EP2345615A1 - Elevator control apparatus - Google Patents

Abstract

An elevator control apparatus which performs velocity control by using a model arithmetic operation part is capable of high-following-capability control by predicting the inertia value of a control target. An elevator control apparatus (110) includes an inertia error prediction unit (80A). A first arithmetic operation part (81a) calculates a pre-convergent inertia error (intermediate value) based on integral arithmetic operation of a velocity deviation between a model velocity (ω_A) and an actual velocity (ω_M) of a period where an acceleration build-up state and a constant acceleration state of an elevator car are consecutive, and outputs the pre-convergent inertia error to a second arithmetic operation part (82a). The second arithmetic operation part (82a) predicts a post-convergent inertia error based on the intermediate value output from the first arithmetic operation part (81). Based on the post-convergent inertia error predicted by the second arithmetic operation part (82a), a parameter correction unit (90) corrects an inertia value (J_A) preset in a model arithmetic operation unit (30).

Description

Technical Field
• The present invention relates to an elevator control apparatus and, more particularly, to an elevator control apparatus which calculates an inertia as a control target in velocity control having a model arithmetic operation unit.
Background Art
• Fig. 9 is a diagram showing the outline of a conventional generally-used elevator. In Fig. 9, the elevator is provided with, e.g., a commercial power supply 310, a car 350 to vertically move a human or a load, a counterweight 360 having a weight that balances the weight of the car and load to be transported, an inverter 330 which supplies power to drive a power device, a converter 320 which supplies power from the commercial power supply to the inverter, an elevator control apparatus 340 which controls the inverter 330, the converter 320, a power device 301, and the like.
• The conventional elevator control apparatus which calculates the inertia of a control target includes a velocity instruction input means, a model arithmetic operation unit, a velocity detector, a compensation arithmetic operation unit, a torque instruction calculation unit, a torque controller, and an inertia calculator, as described in patent literature 1. When accelerating the car, the inertia of the control target is calculated by the following equation (A):
• $$\begin{array}{c}\left[\mathrm{Formula}1\right]\\ J_M=T_{\mathrm{\alpha }}\frac{1}{\mathrm{\alpha }}\frac{\mathrm{<figure-callout\; id="2"\; label="D"\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">D</figure-callout>}}{2}\frac{1}{\mathit{KL}}\end{array}$$

  
• (J_M: inertia of the control target; Tα: torque necessary for the elevator car to accelerate; D: sheave coefficient; and KL: roping coefficient)
• More specifically, in patent literature 1, when accelerating the car or moving the car at a constant velocity, the inertia of the control target is calculated using a force equilibrium equation, i.e., an equation (A), and the result is reflected in the model arithmetic operation unit and the compensation arithmetic operation unit. According to patent literature 1, this enables high-following-capability control.
• Patent Literature 1: JP 2003-128352
Disclosure of Invention Technical Problem
• In the elevator control apparatus of patent literature 1, a correct inertia can be calculated in a steady state where the car acceleration is constant and the above equation (A) is established. A correct inertia, however, cannot be calculated in a transient state where the acceleration time is short, the feedback does not converge sufficiently, the acceleration is not constant, and accordingly the above force equilibrium equation (A) is not established. Therefore, when calculating the inertia of a low-velocity elevator, identification must be performed a number of times to decrease the error, or the acceleration time must be increased so the acceleration converges.
• In the elevator control apparatus of patent literature 1, a correct inertia can be calculated in a steady state where the car acceleration is constant and the above equation (A) is established. However, the inertia cannot be calculated in cases where a constant acceleration time does not substantially exist, e.g., when the inertia error is excessively large and the control system diverges to stop during acceleration build-up, or when the velocity is excessively low and a velocity pattern of constant acceleration build-up and constant deceleration build-up results, as in manual operation.
• It is an object of the present invention to provide a high-following-capability elevator control apparatus that calculates an accurate inertia quickly by calculating the convergence rate of the inertia error even during a transient response before convergence of the car acceleration, and that uses the calculated accurate inertia for normal operation. It is another object of the present invention to provide a high-following-capability elevator control apparatus that calculates the inertia of a control target even in a situation where constant acceleration cannot be performed, and that uses the calculated accurate inertia for normal operation.
• More specifically, according to patent literature 1, when the acceleration time is short (that is, when the constant velocity is low), the feedback does not converge, and the inertia error does not converge, so a correct inertia cannot be obtained. Then, as compared to a case where an accurate inertia is used as a control parameter, the control performance degrades, and the following capability becomes poor. It is, therefore, an object of the present invention to provide an elevator control apparatus that can obtain an accurate inertia even when the acceleration time is short.
