way, in research on ultra-high cycle fatigue, the fatigue fracture of high-strength materials of interest is low.

• It is more sensitive to surface flaws and stress concentration of inclusions present on the surface than fatigue fracture of medium strength materials. If an unbalanced load is applied to the test piece, even if it originally breaks internally, it may cause surface destruction. In such a case, the reliability of the test results may be significantly reduced. Therefore, it is necessary to pay close attention to the alignment of the test pieces, and to avoid this, simplify the alignment work (alignment free). ), A tester and a test piece for ultra-high cycle fatigue research require a mechanism to prevent uneven load. Therefore, in the fatigue tester described in Patent Document 1 described above, a spherical bearing is incorporated in the test piece mounting portion so that self-alignment is performed when a tensile load is applied to the test piece. Alignment free.

However, when a spherical bearing is incorporated as described above, the test conditions are limited to a fatigue test in which only a tensile load is applied (tensile-tensile fatigue test), and a fatigue test involving a compressive load is performed. I can't. In fatigue research, the tensile-compression fatigue test with a stress ratio (minimum stress Z maximum stress) of -1 is the basic data, so the axial load fatigue tester used for ultra-high cycle fatigue research has an alignment-free function. And can apply tension-compression load Is required. On the other hand, in order to obtain high-precision data without such a mechanism, it is necessary to attach a strain gauge to the test piece and perform alignment work to prevent uneven load. However, this alignment work needs to be performed for each test, which consumes a great deal of time and effort. Therefore, it is desired to improve the test accuracy by preventing alignment loads while realizing alignment-free.

An object of the present invention is to provide a fatigue tester capable of improving test accuracy.

 Means for solving the problem

[0013] The present invention provides a main body having a test piece mounting part for mounting a test piece and an actuator part for applying a load to the test piece, and a detection signal corresponding to the load applied to the test piece to feed back the actuator. And a control unit for performing feedback control for sending a control signal for controlling the operation of the unit to the main body.The control unit includes a digital controller for performing feedback control by digital processing. This digital camera controller is provided with a control signal generating means for generating a control signal corresponding to a deviation amount between a detection signal and a target value signal, and compares a waveform of the detection signal with a waveform of the target value signal. Based on the comparison result, an amplification factor correction Z for performing a calculation process relating to an amplification factor and a zero correction used for a control signal generation process, and a zero-point correction calculating means, An amplification factor correcting means for performing an amplification factor correction process based on the calculation result by the positive / zero correction calculation device; and a zero point correction device for performing a zero point correction process based on the calculation result by the amplification factor correction / zero correction device. The amplification factor correction / zero correction calculation means is configured to switch the algorithm of the calculation process for the amplification factor correction stepwise according to the degree of convergence of the detection signal to the target value signal. It is characterized by that.

 Here, the “calculation processing for the amplification factor and the correction of the zero point” includes the processing of calculating the amount of increase or decrease with respect to the original value (the value before correction) of the amplification factor and the zero point, and the processing after the correction. It also includes the process of calculating the zero point of and the value of the amplification factor itself (corrected value itself).

[0015] The "configuration for switching algorithms stepwise" includes both a configuration for switching two algorithms and a configuration for switching three or more algorithms. In this case, one-way switching, which does not return to the original algorithm after switching once, And one-way traffic switching that can return to the original algorithm again after one switching.

 Further, “switching” the “algorithm” includes, for example, the following. That is, (1) switching of the form of the mathematical expression used for the calculation process itself is included. For example, a change from a linear function to a quadratic function, a change in the exponent, a change from an addition term to a subtraction term or vice versa, a change in the number of terms, and the like. (2) Switching the values of coefficients and constants included in the formulas used in the calculation process is included. For example, a process of multiplying by 0.1% is changed to a process of multiplying by 0.05%. (3) Switching data and data groups used for calculation processing is included. (4) Switching the structure of selection processing (for example, processing of IF statements, etc.) and branch processing (for example, processing of CASE statements, etc.) are included. For example, the order and position of the selection processing in the program are changed, the number of branches is increased / decreased, and the degree of layering of the selection processing and the branch processing is changed. (5) This involves switching to a completely different algorithm. For example, switching from an algorithm that performs a calculation process using a mathematical expression to an algorithm that performs a data selection process, or from an algorithm that performs a calculation process using a mathematical expression, to a routine having a different processing content (for example, a routine using a different mathematical formula) Switching to an algorithm that performs selection processing of different data, a routine that uses different data, etc.).

 In such a fatigue tester of the present invention, the amplification factor correction / zero correction calculation means compares the waveform of the detection signal with the waveform of the target value signal, and based on the comparison result, determines the amplification factor and When performing the calculation process related to the correction of the zero point, the algorithm of the calculation process related to the correction of the amplification factor is switched stepwise according to the degree of convergence of the detection signal to the target value signal.

 [0018] For this reason, when the detection signal converges to the target value signal to some extent, by switching the algorithm related to the correction of the amplification factor, it is possible to avoid the state where the convergence reaches a plateau, and to switch the detection signal to the target value signal. Can be increased. Therefore, the test accuracy is improved, and the object is achieved.

Further, in the fatigue tester described above, the amplification factor correction Z zero point correction calculation means switches the algorithm of the calculation process for the zero correction stepwise according to the degree of convergence of the detected signal to the target value signal. It is desirable that it is.

[0020] A configuration in which the algorithm of the calculation process relating to the correction of the zero point is switched stepwise as described above In this case, the algorithm can be switched not only for the amplification factor but also for the calculation processing for the correction of the zero point, so that the degree of convergence of the detection signal to the target value signal is further increased, and the test accuracy is further improved. It becomes possible.

 [0021] Further, in the fatigue tester described above, the amplification factor correction Z-zero correction calculating means converts the waveform of the periodically changing detection signal and the waveform of the periodically changing target value signal into one waveform. Comparing every cycle, based on the amount of deviation of both waveforms obtained as a result of this comparison, perform calculation processing for the amplification factor and zero point correction for each waveform cycle, and switch the algorithm when performing this calculation processing It is desirable that the configuration be such that it is determined at each cycle of the waveform.

 [0022] In the case where the determination as to whether or not to switch the algorithm is made in each cycle of the waveform, the algorithm can be switched at an appropriate timing, the convergence speed is increased, and the test is performed. The accuracy is further improved.

In the specification of the present application, the control that involves the stepwise switching of the algorithm of the calculation processing relating to the correction of the amplification factor and the zero point as described above is called multi-mode adaptive proportional control (MAP). Shall be. Such multi-stage adaptive proportional control can be applied to other controlled devices other than the fatigue tester of the present invention. That is, the output signal (corresponding to the detection signal corresponding to the load applied to the test piece) from the controlled device (corresponding to the body of the fatigue tester) is fed back and the controlled device is controlled. In order to realize a feedback control system that sends a control signal to control a device to be controlled, a digital controller that performs digital feedback control is provided. A control signal generating means for generating a control signal corresponding to the deviation amount; a comparison between a waveform of the output signal and a waveform of the target value signal; and, based on the comparison result, an amplification factor and a zero point used for the control signal generation processing. A gain correction / zero correction calculation method that performs calculation processing related to correction, and an amplification correction based on the calculation result by the amplification correction Z zero correction calculator A gain correction means for performing processing, and a zero point correction means for performing zero point correction processing based on a calculation result by the gain correction / zero point correction calculation means. The amplification factor according to the degree of convergence to the target value signal The algorithm of the calculation process may be switched stepwise. Further, in addition to the amplification factor, a configuration may be adopted in which the algorithm of the calculation process relating to the correction of the zero point is switched stepwise.

 Then, the present invention feeds back a detection signal according to a load applied to the test piece, and a main body having a test piece mounting portion for mounting the test piece and an actuator unit for applying a load to the test piece. The fatigue tester includes a control unit that sends a control signal for controlling the operation of the actuator unit to the main body and a control unit that performs feedback control.The control unit includes a digital controller that performs feedback control by digital processing. And an external processing device connected to the digital controller for transmitting and receiving information to and from the digital controller to perform processing other than processing related to feedback control. The process of displaying the standby screen used when operating the actuator unit in a state other than medium, and using this standby screen Display of a standby screen for receiving the operation input to the actuator section by the experimenter performed by the user.Process for displaying the input reception processing means, the test condition setting screen for setting the test conditions, and this test condition setting screen. A test condition setting screen display that accepts the test condition setting input by the experimenter by using the input input processing means, a process that displays a test-in-progress screen that monitors the test status during the test, and this test. It is characterized by comprising: a screen display during a test for performing a process of receiving an input by an experimenter performed by using a middle screen; and an input reception processing means.

