FORM 2 THE PATENT ACT 1970 (39 of 1970) The Patents Rules, 2003 COMPLETE SPECIFICATION See Section 10, and rule 13 1. TITLE OF INVENTION LOAD INERTIA ESTIMATION METHOD AND CONTROL PARAMETER ADJUSTMENT METHOD 2. APPLICANT(S) a) Name b") Nationality c) Address MITSUBISHI HEAVY INDUSTRIES, LTD. JAPANESE Company 16-5, -KONAN 2-CHOME, MINATO-KU, TOKYO 1088215, JAPAN 3. PREAMBLE TO THE DESCRIPTION The following specification particularly describes the invention and the manner in which it is to be performed : - TECHNICAL FIELD The present invention relates to a load inertia estimation method and a control parameter adjustment method applicable to industrial machines such as machine tools. BACKGROUND ART Feedback control which is a classical control theory is generally used for load position control of a feed system in an industrial machine such as a machine tool. Fig. 4 shows an example of a machine tool. The machine tool of the illustrated example is a double column type machining center which includes a bed 1, a table 2, a gate-shaped column 3, a crossrail 4, a saddle 5, a ram 6, and a main spindle 7. The table 2 is disposed on the bed 1 and the column 33 is disposed in such a manner as to straddle the table 2. A workpiece W is mounted on the table 2 at the time of machining, and the table 2 moves linearly in an X-axis direction along guiderails la on the bed 1 with the assistance of a feed system (not shown in Fig. 4, see Fig. 5). The crossrail 4 moves linearly in a Z-axis direction along guiderails 3b on a column front face 3a with the assistance of a feed system (not shown). The saddle 5 moves linearly in a Y-axis direction along guiderails 4b on a crossrail front face 4a with the assistance of a feed system (not shown). The ram 6 is provided on the saddle 5 and moves linearly in the Z-axis direction with the assistance of a feed system (not shown). The main spindle 7 is supported rotatably inside the ram 6, and a tool 9 is fitted onto a tip of the main spindle 7 via an attachment 8. Accordingly, when the workpiece W is machined with the tool 9, the tool 9 is driven to rotate by the main spindle 7. The main spindle 7 and the tool 9 move linearly in the Z-axis direction together with the crossrail 4 or the ram 6 and move linearly in the.Y-axis direction together with the saddle 5, and the table 2 and the workpiece W move linearly in the X-axis direction. In order to achieve high-precision machining of the workpiece W at this time, positions to which the main spindle 7 (the tool 9) and the table 2 (the workpiece W) are moved are required to be precisely controlled by the feedback control. Fig. 5 shows a general configuration example of a feedback control system and a feed system. Although detailed description is omitted herein, a feed system 11 for the table 2 shown in Fig. 5 includes a servo motor 12, a reduction gear unit 13, brackets 14, a ball screw 15 (a screw portion 15c and a nut portion 15b), and so forth. The feed system 11 moves the table 2 and the workpiece W linearly in the X-axis direction. A feedback control system 16 controls this feed system 11 as follows. Specifically, the feedback control system 16 controls rotation of the servo motor 12 in such a way that a load position 6L, which is a position of the table 2 (the workpiece W) detected with a position detector 6, follows a position command 8 issued from a numerical control (NC) device 17. However, it is difficult to achieve a sufficient following performance with the feedback control system 16 as in the illustrated example, and a delay of the load position 0L in following the position command 0 (namely, a delay in the load position) occurs as a consequence. In order to deal with the follow delay (the delay in the load position), it is a common practice to add, to the feedback control system 16,. a feed-forward control function, which is not illustrated, to differentiate the position command 9 and compensate for a position delay. However, addition of the feed-forward control function to the feedback control system cannot compensate for a position delay or vibration caused by dynamic deformation such as deflection or torsion that occurs in a mechanical element in a controlled object. For example, in the case of the feed system 11 in Fig. 5, rigidity of the screw portion 15c of the ball screw 15 has a limitation and thus torsion or deflection corresponding to load inertia (the weight of a workpiece) or the load position 0L occurs in the screw portion 15c at the time of moving the table 2. The feed-forward control function cannot compensate for the follow delay of the load position 8L thus caused. In .this context, Patent Document 1 listed below discloses a technique for compensating for a delay in a load position or a delay in a velocity caused by torsion or deflection of a ball screw in a feed system by finding a characteristic model (a transfer function) that approximates a characteristic of the feed system, then finding an inverse characteristic model (an inverse transfer function) of the characteristic model, and adding the inverse characteristic model to a feedback control system (see Fig. 1 and Fig. 2: to be described later in detail). Meanwhile, such techniques for adding an inverse characteristic model of a controlled object to a control system are also disclosed in Patent Documents 2 and 3 listed below, for instance. PRIOR ART DOCUMENTS PATENT DOCUMENTS Patent Document 1: Japanese Patent Application Publication No. 2009-201169 Patent Document 2: Japanese Patent No. 3351990 Patent Document 3: Japanese Patent No. 3739746 Patent Document 4: Japanese Patent No. 4137673 SUMMARY OF THE INVENTION PROBLEMS TO BE SOLVED BY THE INVENTION However, in Fig. 5, the weight of the table 2 remains constant