Solution to Problem
• According to the present invention, an elevator control apparatus which controls an elevator as a control target includes:
• a model arithmetic operation unit to which a velocity instruction for an electric motor provided to the elevator is input and which obtains a model velocity and a model torque predicted for the control target by arithmetic operation using a preset inertia such that the model velocity follows the velocity instruction;
• a velocity detector which detects an actual velocity being an actual velocity of rotation of the electric motor;
• a compensation arithmetic operation unit which calculates an error compensation torque by using a plurality of predetermined parameters and a velocity deviation between the model velocity calculated by the model arithmetic operation unit and the actual velocity detected by the velocity detector;
• a torque instruction calculation unit which calculates a torque instruction from the model torque calculated by the model arithmetic operation unit and the error compensation torque calculated by the compensation arithmetic operation unit;
• a torque controller which controls and drives the electric motor such that a torque generated by the electric motor coincides with the torque instruction calculated by the torque instruction calculation unit;
• an inertia error prediction unit, which calculates an intermediate value representing a pre-convergent inertia error, being an error of the preset inertia value with respect to an actual inertia value, based on the velocity deviation between the model velocity calculated by the model arithmetic operation unit and the actual velocity detected by the velocity detector, and which predicts a post-convergent inertia error, being the convergence value, based on the intermediate value calculated; and
• a parameter correction unit which corrects the preset inertia value to be used by the model arithmetic operation unit, by using the post-convergent inertia error predicted by the inertia error prediction unit.
• The inertia error prediction unit includes
  a first arithmetic operation part which calculates the intermediate value based on integral arithmetic operation of the velocity deviation of a period where an acceleration build-up state and a constant acceleration state or a car of the elevator are consecutive, and
  a second arithmetic operation part which predicts the post-convergent inertia error based on the intermediate value calculated by the first arithmetic operation part.
• The second arithmetic operation part, by using inverse Laplace transform, calculates a rate-corresponding value, corresponding to a rate of the intermediate value calculated by the first arithmetic operation part to the value of the inertia error that should converge, and predicts the post-convergent inertia error based on the intermediate value and the rate-corresponding value.
• According to the present invention, an elevator control apparatus which controls an elevator as a control target includes:
• a model arithmetic operation unit to which a velocity instruction for an electric motor provided to the elevator is input and which obtains a model velocity and a model torque predicted for the control target by arithmetic operation using a preset inertia such that the model velocity follows the velocity instruction;
• a velocity detector which detects an actual velocity being an actual velocity of rotation of the electric motor;
• a compensation arithmetic operation unit which calculates an error compensation torque by using a plurality of predetermined parameters and a velocity deviation between the model velocity calculated by the model arithmetic operation unit and the actual velocity detected by the velocity detector;
• a torque instruction calculation unit which calculates a torque instruction from the model torque calculated by the model arithmetic operation unit and the error compensation torque calculated by the compensation arithmetic operation unit;
• a torque controller which controls and drives the electric motor such that a torque generated by the electric motor coincides with the torque instruction calculated by the torque instruction calculation unit;
• an inertia error prediction unit which predicts a post-convergent inertia error, being an error of the preset inertia value with respect to an actual inertia value, based on a velocity deviation, being a velocity deviation between the model velocity calculated by the model arithmetic operation unit and the actual velocity detected by the velocity detector, of a period of an acceleration build-up state of a car of the elevator; and
• a parameter correction unit which corrects the preset inertia value to be used by the model arithmetic operation unit, by using the post-convergent inertia error predicted by the inertia error prediction unit.
• The inertia error prediction unit predicts the post-canvergent inertia error by applying the final-value theorem to the velocity deviation of the period of the acceleration build-up state.
• The inertia error prediction unit predicts the post-convergent inertia error by applying inverse Laplace transform to the velocity deviation of the period of the acceleration build-up state.
Advantageous Effects of Intention
• According to the present invention, a high-following-capability elevator control apparatus can be provided that calculates an accurate inertia even during a transient response before convergence of the car acceleration or even in a situation where the car cannot accelerate constantly, and that uses the calculated inertia.
Description of Embodiments Embodiment 1
• An elevator control apparatus 110 according to Embodiment 1 will be described with reference to Figs. 1 to 4. Embodiment 1 relates to the elevator control apparatus 110 which, even during a transient response before convergence of car acceleration, calculates the convergence rate (E_s(t) to be described later) of the acceleration so as to calculate an accurate inertia quickly, and operates an elevator car with high following capability. The characteristic feature of the elevator control apparatus 110 of Embodiment 1 resides in an inertia error prediction unit 80A, and particularly a second arithmetic operation part 82a of the inertia error prediction unit 80A. The second arithmetic operation part 82a has a function of receiving a pre-convergent intermediate inertia error from a first arithmetic operation part 81a and predicts a post-convergent inertia error. Note that "during a transient response" or "a transient state" in Embodiment 1 refers to a state before the feedback converges not only during acceleration build-up but also during constant acceleration.