 Here, “processing related to feedback control” includes, in addition to the processing of the main feedback loop, for example, the above-described calculation processing relating to the correction of the amplification factor and the zero point, and the correction processing of the amplification factor and the zero point. Is included.

 [0026] In such a fatigue tester of the present invention, the display processing of three screens, a standby screen, a test condition setting screen, and a test screen, is performed by an external processing device provided separately from the digital controller. And an input receiving process using these screens. Therefore, these screens are switched for each situation of the test machine operation, that is, for each scene up to the test preparation stage and the actual stage.

[0027] For this reason, in each scene of the test and its preparation, the experimenter has to use any of the three screens. While referring to these, check only the display of information required in each scene, and perform only the input work required in each scene. Therefore, the experimenter only needs to perform the minimum necessary operations in each stage of the test and its preparation, and eliminates the need to perform extra operations or refer to extra information and think about extra things. As a result, operability is improved. Since the operability is improved, the occurrence of erroneous recognition or erroneous operation is avoided or suppressed, which leads to an improvement in test accuracy.

 [0028] In addition to the standby screen display, the input reception processing means, the test condition setting screen display, the input reception processing means, and the screen display during the test. It is also possible to provide reciprocal operation forced invalidation means for forcibly invalidating operations that should or should not be performed, or input parameter auditing means for auditing whether parameters (test conditions etc.) input by the experimenter are abnormal. Providing these blocks prevents erroneous operation, which leads to further improvement in test accuracy.

 [0029] In addition, standby screen display · input reception processing means, test condition setting screen display · input reception processing means, and screen display during test · input reception processing means are all provided in the external processing device, the digital controller is However, the processing for displaying three screens and the processing for accepting input on three screens performed by these means are not burdened. In other words, all functions of the man-machine interface between the experimenter and the test machine are realized by the external processing device, and the digital controller bears only the arithmetic processing directly related to the control performance. As a result, the processing load on the digital controller is reduced, and the speed of arithmetic processing directly related to control performance such as responsiveness, load accuracy, and convergence to disturbances is improved. The above objectives are achieved by these.

 [0030] In the fatigue tester described above, a test condition setting screen display / input reception processing means sets a combination of a maximum stress and a minimum stress, a combination of a maximum stress and a stress ratio, as a test condition on the test condition setting screen. It is configured to accept the input of any combination of the minimum stress and the stress ratio and the diameter of the test piece. It is desirable to have a configuration to display.

[0031] Thus, the input of the stress “stress ratio” based on the test piece diameter on the test condition setting screen and When the tester is configured to display a graph of the stress on the screen during the test, the experimenter only needs to use the commonly used parameters of stress, stress ratio, and specimen diameter. The performance is improved. In addition, since the stress display on the screen during the test is graphed, the experimenter can intuitively grasp the test status, and the operability of the test machine is further improved. For this reason, even an experimenter who does not have knowledge of a fatigue testing machine or control can easily perform a fatigue test.

 As described above, a program for causing a computer to function as the external processing device is also a distribution target and a transaction target even in the program itself, as described below. That is, as the external processing device described above, a program or a part thereof for causing a computer to function is, for example, a read-only memory (M メ モ リ) or a compact disk (CD) using a read-only memory (CD). CD-ROM), CD recordable (CD_R), CD rewritable (CD-RW), read-only memory (DVD-ROM) using digital 'versatile' disc (DVD), random 'access' memory using DVD ( DVD—RAM), flexible disk (FD), magnetic tape, hard disk, read-only memory (ROM), electrically erasable and rewritable read-only memory (EEPROM), flash 'memory, random' access' memory (RAM) It is possible to record and store and distribute on a recording medium such as a local area network (LAN), metropolitan area network, etc. It can be transmitted using a transmission medium such as a park (MAN), a wide area network (WAN), a wired network such as the Internet, an intranet, an extranet, or a wireless communication network, or a combination of these. It is also possible to carry on a carrier wave. Further, the program described above may be a part of another program, or may be recorded on a recording medium together with a separate program.

[0033] Further, the present invention includes an actuator section for applying a load to the test piece, and a test piece mounting section for mounting the test piece in a state of being arranged in a direction along the load direction of the actuator section. The grips are provided with grips for gripping both ends of the test piece, and each of these grips is formed with a test piece contact surface with which both end faces of the test piece abut. In a fatigue tester, the end faces on both sides of the test piece satisfy the parallelism of 0.01 and the test piece abuts on each of the grips facing each other among the faces of the component parts of the tester main body. All the surfaces that play a role in defining the parallelism between the surfaces are all finished so as to satisfy the parallelism of 0.01, and each end of the test piece with the test piece attached to the test piece mounting part. A gap is formed between the outer peripheral surface of each of the grips and each of the gripping portions.

 Here, the “parallelism” is the parallelism in drafting science (parallelism described in the production drawing), that is, the parallelism of the plane portion with respect to the reference plane. “Parallelism 0.01” means that the space between two planes that are parallel to the reference plane and that have a distance of 0.01 mm is allowed. The reference plane may be, for example, a surface on the opposite side of each component.

 Further, “each surface that plays a role in defining the parallelism between the test piece abutting surfaces of the opposing gripping portions” refers to a case where a plurality of test machine main body components are assembled to form a test machine main body. Next, the surfaces that will affect the relative posture of the test piece contact surfaces of each gripping portion as a result are described.

 [0036] In this way, when the parallelism of each surface of the test piece and the components of the testing machine main body is improved, and when the fitting between the end portion of the test piece and the grip portion is loosened, the parallelism is improved. Because of this, the occurrence of unbalanced load is suppressed, and since the loose fit is applied, the restraining force from the side surface (outer peripheral surface of the end of the test piece) is not applied to the test piece. It is possible to suppress the occurrence of eccentric loads. Therefore, the test accuracy is improved. In addition, loosening the fit usually improves the parallelism of each surface of the test specimen and the components of the test machine main body, which may cause lateral displacement of the test specimen during the fatigue test. Does not apply any force that causes lateral displacement.

 Further, in the above-described fatigue testing machine, the gap between the fitting of the outer peripheral surface of each end of the test piece used in the actual test and each grip portion is a dummy for alignment used in the alignment work performed before the test. It is desirable that the gap be larger than the clearance between the outer peripheral surface of each end of the test piece and each gripping portion.

[0038] In the case where the alignment can be performed using the dummy alignment test specimen having a larger end than the test specimen used in the actual test, the alignment before the test is performed. After performing alignment work using the core dummy test piece, remove the core alignment dummy test piece and attach a test piece to be used in the actual test. This makes it possible to perform alignment work with high accuracy. In addition, the loose fitting described above can be easily realized. In addition, if the alignment work is performed once using the alignment dummy test piece, it is possible to perform a plurality of tests as it is, thereby reducing the labor of the experimenter.

 Other objects, features and advantages of the present invention will become apparent from the following description of embodiments of the present invention with reference to the accompanying drawings.

 BEST MODE FOR CARRYING OUT THE INVENTION

 Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 shows the overall configuration of a fatigue tester 10 of the present embodiment. FIG. 2 is a cross-sectional view of a test piece mounting portion 20 constituting the main body 11 of the fatigue tester 10, and FIG. 3 is an enlarged cross-sectional view of a main part of the test piece mounting portion 20. 4 to 6 show examples of screens associated with the processing by the external processing device 90 constituting the control means 60 of the fatigue tester 10. The fatigue tester 10 is a fatigue tester for acquiring ultra-high cycle range data.

 In FIG. 1, a fatigue tester 10 includes a main body 11 for attaching a test piece 1 and applying a load to the test piece 1, and a control means 60 for controlling an operation for applying a load on the main body 11. It is configured with.

 The main body 11 includes a test piece mounting section 20 for mounting the test piece 1, an actuator section 40 for applying a load to the test piece 1, and a servo valve 50 for switching a hydraulic pressure supplied to the actuator section 40. It is provided with.

 In FIG. 2, a test piece mounting portion 20 is connected to the upper end 1 A of the test piece 1 via a plurality of components, and detects a load applied to the test piece 1, and a load sensor 21. , A load cell holder 23 for holding the load cell 21, an upper (load cell 21 side) gripper 24 for gripping the upper end 1A of the test piece 1, and a lower side of the test piece 1 A lower (piston 41 side) grip portion 25 for gripping the end portion 1B.

 The load sensor 21 is fixed to the upper surface 23 A of the load cell holder 23 by a plurality of bolts 26. The load cell 21 undergoes minute deformation by receiving a force from the test piece 1, and the strain at that time is detected by the strain gauge 22, whereby the load applied to the test piece 1 is detected.