whereas the weight of the workpiece W varies depending on the type of a machined product and the like. Accordingly, the load inertia to be determined by the weight of the table 2 and the weight of the workpiece W also varies with a change in the weight of the workpiece W. As a consequence, if the load inertia included in the inverse characteristic model (the inverse transfer function) of the feed system is always set to a constant value, then the load inertia included in the inverse characteristic model of the feed system differs from actual load inertia of the feed system when the workpiece W having a different weight from the constant value is mounted on the table 2 for machining. Accordingly, even when the inverse characteristic model of the ieed system is added to the feedback control system, the inverse characteristic model cannot sufficiently compensate for the follow delay of the load position 6L caused by torsion, deflection or the like of the ball screw 15 when the workpiece W having a different weight from the constant value is machined. Hence, a position deviation between the position command 9 and the load position 8L is increased. As a consequence, the workpiece W cannot be machined at high precision. For this reason, in order to enable the feedback control system, to which the inverse characteristic model of the feed system is added', to perform high-precision machining on the workpiece W having any weight, it is necessary to estimate the load inertia corresponding to the weight of the workpiece W and to adjust the load inertia included in the inverse characteristic model of the feed system based on the estimated load inertia. In view of the aforementioned circumstances, it is an object of. the present invention to provide a load inertia estimation method of estimating load inertia corresponding to the weight of a workpiece, and a control parameter adjustment method of adjusting load inertia included in an inverse characteristic model of a feed system based on the estimated load inertia. Incidentally, the above-mentioned Patent Document 4 discloses a method of calculating the weight of a load by using a difference between a motor torque when no load is applied and a motor torque when a load is applied. In contrast, the method of the present invention estimates the load inertia based on a position deviation and so forth. MEANS FOR SOLVING THE PROBLEMS A load inertia estimation method according to a first aspect of the invention for solving the above problems is a load inertia estimation method of estimating load inertia of a feed system for a load position control system configured to cause a feedback control system, to which an inverse characteristic model of the feed system is added, to control a load position of the feed system on the basis of an amount of compensation outputted from the inverse characteristic model and used for compensating for a dynamic error factor of the feed system. The method is characterized in that the method comprises: in the load position control system, conducting a load position control test using the feedback control system by issuing a position command to the feedback control system, and measuring a position deviation between the position command and the load position arising at a prescribed load position at this time; and in a load inertia estimation model being a model of the load position control system, conducting load position control simulation on a model of the feed system using a model of the feedback control system by issuing the position command to the model of the feedback control system, repeating the load position control simulation while the load inertia included in the model of the feed system is adjusted until a position deviation between the position command and the load position arising at the prescribed load position in the load position control simulation becomes equal to the position deviation measured in the load position control test, and as a consequence, if the position deviation arising at the prescribed load position in the load position control simulation becomes equal to the position deviation measured in the load position control test, estimating the load inertia included in the model of the feed system at this time as the load inertia of the feed system. In addition, a load inertia estimation method according to a second aspect of the invention is a load inertia estimation method of estimating load inertia of a feed system for a load position control system configured to cause a feedback control system, to which an inverse characteristic model of the feed system is added, to control a load position of the feed system on the basis of an amount of compensation outputted from the inverse characteristic model and used for compensating for a dynamic error factor of the feed system. The method is characterized in that the method comprises: in the load position control system, conducting a load position control test using the feedback control system by issuing a position command to the feedback control system and measuring a position deviation between the position command and the load position arising at a prescribed load position at this time, or in a model of the load position control system, conducting load position control simulation on a model of the feed system using a model of the feedback control system by issuing the position command to the model of the feedback control system and measuring the position deviation between the position command and the load position arising at the prescribed load position at this time; and finding load inertia corresponding to the position deviation measured in the load position control test or the load position control simulation on the basis of