• Fig. 1 is a block diagram showing the configuration of the elevator control apparatus 110 of Embodiment 1. Fig. 2 is a diagram showing the detailed configuration of the elevator control apparatus 110 of Embodiment 1.
• As shown in Fig. 1, the elevator control apparatus 110 includes a velocity instruction input unit 10, a parameter setting unit 20, a model arithmetic operation unit 30, a velocity detector 40, a compensation arithmetic operation unit 50, a torque instruction calculation unit 60, a torque controller 70, the inertia error prediction unit 80A, and a parameter setting unit 90. The inertia error prediction unit 80A includes the first arithmetic operation part 81a and the second arithmetic operation part 82a. The torque controller 70 controls a power device 95, and the velocity detector 40 detects an actual velocity ω_M of the power device 95.
• Fig. 2 will be described. Fig. 2 does not show the parameter setting unit 20 shown in Fig. 1.
• Referring to Fig. 2, a control target 200 incorporates the power device 95, the torque controller 70 which controls the torque of the power device, the velocity detector 40 which detects the velocity of the power device, a mechanical system 201 as a load, and a disturbance torque τ _L 202 which acts on the control target.
1. (1) The velocity instruction input unit 10 inputs a velocity instruction for the power device (electric motor) provided to the elevator which is the control target.
2. (2) The parameter setting unit 20 sets parameters used for arithmetic operation based on an inertia value estimated in the model arithmetic operation unit 30 in advance.
3. (3) Using an inertia value J_A estimated in advance, the model arithmetic operation unit 30 calculates a torque q_A (may also be called a model torque q_A) which is necessary to cause the control target 200 to follow a velocity instruction ω_ref, and a velocity ω_A (may also be called a model velocity ω_A) estimated for the control target 200 when the torque q_A is) input, and outputs the torque q_A and velocity ω_A. In other words, the model arithmetic operation unit 30 obtains the model velocity ω_A and model torque q_A estimated for the control target by arithmetic operation such that the velocity ω_A follows the velocity instruction ω_ref.
4. (4) The velocity detector 40 detects the actual velocity which is the velocity of rotation of the power device (electric motor).
5. (5) The compensation arithmetic operation unit 50 calculates an error compensation torque q_c based on the difference between the model velocity ω_A and actual velocity ω_M. More specifically, the compensation arithmetic operation unit 50 multiplies a velocity deviation E between the estimated velocity ω_A and actual velocity ω_M, and the integral of the velocity deviation E, by a preset proportion gain K_sp2 and an integral gain K_si2, and outputs a compensation value (error compensation torque q_c).
6. (6) The torque instruction calculation unit 60 calculates a torque instruction q_r from the model torque q_A and error compensation torque q_c. More specifically, the torque instruction calculation unit 60 inputs the torque q_A from the model arithmetic operation unit 52 and the error compensation torque q_c from the compensation arithmetic operation unit 53, and determines the torque instruction q_r to be input to the control target.
7. (7) The torque controller 70 drives the power device 95 by controlling it such that a torque generated by the power device 95 (electric motor) coincides with the torque instruction q_r.
8. (8) The first arithmetic operation part 81a of the inertia error prediction unit 80A outputs the error (ΔJ_M to be described later) between the estimated inertia value J_A and an actual inertia J_M based on the integral value of the deviation E between the estimated velocity ω_A and the actual velocity ω_M of the control target 200. The second arithmetic operation part 82a inputs pre-convergent data from the first arithmetic operation part 81a, and calculates a convergence prediction value of the inertia error ΔJ_M.
9. (9) Based on the output from the second arithmetic operation part 82a, the parameter correction unit 90 corrects the estimated inertia value J_A to be used by the model arithmetic operation unit 30, and the gains K_sp2 and K_si2 of the compensation arithmetic operation unit 50.
(Outline of Operation)
• The outline of the operation will be described with reference to Fig. 2. First, upon input of the torque instruction q_r to the control target, in the control target, the torque of the power device is controlled by the torque controller 70 so as to coincide with the input torque instruction q_r. Then, the electric motor, and the mechanical system as the load, are driven, and the actual velocity ω_M of the electric motor is detected by the velocity detector 40, and output to the outside. In other words, the transfer characteristics from the torque instruction q_r to the actual velocity ω_M may be described as G(s).