[0044] The load cell holder 23 is mounted on the actuator 40 by a plurality of bolts (not shown). It is fixed to the upper surface 42A of the cylinder 42. The load cell holder 23 has a configuration in which the upper and lower flange portions are connected by a plurality of (for example, two) pillar portions, and is formed of, for example, a lump of metal material.

 [0045] The upper grip portion 24 is provided with a round bar-shaped shaft portion 27 extending upward, and a male screw 28 is cut on the upper portion of the shaft portion 27. A bush 29 is sandwiched between the shaft portion 27 and the load sensor 21, and thereby the relative positioning between the upper grip portion 24 and the load cell 21 in a direction orthogonal to the load direction (vertical direction in FIG. 2). Has been done. The external thread 28 of the shaft 27 is engaged with a lock nut 30 by force S. By tightening the lock nut 30, the relative position of the upper grip portion 24 with respect to the load cell 21 in the load direction is adjusted. It is fixed.

 The lower grip 25 is screwed to a piston screw 41D provided at the tip of the first piston rod 41B of the actuator 40, and is fixed to the first piston rod 41B. As a result, the test piece 1 is loaded with a load due to the reciprocating motion of the piston 41 of the actuator section 40.

 In FIGS. 2 and 3, each of the grasping portions 24 and 25 for grasping the upper and lower ends 1A and 1B of the test piece 1 is, for example, a circular insertion for inserting the upper and lower ends 1A and 1B of the test piece 1. Specimen holders 24A and 25A with holes 31 and 32 formed, and upper and lower ends 1A and 1B of specimen 1 inserted into insertion holes 31 and 32 by tightening multiple bolts 33 and 34 The test piece fixing lids 24B and 25B are configured to hold down and fix the test piece. Each test piece fixing lid 24B, 25B has a disk shape having a through hole through which the test piece 1 is inserted in the center, that is, a donut shape. It consists of a lid piece.

[0048] Each of the test piece fixing lids 24B and 25B may be divided into three parts instead of two parts. For example, in the case of dividing into three parts, each part is divided into 120-degree sections to form a substantially fan-shaped lid piece, and each of these lid pieces is attached to each test piece holder 24A, 25A using two bolts 33, 34, respectively. Fix it. That is, the test piece fixing cover 24B is fixed with a total of six bolts 33, and the test piece fixing lid 25B is fixed with a total of six bolts 34. As a result, even if the bolts are fixed with the same total of six bolts 33 and 34, the test piece 1 in the three divisions is better than that in the two divisions. The load applied to each end 1A, IB becomes even. Therefore, for the test piece 1, a more reasonable load can be achieved by using the three-piece test piece fixing lids 24B and 25B than by the two-piece test piece fixing lids 24B and 25B. However, there is an advantage that the two-parting is easier to install. For this reason, for example, a material that may break due to fatigue from the contact between the upper and lower ends 1A and IB of the test specimen 1 and the test specimen fixing lids 24B and 25B in the long life region is divided into three parts. For materials that do not cause breakage such as the above, such as high-strength steel, for example, using two-piece test piece fixing covers 24B and 25B, Depending on the material of 1, the test piece fixing lids 24B and 25B having different numbers of divisions may be used.

 In FIG. 3, the bottom surfaces of the insertion holes 31 and 32 formed in the test piece holders 24A and 25A are the test piece contact surfaces 31A on which the upper and lower end faces 1C and 1D of the test piece 1 are respectively contacted. , 32A. When the test piece 1 is mounted on the test piece mounting part 20, the outer peripheral surfaces IE and 1F of the upper and lower ends 1A and 1B of the test piece 1 and the test piece holders 24A and 25A of the upper and lower grip portions 24 and 25 are provided. The gaps 35 and 36 are formed between the inner peripheral surfaces (side surfaces) 31B and 32B of the insertion holes 31 and 32 formed in FIG.

 In FIG. 3, the upper and lower ends 2A, 2B of the alignment test specimen used for the alignment operation performed before the test are indicated by the two-dot chain line in the figure. Except for the upper and lower ends 2A and 2B, this dummy test piece for centering is the same as the test piece 1 used in the actual test. The outer diameters (diameters) Wl and W2 of the upper and lower ends 2A and 2B of the centering dummy test piece are the outer diameters of the upper and lower ends 1A and 1B of the test piece 1 used in the actual test. The dimensions (diameter) are larger than Dl and D2, respectively. Therefore, assuming that the inner diameters (diameters) of the insertion holes 31 and 32 formed in the test piece holders 24A and 25A of the upper and lower grip portions 24 and 25 are Hl and H2, the upper and lower ends of the test piece 1 used in the actual test are used. The gaps (H1-D1) and (H2-D2) between the outer surfaces 1E and IF of the parts 1A and 1B and the inner surfaces (side surfaces) 31B and 32B of the insertion holes 31 and 32 are for alignment. The gaps (HI—Wl), (H2—) between the outer peripheral surfaces 2E, 2F of the upper and lower ends 2A, 2B of the dummy specimen and the inner peripheral surfaces (side surfaces) 31B, 32B of the insertion holes 31, 32. Each is larger than W2).

Furthermore, the processing accuracy of each component (see FIG. 2) that plays a role in defining the parallelism between the test piece contact surfaces 31A and 32A (see FIG. 3) of the upper and lower grip portions 24 and 25 is high. Each component Each surface (upper surface or lower surface) is finished to satisfy the parallelism of 0.01. Further, the surface roughness of each of these surfaces is equivalent to the roughness of the polished finish. Note that the parallelism of these surfaces is the parallelism after securing a perpendicularity to the load direction (the direction in which the piston 41 advances and retreats).

 Specifically, the following surface is processed to have a parallelism of 0.01 and a roughness equivalent to a polished finish. That is, in FIG. 3, (1) the upper and lower end surfaces 1C and 1D of the test piece 1, (2) the test piece contact surfaces 31A and 32A of the upper and lower grip portions 24 and 25, and (3) the upper and lower ends of the test piece 1. Opposite surfaces 1G, 1H of parts 1A, 1B, (4) Upper and lower test specimen fixing lids 24B, 25B contacting these opposing surfaces 1G, 1H. Bottom bottom surfaces 24C, 25C, (5) Upper and lower grips 24, The surface 24D, 25D of 25 specimens Honoreda 24A, 25A, and (6) the above conditions were applied to the upper and lower joint surfaces 24E, 25E of the upper and lower specimen fixing lids 24B, 25B joined to these surfaces 24D, 25D. High-precision processing that satisfies is performed.

 In FIG. 2, (7) the upper surface 42 A of the cylinder 42 of the actuator section 40, (8) the lower surface 23 B of the load cell holder 23, (9) the upper surface 23 A of the load cell holder 23, and (10) the lower surface of the load cell 21. 21A, (11) Lower surface 30A of lock nut 30, (12) Upper surface 29A of bush 29, (13) Lower surface 29B of bush 29, (14) Upper surface 24F of upper grip 24, (15) Lower grip The lower surface 25F of 25, and (16) the upper end surface 41E of the first piston rod 41B is subjected to high-precision processing that satisfies the above conditions.

 [0054] The actuator section 40 is the same as the above-described actuator section of the fatigue tester described in Patent Document 1 by the present applicant, and therefore, detailed illustration is omitted. In FIG. 1, the actuator section 40 includes a piston 41 that moves forward and backward to apply a load to the test piece 1, and a cylinder 42 that is arranged around the piston 41 and guides the sliding of the piston 41. I have. The piston 41 has a head portion 41A that moves forward and backward due to a difference in front and rear oil pressure, a first piston rod 41B provided on the forward side of the head portion 41A, and a first piston rod 41B provided on the retreat side of the head portion 41A. It is configured to include a two-piston rod 41C.

[0055] In a cylindrical space inside the cylinder 42, a first pressure chamber 43 is formed on the forward side of the head 41A of the piston 41, and a second pressure chamber 44 is formed on the retreat side of the head 41A. ing . The first pressure chamber 43 and the second pressure chamber 44 have a first pressure chamber Oil for applying oil pressure to the head portion 41A is supplied through the flow path 45 and the second flow path 46.

 The servo valve 50 is similar to the servo valve of the fatigue tester described in Patent Document 1 by the present applicant, and has four ports (not shown), ie, a P port, an R port, a C1 port, and a C2 port. It has a port. The P port is connected to a supply pressure pipe of a hydraulic unit (not shown), and the R port is connected to a return pipe connected to an oil tank (not shown). The C1 port and the C2 port are connected to a first flow path 45 and a second flow path 46 formed in the cylinder 42 of the actuator section 40, respectively.