position deviation characteristic data which is preset based on the position deviation between the position command and the load position being measured in advance and arising at the prescribed load position when no load is applied and on the position deviation between the position command and the load position being measured in advance and arising at the prescribed load position when a certain load is applied and which increases linearly in proportion to an increase in the load inertia, and estimating the load inertia thus found as the load inertia of the feed system. Further, a control parameter adjustment method according to a third aspect of the invention is a control parameter adjustment method of adjusting load inertia included in an inverse characteristic model for a load position control system configured to cause a feedback control system, to which the inverse characteristic model of a feed system is added, to control a load position of the feed system on the basis of an amount of compensation outputted from the inverse characteristic model and used for compensating for a dynamic error factor of the feed system. The method is characterized in that the method comprises adjusting the load inertia included in the inverse characteristic model on the basis of the load inertia estimated by the load inertia estimation method according to the first or second aspect. EFFECT OF THE INVENTION The load inertia estimation method of the first aspect of the invention provides the method of estimating the load inertia of the feed system for the load position control system configured to cause the feedback control system, to which the inverse characteristic model of the feed system is added, to control the load position of the feed system on the basis of the amount of compensation outputted from the inverse characteristic model and used for compensating for the dynamic error factor of the feed system. Here, the method is characterized in that the method includes, in the load position control system, conducting a load position control test using the feedback control system by issuing a position command to the feedback control system, and measuring a position deviation between the position command and the load position arising at a prescribed load position at this time; and in a load inertia estimation model being a model of the load position control system, conducting load position control simulation on a model of the feed system using a model of the feedback control system by issving the position command to the model of the feedback, control system, repeating the load position control simulation while the load inertia included in the model of the feed system is adjusted until a position deviation between the position command and the load position arising at the prescribed load position in the load position control simulation becomes equal to the position deviation measured in the load position control test, and as a consequence, if the position deviation arising at the prescribed load position in the load position control simulation becomes equal to the position deviation measured in the load position control test, estimating the load inertia included in the model of the feed system at this time as the load inertia of the feed system. For this reason, even when the weight of a load on the feed system (such as the weight of a workpiece mounted on a table of a machine tool) varies, the load inertia corresponding to the load weight can easily be estimated. The load inertia estimation method of the second aspect of the invention provides the method of estimating the load inertia of the feed system for the load position control system configured to cause the feedback control system, to which the inverse characteristic model of the feed system is added, to control the load position of the feed system on the basis of the amount of compensation outputted from the inverse characteristic model and used for compensating for the dynamic error factor of the feed system. Here, the method is characterized in that the method includes, in the load position control system, conducting a load position control test using the feedback control system by issuing a position command to the feedback control system and measuring a position deviation between the position command and the load position arising at a prescribed load position at this time, or in a model of the load position control system, conducting load position control simulation on a model of the feed system using a model of the feedback control system by issuing the position command to the model of the feedback control system and measuring the position deviation between the position command and the load position arising at the prescribed load position at this time; and finding load inertia corresponding to the position deviation measured in the load position control test or the load position control simulation on the basis of position deviation characteristic data which is preset based on the position deviation between the position command and the load position being measured in advance and arising at the prescribed load position when no load is applied and on the position deviation between the position command and the load position being measured in advance and arising at the prescribed load position when a certain load is applied and which increases linearly in proportion to an increase in the load inertia, and estimating the load inertia thus found as the load inertia of the feed system. For this reason, even when the load weight on the feed system (such as the weight of the workpiece mounted on the table of the machine tool) varies, the load inertia corresponding to the load weight can easily be estimated. The control parameter