 [0057] The servo valve 50 switches the oil that has entered from the P port to the C1 or C2 port according to the current signal or the voltage signal that is sent from the DSP controller 70. For example, when switching to the C1 port, the oil that has flowed out of the C1 port enters the first pressure chamber 43 through the first flow path 45 and pushes the head 41A of the piston 41 to move the piston 41 backward. Due to the backward movement of the piston 41, the oil in the second pressure chamber 44 is pushed by the head portion 41A and enters the C2 port through the second flow path 46. Then, the oil that has entered the servo valve 50 from the C2 port is guided to the R port in the servo valve 50 and finally flows to the return pipe of the hydraulic unit. The oil circulates through such a route. The same is true when switching to the C2 port, in which case the piston 41 moves forward.

 [0058] Accordingly, the switching control of the servo valve 50 causes the piston 41 to move forward and backward. At this time, as shown in FIG. 2, the first piston rod 41B of the actuator section 40 is connected to the lower end 1B of the test piece 1 via the lower grip 25. For this reason, with the reciprocating movement of the piston 41, the lower end 1B of the test piece 1 is displaced in the longitudinal direction of the test piece 1, and as a result, a load is applied to the test piece 1. It is as follows.

[0059] As the servo valve 50, for example, a so-called nozzle flapper type servo valve can be suitably used, and from the viewpoint of further improving responsiveness, a so-called direct drive type servo valve may be used. Here, the former nozzle flapper type servo valve is formed by a pressure difference between a nozzle flapper and a flapper in a portion called a first stage. The latter moves the pool, and the latter direct drive type servo valve drives the spool directly by a driving element such as a voice coil or a giant magnetostrictive element / electrostrictive element.

 In FIG. 1, the control means 60 feeds back a detection signal of the strain gauge 22 corresponding to the load applied to the test piece 1 to control the operation of the actuator section 40 and sends a control signal to the servo valve 50. The DSP controller 70 performs digital processing of the processing directly related to control performance, and transmits and receives information between the DSP controller 70 and the DSP controller 70. An external processing device 90 that performs processing that is not directly related to control performance is provided.

 In FIG. 1, the DSP controller 70 includes a function generator 71, a servo amplifier 72 constituting control signal generation means, a D_A converter 73, an A_D converter 74, a counter 75, and a limiter 76. Reply The DSP controller 70 also includes a peak-bottom Hornoled 77, an amplification factor correction / zero correction calculation unit 78, an amplification factor correction unit 79, and a zero correction unit 80.

 [0062] Each component of the DSP controller 70 includes one or a plurality of central processing units (CPUs) provided inside the DSP controller 70 and one that defines the operation procedure of the CPU. Alternatively, it is realized by a plurality of control programs and various memories such as a ROM and a RAM. Further, the hardware constituting the DSP controller 70 may include, for example, a product-sum operation unit in addition to the CPU. The control program is such that an execution program created and compiled by the external processing device 90 is mounted on the DSP controller 70. This control program is created by combining a program created by the external processing unit 90 with various instructions (for example, an instruction for outputting a sine waveform) provided by a DSP board built in the DSP controller 70. Generally, all parts of the control program may be created by the external processing device 90.

The DSP controller 70 performs the following fully digitalized feedback control. Explaining from the outline of the control, first, a signal corresponding to the target load is generated in the function generator 71. This signal is amplified at a predetermined amplification rate by the servo amplifier 72 and is converted into an analog current signal or a voltage signal through the DA converter 73. . The analog current signal or the voltage signal is replaced in the servo valve 50 with oil having a flow rate proportional to the magnitude of the signal, and guided to the first pressure chamber 43 or the second pressure chamber 44 of the actuator section 40. As a result, the piston 41 moves, and a load is applied to the test piece 1.

 Next, the load applied to the test piece 1 is converted into an analog voltage signal or a current signal in proportion to the load by the load cell 21 and the strain gauge 22, and further converted into a digital signal through the AD converter 74. Is done. Then, a deviation value between the target value signal from the function generator 71 and this load signal is calculated, and this value is sent to the servo amplifier 72 again. The deviation finally converges by such a feedback loop, and a load corresponding to the target value signal is applied to the test piece 1.

 On the other hand, since the fatigue tester 10 is affected by various disturbances as a whole system, the load signal may not follow the target value signal when the gain is set to a constant value. In such a case, a correction operation for disturbance is performed by the peak bottom hold 77, the amplification factor correction / zero correction calculation unit 78, the amplification factor correction unit 79, and the zero correction unit 80. That is, the peak value and the minimum value of the load signal are monitored in real time by the peak bottom hold 77, and the amplification factor correction / zero correction calculation means 78 determines the optimum amplification factor based on the deviation between the monitoring result and the target value signal. A correction amount and a zero-point correction amount are calculated. Then, based on this calculation result, the signal of the function generator 71 is corrected by the zero point correction means 80, and the amplification factor of the servo amplifier 72 is corrected by the amplification factor correction means 79.

 Further, when the load signal deviates from a preset range, such as when the test piece 1 is broken or when an emergency occurs, the fatigue tester 10 is stopped by the limiter 76. The number of repetitions required for the fatigue test is counted by the counter 75.

 Next, the details of the control, that is, the details of the processing of each component of the DSP controller 70 will be described.

 [0068] The function generator 71 performs a process of generating a target value r (k). The target value r (k) is a sign waveform and is represented by the following equation (1).

[0069] r (k) = Α Χ 8ϊη (2 π X (F X k / F)) + OFF · · · (1)

[0070] Here, A is the sine wave amplitude (amplitude) of the attained target load, and F is the value of the fatigue test. The repetition frequency (frequency), F is the operating frequency (sampling frequency) of the A / D converter 74 of the DSP controller 70, and OFF is the zero point, that is, the sine wave average value (offset) of the target load. k is a value indicating a certain time in the discretized time.For example, if the sampling frequency F is 50 Hz, 50 times of time data can be measured per second. , K = l data represent the first time data of the 50 data.

 Further, the sine wave amplitude A and the zero point OFF of the attained target load are shown in FIG.

 t t

 It is determined from the maximum stress and stress ratio input by the experimenter using the case setting screen 200. This determination process is performed by the processing means 90A of the external processing device 90, and the result of the determination is delivered to the function generator 71, and is used for the processing in the function generator 71. For example, suppose that the mouthpiece 21 is calibrated to output 10 V at 100 kgf (9.81 kN). At this time, if you want to apply a repetitive sine waveform load of 1000kgf (9.81kN), maximum / J, and a value of 100kgf (0.98kN) to test piece 1, the amplitude of the target load is (1000-100 ) / 2 = 450kgf (4.41kN), and the average sine wave of the target load is (1000 + 100) / 2 = 550kgf (5.39kN). In other words, this waveform is a sine waveform having an amplitude of 450 kgf (4.41 kN) above and below 550 kgf (5.39 kN). At this time, lkgf (9.81N) corresponds to 0.01 V, so if the above target load is converted to a target voltage signal, the amplitude A = 450 X 0.01 = 4.5 V, zero point OFF = 550 X 0.01 = 5.5V. In the fatigue testing machine 10, the setting input based on the stress value is received on the test condition setting screen 200 in FIG. 5 by the test condition setting screen display 'input reception processing means 92' of the external processing device 90 as described later. The experimenter does not need to perform the conversion work from the stress value to the load value.The external processing device 90 converts the stress value to the load value, calculates the target load amplitude and zero point, determines the load, and converts the voltage value to the voltage value. Performs all conversion processing automatically.

 Further, the repetition frequency F of the fatigue test is input by the experimenter using a test condition setting screen 200 in FIG. 5 described later.

[0073] The servo amplifier 72 performs a process of generating a control signal to be sent to the servo valve 50. The control by the DSP controller 70 is proportional control, and the manipulated variable u (k) of the proportional control sent to the servo valve 50 as a control signal is expressed by the following equation (2). [0074] u (k) = GX e (k) (2)

 Here, G is an amplification rate, that is, a proportional gain (gain). e (k) is the deviation between the target value r (k) and the response value (the load value detected by the strain gauge 22) X (k), and e (k) = r (k)-x (k ).

 Therefore, considering the calculation processing of the deviation e (k) and the processing by the servo amplifier 72 based on the deviation e (k), the value of the control signal finally sent to the servo valve 50 becomes The following equation (3) is obtained by substituting equation (1) into equation (2).

 [0077] u (k)

 = G X (r (k) -x (k))

 = G X (A X sin (2 X (F X k / F)) + OFF -x (k))

 (3)

 [0078] The peak bottom hold 77 measures the maximum value and the minimum value of each cycle of the waveform of the actually measured load value detected by the strain meter 22, and performs a process of storing and storing the maximum value and the minimum value.

 The amplification factor correction / zero correction calculation means 78 compares the waveform of the target value signal with the waveform (feedback waveform) of the detection signal of the strain meter 22, and based on the comparison result, calculates the amplification factor (gain). And a calculation process for correcting the zero point (offset). This is a process for performing adaptive proportional control. The comparison between the target value signal waveform and the feedback waveform includes comparing the amplitude of the target value signal waveform with the amplitude of the feedback waveform, and comparing the zero point (offset) of the target value signal waveform with the zero point (offset) of the feedback waveform. This is a comparison. Here, the amplitude and offset of the waveform of the target value signal are values calculated and determined by the external processing device 90 when the test conditions are first set on the test condition setting screen 200 in FIG. 5 as described above. On the other hand, the amplitude and offset of the feedback waveform are calculated by the amplification factor correction / zero correction calculation means 78 based on the maximum value and the minimum value measured by the peak bottom hold 77. It should be noted that the comparison processing of these waveforms and the calculation processing relating to the correction based on the comparison result are basically performed for each cycle of the waveform, but may be performed for a plurality of cycles.

Further, the amplification factor correction / zero correction calculation means 78 outputs a response signal (a detection signal of the strain meter 22). According to the degree of convergence to the target value signal in step ()), the processing for switching the algorithm of the calculation processing relating to the correction of the amplification factor (gain) and the zero point (offset) is performed stepwise. This is a process for performing multi-stage adaptive proportional control. In principle, the process of determining whether or not to switch is performed for each cycle of the waveform, but may be performed for a plurality of cycles.

 [0081] A specific example of the process of switching the algorithm stepwise includes the following two-stage switching process. The number of switching stages, the content of the algorithm, the switching determination method, and the like are as follows. It is not limited.

For example, regarding the amplification factor (gain), the following algorithm A and algorithm B are used.

, Can be automatically switched based on the following conditions.

[0083] <Anoregorism A>

 At a certain time during the fatigue test, the amplitude of the measured load is A, and the proportional gain at that time is G. Here, assuming that the amplitude of the attained target load is A, the proportional gain to achieve this is

 t

 The in-predicted value G can be thought of as G = (A / A) XG. This value of G and the current ratio

 t t t c c t

 Assuming that the difference from the gain G is AG, AG is given by the following equation (4).

[0084] AG = G— G = ((A / A) _1) XG (4)

 [0085] In the calculation processing (AGC: Auto Gain Control) for the amplification factor correction in the normal adaptive proportional control, based on the AG of the above equation (4), the algorithm shown in the following equation (5) is optimal. The calculation for gain is performed.

 [0086] G (i + 1)

 c

 = G (i) + gtpx AG

 = G (i) + gtpX ((A A) -l) XG (i) '' (5)

Here, i indicates the i-th sampling data, and corresponds to each cycle. Therefore, the value of i changes at a longer time interval than k in Equation (1) and Equation (3). Gtp is a parameter (gain tuning parameter) used to smooth the change in proportional gain. As the value of gtp, for example, an experience value of 0.02 or the like can be used, but the value is not limited to this. The above method is Algorithm A.

[0088] Kwanoregorism B>

In normal adaptive proportional control, AGC is performed using only algorithm A described above. However However, when the measured load value converges to the target value to some extent, the value of gtp XAG becomes extremely small, and the convergence peaks out, making it difficult to converge the error to a value smaller than, for example, 1%.

 Therefore, in the multi-stage adaptive proportional control, the AGC algorithm is changed from the above algorithm A to the following algorithm B when the actually measured load value converges to a target value to some extent after the start of the test.

 (1) When A <A, the proportional gain is increased by 0.0002 calories.

 C t

 (2) If A> A, decrease the proportional gain by 0.0002.

 c t

 (3) When A = A, the proportional gain is left as it is.

 c t

 [0091] In the algorithm B, the amount of increase or decrease is not limited to the above-mentioned 0.0002, but may be another value, or the increase and decrease amounts may be different values.

 [0092] The timing of switching from algorithm A to algorithm B is the first time gtp X A G <0.0006 is satisfied during control by algorithm A. Therefore, in each cycle, gtp X A G <0.0006 is satisfied, and the repulsive force is cut off by half IJ. The numerical value of 0.0006 is a value obtained empirically through experiments, but the criterion value is not limited to this.

 The error can be converged to ± 0.1% or less by the multi-stage adaptive proportional control that switches from algorithm A to algorithm B as described above. Note that the value of gtp in the algorithm A is dynamically changed while monitoring the degree of convergence, or a switching criterion value (boundary value of gtp XAG) which is only 0 · 0006 in the above example. A multi-stage adaptive proportional control that switches over three or more stages by changing the amount of increase / decrease in algorithm B, etc.

 [0094] For the zero point (offset), the following algorithm C and algorithm D can be automatically switched based on the following conditions.

 [0095] Kwanoregorism C>

 When the error is greater than 1.5% of the rated output, use the following algorithm C.

[0096] (1) If the current load average value is larger than the target load average value, decrease the offset value. The amount of adjustment is 0.15% of the rated output. (2) If the current load average value is smaller than the target load average value, increase the offset value. The adjustment amount is 0.15% of the rated output.

 (3) If the current average load value is the same as the target average load value, leave the offset as it is.

 [0097] Cyanoregorism D>

 When the error is less than 1.5% of the rated output, the following algorithm D is used.

[0098] (1) If the current load average value is larger than the target load average value, reduce the offset value. The adjustment amount is 0.006% of the rated output.

 (2) If the current load average value is smaller than the target load average value, increase the offset value. The adjustment amount is 0.006% of the rated output.

 (3) If the current average load value is the same as the target average load value, leave the offset as it is.

 [0099] Here, the rated output refers to the larger of the absolute value of the maximum value of the target load and the absolute value of the minimum value of the target load.

 [0100] Therefore, the flow of processing for determining the offset correction amount is the same for both algorithms C and D, and the only difference is the numerical value of the adjustment amount. As described above, in the present invention, even when there is only a difference between the used data, the algorithm is regarded as different, and the “algorithm” is defined. Note that the values of each adjustment amount of 0.15% and 0.006% of the rated output are forces S, which are empirically obtained by experiments, and are not limited to these values. Also, the error value of 1.5% of the rated output, which serves as a criterion value for switching between the algorithms C and D, is not limited to this, and other values may be used.

 [0101] Note that, in the above example, a plurality of switching reference values (error boundary values) that are 1.5% of the rated output should be provided, and another adjustment value should be prepared. Thus, multi-stage adaptive proportional control that performs switching in three or more stages is possible.

[0102] Further, the amplification factor correction / zero correction calculation means 78 always keeps the automatic offset control (A〇C: Auto offset control) valid even when the AGC is invalidated. That is, during the test, the load average value always or substantially coincides with the target value. The experimenter uses the “AGC enable” button 360 on the screen 300 during the test shown in FIG. Effect can be selected. In the fatigue tester 10 of the present embodiment, the average value is adjusted once before the test, so that the error of the average value can be kept small immediately after the start of the test. Therefore, the offset adjustment amount actually used may be almost the value of algorithm D.

 [0103] The amplification factor correction means 79 and the zero point correction means 80 store the calculation results obtained by the amplification factor correction / zero point correction calculation means 78 in a memory, and perform amplification factor and zero point correction processing based on the calculation results. Each one is to do. Since the processing by the amplification factor correction Z zero point correction calculation means 78 is performed for each cycle of the waveform, the contents of the correction processing by the amplification factor correction means 79 and the zero point correction means 80 are updated for each waveform cycle.

 The counter 75 performs a process of counting the number of repetitions necessary for a fatigue test and a process of a counter limit. Counter limit is a function that stops the test when the number of test repetitions reaches a specified value. The specified value is a value that the experimenter has input in the input section 233 of the number of repetition of the discontinuation on the test condition setting screen 200 in FIG.

 [0105] The limiter 76 performs each processing of a maximum limiter, a minimum limiter, a gain oscillation prevention limiter, and a test piece breakage detection mechanism.