adjustment method according to the third aspect of the invention provides the control parameter adjustment method of adjusting the load inertia included in the inverse characteristic model for the load position control system configured to cause the feedback control system, to which the inverse characteristic model of the feed system is added, to control the load position of the feed system on the basis of the amount of compensation outputted from the inverse characteristic model and used for compensating for the dynamic error factor of the feed system. Here, the method is characterized in that the method includes adjusting the load inertia included in the inverse characteristic model on the basis of the load inertia estimated by the load inertia estimation method according to the first or second aspect of the invention. Therefore, even when the load weight on the feed . system (such as the weight of the workpiece mounted on the table of the machine tool) varies, it is possible to cause parameters of the feed system to match parameters of the inverse characteristic model (such as coefficients (to be described later in detail) in differential terms of third and higher orders including the term of the load inertia). For this reason, it is possible to perform precise control over the load position such that the load position follows the position command, and thereby to cause, for example, a machine tool to perform high-precision machining. BRIEF DESCRIPTION OF THE DRAWINGS Fig! 1 is a view showing a configuration of a load position control system which embodies a load inertia estimation method and a control parameter adjustment method according to a first embodiment of the present invention. Fig. 2 is a view showing a configuration of a load inertia estimation model. Fig. 3 is a view showing a configuration of a load position control system which embodies a load inertia estimation method and a control parameter adjustment method according to a second embodiment of the present invention. ' Fig. 4 is a view showing a configuration of a conventional machine tool. Fig. 5 is a view showing a configuration of a conventional load position control system (a feedback control system and a table feed system). MODES FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below in detail based on the drawings. Now, the calculation method of setting (calculating) the coefficients al to a5 in the inverse characteristic model 50 will be described. In the mechanical system model shown in Fig. 2, the transfer functions for the inverse characteristic model involving the torque and the velocity can be calculated as follows. First, Formula (1) and Formula (2) shown below are found from equations of motion. Here, Formula (1) is an equation of motion representing an input-output relation concerning a motor transfer function that models a characteristic of the servo motor 23, and Formula (2) is an equation of motion representing an input-output relation concerning a load transfer function that models a characteristic of the table 2 and the workpiece W collectively serving as the load. [Expression 2] {eu-0'L)iCLs+KL)={?Ls* + DLs)-$L • * • (2) The following Formula (3) „and Formula (4) are derived from Formula (1) and Formula (2) shown above. [Expression 3] (Jis*+D,s* ] „ In order to move the load (the table 2 and the workpiece W) with no error, compensation control should be performed such that the load position 9L matches the position command θ, i.e., such that Θ=ΘL is satisfied. In order to satisfy Θ=ΘL, the torque command T should be subjected to feed-forward compensation control in accordance with a formula in braces {} (a first transfer function formula) on the right side of Formula (3), and the velocity command V should be subjected to feed¬forward compensation control in accordance with a formula in parentheses () (a second transfer function formula) on the right side of Formula (4). Note that 6MS in Formula (4) is equivalent to the motor velocity VM. In Formula (3), ΘL is replaced with 9 and then the formula is translated into a command velocity Vx. Thus, Formula (3) is converted into Formula (5). Formula (5) is equivalent to Formula (3) multiplied by an inverse operation expression of a proportional integral operation expression set in the proportional integral operating unit 34. In other words, Formula (5) is equivalent to Formula (3) divided by the proportional integral operation expression set in the proportional integral operating unit 34. A portion on the right side of Formula (5) excluding 6 constitutes a third transfer function. Meanwhile, Formula (6) shown below is obtained by replacing ΘL with 0 in Formula (4) and then transforming Formula (4). In order to perform the compensation control such that the load position ΘL matches the position command 9, the compensation velocity VH for achieving no error between θ and ΘL should be set equal to a sum of Formula (5) and Formula (6). Such a sum is expressed by Formula (7) below. A portion on the right side of Formula (7) excluding 0 constitutes a fourth transfer function. [Expression 4] It is not possible to organize the original Formula (7) in terms of the differential orders. However, the following Formula (8) is obtained by deleting the term CL, which has little effect on accuracy, from Formula (7). A portion on the right side of Formula (8) excluding θ constitutes a transfer function for compensation control. The following Formula (9) is obtained by replacing Formula (8) with the coefficients al to a5. In this way, the coefficients al to a5 are obtained from Formula (8) and Formula (9). [Expression 5] INDUSTRIAL APPLICABILITY The present invention relates to a load inertia estimation method and a control parameter