 [0106] The maximum limiter is the maximum force detected by the peak bottom hold 77 When the value exceeds a preset value before the test, the minimum limiter is the minimum value detected by the peak bottom hold 77 This function stops the test operation when the value falls below the previously set value, and prevents the abnormal load from being repeated. The value set in advance before the test is a value calculated based on the value entered by the experimenter in the input section 231 of the allowable excessive error on the test condition setting screen 200 in FIG.

 [0107] The gain oscillation prevention limiter has a function of preventing the output from increasing without limit by auto tuning and making the test unstable. The experimenter applies a limiter based on the value entered in the input section 232 for the maximum allowable gain on the test condition setting screen 200 in FIG.

 [0108] The test piece fracture detection mechanism detects the fracture of the test piece 1 by sensing a sudden change in the response load waveform, and stops the fatigue test at the moment when the test piece 1 breaks, thereby detecting the fracture of the test piece. And a function to protect the fatigue tester 10 itself.

[0109] The external processing device 90 is configured by a computer, and has various functions not directly related to control performance. It is provided with processing means 90A for performing seed processing, input means 96 such as a keyboard and a mouse, and display means 97 such as a liquid crystal display and a CRT display. Further, for example, an output unit such as a printer or a plotter may be appropriately provided.

 [0110] The processing means 90A includes a standby screen display, an input reception processing means 91, a test condition setting screen display, an input reception processing means 92, a screen display during the test, an input reception processing means 93, and a reciprocal operation forced invalidation. Means 94 and input parameter auditing means 95.

 [0111] The standby screen display 'input reception processing means 91 is a processing for displaying the standby screen 100 of Fig. 4 used when operating the actuator unit 40 in a state other than the test, and the standby screen 100 This is a process for receiving an operation input to the actuator section 40 by the experimenter performed by using.

 [0112] Display of test condition setting screen · Input reception processing means 92 is a process of displaying test condition setting screen 200 of FIG. 5 for setting test conditions, and an experimenter performing using this test condition setting screen 200. The processing for accepting the input of the setting of the test condition according to. If the experimenter inputs necessary values on the test condition setting screen 200 of FIG. 5 and instructs the decision, the input test conditions are displayed on the test condition setting screen display ' The parameters are automatically converted into parameters that can be interpreted by the control program of 70 and transferred to the DSP controller 70.

 [0113] The screen display during test 'input reception processing means 93 is a process for displaying the screen 300 during the test shown in Fig. 6 for monitoring the test status during the test, and the input performed by the experimenter using the screen 300 during the test. This is a process for accepting the request. In addition, when an error occurs during the test and the test is stopped, the input reception processing means 93 communicates with the control program of the DSP controller 70 and investigates the cause. Perform the process of reporting.

 [0114] The reciprocal operation forcibly invalidating means 94 performs a process of forcibly invalidating an operation that should not be performed or an operation that should not be performed while a certain operation is being performed. For example, during the test, the parameters of the test conditions cannot be changed.

[0115] The input parameter auditing means 95 performs a process of monitoring whether a parameter (test condition or the like) input by an experimenter is abnormal. For example, monitor for inconsistent parameter inputs on the test condition setting screen 200 in FIG. [0116] The means 91-95 included in the processing means 90A are provided inside a computer main body (including not only personal computers but also higher-order models) constituting the external processing device 90. This is realized by a central processing unit (CPU) and one or more programs that define the operation procedure of the CPU.

 In the present embodiment as described above, a fatigue test for acquiring ultra-high cycle range data is performed using the fatigue tester 10 as follows.

 First, before performing the fatigue test, the experimenter performs the centering operation of the main body 11 of the fatigue testing machine 10 using the dummy test piece for centering. In FIGS. 2 and 3, in the alignment operation, first, a dummy test piece for alignment is mounted on the grip portion 25 on the lower side (the piston 41 side). In this case, insert the lower end 2B of the alignment test piece into the insertion hole 32 formed in the test piece holder 25A, tighten the bolt 34, and use the test piece fixing lid 25B. Press down and fix the lower end 2B of the alignment test specimen.

 [0119] Next, the horizontal position of the load cell 21 is adjusted so that the upper end 2A of the dummy test piece for centering fits into the insertion hole 31 formed in the test piece holder 24A of the grip portion 24 on the upper side (the load cell 21 side). Adjust the position.

 Subsequently, after adjusting the horizontal position of the load cell 21, the load cell 21 is fixed by the bonoleto 26, and the lock nut 30 is tightened. Finally, the dummy specimen for alignment is removed.

 [0121] The above alignment work can be performed in about a few minutes, and the work time is remarkably reduced as compared with the case where the strain gauge is pasted on the test piece and the alignment work is conventionally performed. it can. After this alignment work, no matter how many times the test piece 1 for the fatigue test is attached at random, the test piece 1 will not receive any restraining force from the side. Therefore, unlike the conventional case, this alignment work does not need to be performed for each fatigue test.

 [0122] After performing the alignment operation using the alignment dummy test piece as described above, the experimenter attached a test piece to the test piece mounting portion 20 of the main body 11 of the fatigue tester 10 in order to perform a fatigue test. Attach strip 1. At this time, the experimenter uses the standby screen 100 of FIG. 4 displayed on the screen of the display means 97 by the input screen processing means 91 to display the standby screen of the external processing device 90 and displays the test piece 1 Operate the piston 41 necessary for installation work.

In FIG. 4, the standby screen 100 includes a feedback value display section 110 and a test condition table. Display unit 120, piston control unit 130, "communication end" button 140, "test condition setting" button 150, "to test console" button 160, and "numeric offset input" button 170 are provided. .

 [0124] The feedback value display section 110 is provided with a display section 111 of a load obtained from a detection signal of the strain gauge 22, and a display section 112 of a load stress obtained by converting the load into a test piece diameter. I have. The conversion process to the stress value is performed by the standby screen display 'input reception processing means 91.

 [0125] The test condition display section 120 is provided with display sections 121 124 for the maximum absolute stress, the test stress ratio, the test piece diameter, and the test frequency. These display values are the values entered by the experimenter on the test condition setting screen 200 in FIG. 5 by clicking the “test condition setting” button 150.

 [0126] The piston control unit 130 includes an up button 131 for moving the piston 41 forward (up), a down button 132 for moving the piston 41 backward (down), and a neutral button 133 for moving the piston 41 to the center position. , The advance amount display section 134 that displays the advance amount (up amount) of the piston 41 with a bar, the reverse amount display section 135 that displays the retreat amount (down amount) of the piston 41 with a bar, and the adjustment amount of the zero point of the servo valve 50. There are provided a valve zero adjustment input section 136 for inputting, and a “Set” button 137 for actually setting the value input on the valve zero adjustment input section 136. Since the zero point of the servo valve 50 is normally shifted to the safe side, the zero point adjustment is performed to correct this in advance and perform precise control.

 [0127] The "communication end" button 140 is a button for terminating communication with the DSP controller 70, and the "test condition setting" button 150 is a button for moving to the test condition setting screen 200 in FIG. The “To test console” button 160 is a button for moving to the in-test screen 300 in FIG. 6, and the “Numerical value input” button 170 is used to forcibly set the zero point (offset) at the time of adjustment, such as before starting the test. ) Is a button used to set a certain value.

[0128] The experimenter also sets test conditions before the test. In this case, the experimenter clicks the “test condition setting” button 150 on the standby screen 100 of FIG. Then, on the screen of the display means ₉₇ of the external processing device 90, the test condition setting screen display / input reception processing means 92 displays the test condition setting screen 200 of FIG.

In FIG. 5, a test condition setting screen 200 is used to set main conditions among the test conditions. The main condition setting section 210 for setting advanced conditions of the test conditions, the advanced condition setting section 220 for setting advanced conditions, the limiter setting section 230 for setting various limiter values, and the setting sections 210, 220, 230 A "cancel" button 240 for taking human power and a "set condition" button 250 for actually setting the conditions manually input in each of the settings 220 and 230 are provided.

[0130] The main condition setting section 210 is provided with input sections 211 214 for a maximum absolute stress, a test stress ratio, a test piece diameter, and a test frequency, and is used to input a check when performing a compression-compression test. A compression / compression test selection check input unit 215 is provided. The values of the maximum absolute stress, test stress ratio, and specimen diameter entered here are determined by the sine wave amplitude A and the average sine wave (zero) OFF of the target load in Equation (1) above.

 t t

 Used for fixed processing. This calculation determination processing is performed by the test condition setting screen display 'input reception processing means 92. The value of the test frequency input here is set as the value of F in the above-mentioned equation (1).

[0131] The altitude condition setting section 220 is provided with input sections 221-223 for load cell conversion constant, vacuum bias, and default gain. The load cell conversion constant indicates the relationship between the load applied to the load cell 21 and the output voltage of the strain gauge 22, and is input according to the calibration result of the load cell 21. It is used for load-voltage conversion in various processes such as sine wave amplitude A of target load and sine wave average value (zero) OFF calculation determination process. The default gain sets the initial value of the amplification factor (gain).

 [0132] The limiter setting section 230 is provided with input sections 231-233 for the permissible excessive error, the permissible maximum gain, and the number of repetition of discontinuation. These input values are used for various limit processes performed by the limiter 76 and the counter 75.

 When the experimenter finishes the input in each of the setting sections 210, 220 and 230 and clicks the “set condition” button 250, the input condition is set and the screen returns to the standby screen 100 of FIG. Then, when starting the test, the experimenter clicks the “To test console” button 160 on this screen 100. Then, a screen during test is displayed on the screen of the display means 97 of the external processing device 90. The screen 300 during test is displayed by the input reception processing means 93 in FIG.

In FIG. 6, a test condition display section 310 for displaying test conditions is displayed on a screen 300 during the test. A test status numerical display unit 320 for numerically displaying the test status and a test status graph display unit 330 for displaying the stress based on the actually measured load as a test status in a graph display are provided.

 The test condition display section 310 is provided with display sections 311 314 for the maximum absolute stress, the test stress ratio, the test piece diameter, and the test frequency. These display values are the values entered by the experimenter on the test condition setting screen 200 in FIG.

 [0136] The test status numerical display section 320 includes a count display section 321 for displaying the count value of the counter 75, a gain display section 322 for displaying the current gain, and a percentage of the rated input of the servo valve 50. The output LV display section 323, which displays the flow rate at what percentage of the maximum flow rate, and the output level, and the test error display section 324 are provided. The test error displayed here is (actual load amplitude-target load amplitude) / target load amplitude X 100 (%).

 [0137] The test status graph display section 330 displays a stress waveform (a waveform in which the horizontal axis represents time and the vertical axis represents stress) calculated from the actual load. The conversion process from the load value to the stress value at this time is performed by the screen display / input reception processing means 93 during the test.

 The display contents of the test status numerical display section 320 and the test status graph display section 330 are periodically updated by the screen display under test / input reception processing means 93.

 [0139] Also, on the screen during test 300, an "automatic offset adjustment" button 340 for automatically adjusting the offset, a "test stop" button 350 for stopping the test, and an "AGC enable" button 360 for enabling the AGC are provided. A "Reset counter" button 370 for resetting the count value of the counter 75, a "Unload" button 380 for unloading, and a "Go to main console" button 390 for moving to the standby screen 100 in FIG. Is provided.

 [0140] When the test is actually performed, the offset is statically adjusted to the offset value of the target load, and then the sum is added to superimpose the sine waveform. “Automatic offset adjustment” button 340 is a button for automatically bringing the load statically to the offset value before the test starts.

[0141] If the "AGC enable" button 360 is not clicked, the multi-stage adaptive proportional control is not enabled, and only the proportional control is performed. In this case, the values of the amplification factor (gain) and the zero point (offset) are constant, and even if the actual load deviates from the target load due to external factors, it will be appropriate. Accordingly, the operation of correcting the amplification factor and the value of the zero point is not performed.

 According to the present embodiment, the following effects can be obtained. That is, the amplification factor correction / zero correction calculation means 78 performs amplification processing in accordance with the degree of convergence of the actual load signal (the detection signal of the strain gauge 22) to the target value signal when performing the calculation processing relating to the amplification factor and zero correction. Since the algorithm of the calculation process related to the correction of the gain is switched stepwise, when the detection signal converges to the target value signal to some extent, the algorithm related to the correction of the amplification factor is switched so that the target value of the detection signal is The degree of convergence to a signal can be increased. Therefore, the test accuracy can be improved.

 Further, the amplification factor correction / zero correction calculation means 78 has a configuration in which not only the amplification factor but also the algorithm of the calculation process relating to the correction of the zero is switched stepwise according to the degree of convergence of the detection signal to the target value signal. Therefore, the degree of convergence of the detection signal to the target value signal can be further increased, and the test accuracy can be further improved.

 Further, since the amplification factor correction / zero correction calculation means 78 is configured to determine whether to switch the algorithm for each cycle of the waveform, the algorithm can be switched at an appropriate timing. . For this reason, the speed of convergence can be improved, and the test accuracy can be further improved.

 [0145] Specifically, assuming that the convergence of the error between the target value signal and the detection signal is, for example, about ± 1% in the conventional adaptive proportional control, the multi-stage adaptive proportional control according to the present embodiment uses ± 0%. . Can be converged within 1%.

[0146] Since the fatigue tester 10 includes the DSP controller 70 and the external processing device 90 as the control means 60, the DSP controller 70 includes arithmetic processing (target value signal) directly related to control performance. And the load signal, amplification factor correction, zero point correction, peak bottom hold, and calculation of amplification factor and zero point correction for multi-stage adaptive proportional control). For 90, arithmetic processing that is not directly related to control performance (calculation processing that converts the voltage fed back to load, numerical display of stress' Draft display of stress ・ Bar display of piston travel distance ・ Arrow display of piston operation button , Etc., which can be used by humans to make the display easy to handle. In other words, a program for arithmetic processing directly related to control performance The functions that are not directly related to the functions can be completely separated from the programs for arithmetic processing, and the hardware that executes each of these programs can be separated.

 [0147] For this reason, the processing load on the DSP controller 70 can be reduced, so that the processing speed of arithmetic processing directly related to control performance such as responsiveness, load accuracy, and convergence to disturbance can be improved. However, the test accuracy can be improved.

 The external processing device 90 includes a standby screen display and input reception processing means 91, a test condition setting screen display and input reception processing means 92, and a test screen display and input reception processing means 93. Therefore, in each stage of the test and preparation, the experimenter refers to one of the three screens: the standby screen 100 in FIG. 4, the test condition setting screen 200 in FIG. 5, and the test screen 300 in FIG. At the same time, it is possible to confirm only the display of the information required in each scene, and to perform only the input work required in each scene. Therefore, the experimenter only needs to perform the minimum necessary operations in each stage of the test and its preparation, so that there is no need to perform extra operations or refer to extra information to think about extra things. Operability can be improved. Since the operability can be improved, the occurrence of erroneous recognition or erroneous operation can be avoided or suppressed, and the test accuracy can be improved.

[0149] Then, the waiting button display 131 and the descending button 132 are displayed as arrows in the piston control section 130 of the waiting screen 100 in FIG. Since the bar 134 and the retreat amount display section 135 are displayed as bars, the experimenter can intuitively operate the piston 41. In addition, erroneous operations can be avoided or suppressed. Further, since the feedback value display section 110 is provided and the stress is displayed, the experimenter can easily confirm the load state of the current load.

 [0150] Further, the test condition setting screen display 'input reception processing means 92 can receive the input of the maximum absolute stress, the stress ratio, and the test piece diameter on the test condition setting screen 200 in Fig. 5. For this reason, the experimenter can use only the commonly used parameter of the stress 'stress ratio' specimen diameter, which can improve the operability of the fatigue testing machine 10 and improve the fatigue testing machine and control. Even an experimenter without knowledge can easily perform a fatigue test.

[0151] Further, the display of the screen during test 300 shown in FIG. Since the stress is graphically displayed in the test status graph display section 330, the experimenter can easily and intuitively grasp the current test status. In addition, the experimenter can properly perform the fatigue test under the set conditions simply by pressing the buttons 340 to 390.

 [0152] Since the external processing device 90 includes the reciprocal operation forced invalidation means 94, erroneous operation can be prevented.

 Further, since the external processing device 90 includes the input parameter monitoring means 95, the safety of the test can be improved.

 [0154] Furthermore, in the fatigue tester 10, since the parallelism of each surface of each component of the test piece 1 and the main body 11 is improved, it is possible to suppress the occurrence of an uneven load. At the same time, the fitting between the ends 1A and 1B of the test piece 1 and the upper and lower grips 24 and 25 was loosened (see Fig. 3). It is possible to prevent the restraining force from being applied from the outer peripheral surfaces 1E, IF) of the portions 1A, 1B, thereby suppressing the occurrence of uneven loads. Therefore, test accuracy can be improved. Normally, loosening the fit improves the parallelism of each surface of each component of the test piece 1 and the main body 11 because of the possibility of lateral displacement of the test piece 1 during the fatigue test. 1 can be prevented from being applied with a force that causes lateral displacement.