adjustment method, which is useful for application to the case of adjusting load inertia included in an inverse characteristic model of a feed system that is added to a feedback control system of a machine tool and the like. EXPLANATION OF REFERENCE NUMERALS 1 bed 2 table 21 feedback control system 22 feed system 23 servo motor 24 reduction gear unit 24a motor end gear 24b load end gear 25 bearing 26 bracket 27 ball screw 27a screw portion 27b nut portion 28 position detector 29 pulse encoder 31 position deviation operating unit 32 multiplication unit 33 velocity deviation operating unit 34 proportional integral operating unit 35 current control unit 36 differential operating unit 41 NC device 50 inverse characteristic model 51 first-order differential term operating unit 52 second-order differential term operating unit 53 third-order differential term operating unit 54 fourth-order differential term operating unit 55 fifth-order differential term operating unit 56 addition unit 57 proportional integral inverse transfer function unit 60 load inertia estimation model 61 torque deviation operating unit 62, 63 blocks of transfer functions concerning servo motor 64,65,66 blocks of transfer functions concerning table and ball screw 67 position deviation operating unit 70 position deviation characteristic data unit WE CLAIM: 1. A load inertia estimation method of estimating load inertia of a feed system for a load position control system configured to cause a feedback control system, to which an inverse characteristic model of the feed system is added, to control a load position of the feed system on the basis of an amount of compensation outputted from the inverse characteristic model and used for compensating for a dynamic error factor of the feed system, the method characterized in that the method comprises: in the load position control system, conducting a load position control test using the feedback control system by issuing a position command to the feedback control system, and measuring a position deviation between the position command and the load position arising at a prescribed load position at this time; and in a load inertia estimation model being a model of the load position control system, conducting load position control simulation on a model of the feed system using a model of the feedback control system by issuing the position command to the model of the feedback control system, repeating the load position control simulation while the load inertia included in the model of the feed system is adjusted until a position deviation between the position command and the load position arising at the prescribed load position in the load position control simulation becomes equal to the position deviation measured in the load position control test, and as a consequence, if the position deviation arising at the prescribed load position in the load position control simulation becomes equal to the position deviation measured in the load position control test, estimating the load inertia included in the model of the feed system at this time as the load inertia of the feed system. 2. A load inertia estimation method of estimating load inertia of a feed system for a load position control system configured, to cause a feedback control system, to which an inverse characteristic model of the feed system is added, to control a load position of the feed system on the basis of an amount of compensation outputted from the inverse characteristic model and used for compensating for a dynamic error factor of the feed system, the method characterized in that the method comprises: in the load position control system, conducting a load position control test using the feedback control system by issuing a position command to the feedback control system and measuring a position deviation between the position command and the load position arising at a prescribed load position at this time, or in a model of the load position control system, conducting load position control simulation on a model of the feed system using a model of the feedback control system by issuing the position command to the model of the feedback control system and measuring the position deviation between the position command and the load position arising at the prescribed load position at this time; and finding load inertia corresponding to the position deviation measured in the load position control test or the load position control simulation on the basis of position deviation characteristic data which is preset based on the position deviation between the position command and the load position being measured in advance and arising at the prescribed load position when no load is applied and on the position deviation between the position command and the load position being measured in advance and arising at the prescribed load position when a certain load is applied and which increases linearly in proportion to an increase in the load inertia, and estimating the load inertia thus found as the load inertia of the feed system. 3. A control parameter adjustment method of adjusting load inertia included in an inverse characteristic model for a load position control system configured to cause a feedback control system, to which the inverse characteristic model of a feed system is added, to control a load position of the feed system on the basis of an amount of compensation outputted from the inverse characteristic model and used for compensating for a dynamic error factor of the feed system, the method characterized in that the method comprises adjusting the load inertia included in the inverse characteristic model on the basis of the load inertia estimated by the load inertia estimation method according to claim 1 or 2.