 [0155] Alignment work is performed using a dummy alignment test specimen having ends 2A, 2B that are larger than ends 1A, 1B of test piece 1 used in the actual test. In addition to being able to perform with high accuracy, the loose fitting described above can be easily realized. In addition, if the alignment work is performed once using the alignment dummy test piece, a plurality of tests can be performed as it is, so that the labor of the experimenter can be reduced.

[0156] In order to confirm the effects of the present invention, an experiment for measuring the partial stress Δσ was performed in the following procedure. First, the alignment of the fatigue tester 10 is performed using the dummy test piece for alignment. Next, test piece 1 with a total of three strain gauges affixed to the test piece mounting part 20 at the position where the outer periphery of the test piece center is divided into three parts. At this time, the value of the strain gauge is not referred to. Subsequently, a static stress of ± 400 MPa is applied to the test piece 1. Finally, the partial stress Δσ is calculated from the value of the strain gauge. Then, the above measurement was performed several times. [0157] The above measurement results were as follows. The bias stress Δ σ at a load stress of ± 400 MPa was 3.2.3.9 MPa. The ratio of Δσ is 0.8-2.3% with respect to the applied stress. This value does not affect the fatigue test result at all. Also, in the fatigue tester (fatigue tester equipped with a spherical bearing) by the present applicant described in Patent Document 1 described above, the ratio of Δσ to the applied stress of 400 MPa is 3.5-7.4%. . Therefore, in the fatigue tester 10 of the present embodiment, it was found that the eccentric load was significantly reduced, and the effect of the present invention was remarkably shown.

 [0158] The present invention is not limited to the above-described embodiment, and includes modifications and the like within a range that can achieve the object of the present invention.

 That is, in the above-described embodiment, the amplification factor correction / zero correction calculation means 78 has a configuration in which, for both the amplification factor and the zero, the algorithm of the calculation process relating to the correction is stepwise switched. In the present invention, a configuration may be adopted in which the algorithm is switched stepwise only for the amplification factor. However, if the algorithm is switched stepwise for both the amplification factor and the zero point as in the above embodiment, the test accuracy can be further improved.

 As described above, according to the present invention, the waveform of the detected signal and the waveform of the target value signal are compared by the amplification factor correction / zero correction calculation means, and the amplification factor and the zero value are calculated based on the comparison result. When performing the calculation process related to the correction of the zero point, the algorithm of the calculation process related to the correction of the amplification factor is switched stepwise according to the degree of convergence of the detection signal to the target value signal. When the convergence is achieved, by switching the algorithm relating to the correction of the amplification factor, the degree of convergence of the detection signal to the target value signal can be increased, and the test accuracy can be improved.

 [0161] Further, according to the present invention, the display processing of the three screens of the standby screen, the test condition setting screen, and the test screen, and the use of these screens are performed by an external processing device provided separately from the digital controller. Since the input processing is performed, the operability can be improved, the operation burden on the experimenter can be reduced, and the occurrence of erroneous recognition or erroneous operation can be avoided or suppressed, and the test accuracy can be improved. There is an effect that can be.

Further, according to the present invention, the parallelism of each surface of the test piece and the component parts of the test machine main body is improved. In addition, since the fitting between the end portion of the test piece and the grip portion is loosened, it is possible to suppress the occurrence of an eccentric load and to improve the test accuracy.

BRIEF DESCRIPTION OF THE FIGURES

FIG. 1 is an overall configuration diagram of a fatigue tester according to one embodiment of the present invention.

 FIG. 2 is a cross-sectional view of a test piece mounting portion constituting a main body of the fatigue tester of the embodiment.

FIG. 3 is an enlarged cross-sectional view of a main part of a test piece mounting portion of the embodiment.

FIG. 4 is a view showing an example of a standby screen associated with processing by an external processing device constituting control means of the fatigue tester of the embodiment.

 FIG. 5 is an exemplary diagram of a test condition setting screen associated with processing by an external processing device constituting control means of the fatigue tester of the embodiment.

 FIG. 6 is a view showing an example of a screen during a test associated with processing by an external processing device constituting a control means of the fatigue tester of the embodiment.

Claims

The scope of the claims

 [1] A main body having a test piece mounting part for mounting a test piece and an actuator unit for applying a load to the test piece, and a detection signal corresponding to the load applied to the test piece is fed back to the actuator. Sending a control signal to control the operation of the unit to the main body feedback control means comprising a control means for performing,

 The control means includes a digital controller that performs the feedback control by digital processing,

 This digital controller

 Control signal generating means for generating the control signal according to the deviation between the detection signal and the target value signal;

 An amplification factor correction / zero correction calculation means for comparing the waveform of the detection signal with the waveform of the target value signal and performing a calculation process relating to an amplification factor and a zero correction used in the generation processing of the control signal based on the comparison result. When,

 Amplification factor correction means for performing the amplification factor correction processing based on the calculation result by the amplification factor correction / zero point correction calculation means;

 Zero point correction means for performing the zero point correction processing based on the calculation result by the amplification factor correction / zero point correction calculation means,

 The amplification factor correction / zero-point correction calculation means is configured to switch an algorithm of a calculation process relating to the correction of the amplification factor in a stepwise manner according to the degree of convergence of the detection signal to the target value signal.

 A fatigue test machine characterized by the following.

 [2] The fatigue tester according to claim 1,

 The amplification factor correction / zero correction calculation means is configured to switch the algorithm of the calculation process for the correction of the zero in a stepwise manner according to the degree of convergence of the detection signal to the target value signal.

 A fatigue test machine characterized by the following.

[3] The fatigue tester according to claim 1 or 2,

The amplification factor correction / zero point correction calculation means is configured to generate a periodically changing wave of the detection signal. The shape and the waveform of the target value signal, which changes periodically, are compared for each period of the waveform, and a calculation relating to the correction of the amplification factor and the zero point is performed based on the amount of deviation between both waveforms obtained as a result of the comparison. A fatigue tester characterized in that processing is performed for each cycle of a waveform, and whether or not to switch the algorithm is determined for each cycle of the waveform when performing this calculation.

[4] A main body having a test piece mounting part for mounting a test piece and an actuator unit for applying a load to the test piece, and a detection signal corresponding to the load applied to the test piece is fed back to the actuator. Sending a control signal to control the operation of the unit to the main body feedback control means comprising:

 The control means,

 A digital controller that performs the feedback control by digital processing;

 An external processing device that is connected to the digital controller and transmits and receives information to and from the digital controller to perform processing other than the processing related to the feedback control,

 The external processing device,

 A process of displaying a standby screen used when operating the actuator unit in a state other than during the test, and a process of receiving an operation input to the actuator unit by an experimenter performed using the standby screen are performed. 'Waiting screen display' input reception processing means,

 Displaying a test condition setting screen for setting test conditions and displaying a test condition setting screen for performing a process of receiving the test condition setting input by the experimenter performed using the test condition setting screen. Input reception processing means,

 During the test, a process for displaying a screen during the test that monitors the test status, and a process for receiving the input by the experimenter performed using the screen during the test, a screen display during the test.

 A fatigue tester characterized by comprising:

[5] The fatigue tester according to claim 4,

The test condition setting screen display The test condition is configured to receive any combination of a combination of a maximum stress and a minimum stress, a combination of a maximum stress and a stress ratio, a combination of a minimum stress and a stress ratio, and an input of a specimen diameter. The middle screen display 'input reception processing means is configured to display a graph of stress as the test status on the screen during test.

 A fatigue test machine characterized by the following.

 [6] An actuator part for applying a load to the test piece, and a test piece mounting part for mounting the test piece in a state of being arranged in a direction along the load direction by the actuator part, wherein the test piece mounting part A grip portion for gripping both ends of the test piece is provided, and each of these grip portions is formed with a test piece contact surface with which both end surfaces of the test piece abut. In a fatigue testing machine,

 The end face on both sides of the test piece satisfies the parallelism of 0.01 and plays a role in defining the parallelism between the test piece contact surfaces of the opposing gripping portions among the surfaces of the test machine main body component. All the surfaces that fulfill the cracks are also finished in a state that satisfies the parallelism of 0.01, and the outer peripheral surface of each end of the test piece is attached with the test piece attached to the test piece mounting part. A gap is formed between each of the grip portions.

 A fatigue test machine characterized by the following.

[7] The fatigue tester according to claim 6, wherein

 The gap between the outer peripheral surface of each end of the test piece used in the test production and the fitting of each gripping part is the same as the outer peripheral surface of each end of the dummy test piece for alignment used in the alignment work performed before the test. A fatigue tester characterized in that the gap is larger than the clearance of the fitting with each of the grip portions.