1. (1) First, the parameter setting unit 20 sets parameters (J_A, K_SP1, and the like) to be used for arithmetic operation, in the model arithmetic operation unit 30 and the compensation arithmetic operation unit 50 based on the inertia value J_A estimated in the model arithmetic operation unit 30 in advance.
2. (2) The velocity instruction input unit 10 inputs to the model arithmetic operation unit 30 the velocity instruction ω_ref for the power device 95 (electric motor) provided to elevator as the control target.
3. (3) Upon input of the velocity instruction ω_ref from the velocity instruction input unit 10, the model arithmetic operation unit 30 calculates the model velocity ω_A and torque q_A, which are estimated for the control target, such that the model velocity ω_A follows the velocity instruction ω_ref, and outputs the model torque q_A and model velocity ω_A.
4. (4) The velocity detector 40 detects the actual velocity ω_M of the power device 95, and outputs it.
5. (5) The compensation arithmetic operation unit 50 inputs the model velocity ω_A from the model arithmetic operation unit 30 and the actual velocity ω_M from the velocity detector 40, and calculates the error compensation torque q_c based on the deviation E between the model velocity ω_A and actual velocity ω_M and outputs it. More specifically, a comparative controller 51 outputs a signal obtained by multiplying the deviation E between the model velocity ω_A and actual velocity ω_M by the preset proportion gain K_sp2, and an integral controller 52 outputs a signal obtained by multiplying the deviation E between the model velocity ω_A and actual velocity ω_M by the preset integral gain K_si2 and integrating the obtained product. The sum of the output from the comparative controller 51 and the output from the integral controller 52 is defined as the error compensation torque q_c. In other words, a PI (proportional integration) operation is performed.
6. (6) The torque instruction calculation unit 60 inputs the model torque q_A from the model arithmetic operation unit 30 and the error compensation torque q_c from the compensation arithmetic operation unit 50, and calculates the torque instruction q_r from the model torque q_A and error compensation torque q_c and outputs it. More specifically, the torque instruction calculation unit 60 inputs the sum of the error compensation torque q_c and the model torque q_A which is calculated by the model arithmetic operation unit 102 to the control target as the torque instruction q_r. The mechanical system of the control target 101 is driven by the torque instruction q_r.
7. (7) The torque controller 70 inputs the torque instruction q_r from the torque instruction calculation unit 60, and drives the power device 95 by controlling it such that the torque generated by it coincides with the torque instruction q_r.
8. (8) The first arithmetic operation part 81a inputs the model velocity ω_A calculated by the model arithmetic operation unit 30 and the actual velocity ω_M detected by the velocity detector 40. The first arithmetic operation part 81a calculates an intermediate value representing the pre-convergent-to-convergence-value inertia error ΔJ_M, being an error of the preset inertia value J_A with respect to the actual inertia value J_M, based on the velocity deviation E between the model velocity ω_A and actual velocity ω_M. The second arithmetic operation part 82 predicts a post-convergent inertia error, being a convergence value, based on the calculated intermediate value. The detailed operation of the first arithmetic operation part 81a and second arithmetic operation part 82a will be described later.
9. (9) Using the post-convergent inertia error predicted by the second arithmetic operation part 82a, the parameter correction unit corrects a plurality of parameters including the preset inertia value to be used by the model arithmetic operation unit 30, and a plurality of predetermined parameters to be used by the compensation arithmetic operation unit 50.
(Process by First Arithmetic Operation Part 81a: Transfer Function from Velocity Input to Velocity Deviation)
• The operation of the first arithmetic operation part 81a will now be described. The first arithmetic operation part 81a inputs the actual velocity ω_M from the velocity detector 40 and the model velocity ω_A firm the model arithmetic operation unit 30, and performs an arithmetic operation using the actual velocity ω_M and the model velocity ω_A. In Fig. 2, the transfer function from the velocity instruction ω_ref, which the elevator should follow, to the estimated model velocity ω_A and the actual velocity ω_M, of the hoist (power device 95) can be expressed as the following equation (1). The estimated model velocity ω_A is expressed by equation (1):
• $$\begin{array}{l}\left[\mathrm{<figure-callout\; id="2"\; label="Formula"\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">Formula</figure-callout>}2\right]\\ {\mathrm{\omega }}_A=\frac{K_{\mathit{sp}1}}{J_A+K_{\mathit{sp}1}}{\mathrm{\omega }}_{\mathit{ref}}\end{array}$$

  
• The actual velocity ω_M is expressed by the equation (2):
• $$\begin{array}{l}\\ {\mathrm{\omega }}_M=\frac{J_{\mathrm{<figure-callout\; id="2"\; label="A\; \; s"\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">A</figure-callout>}}{s}^{}{}^2+K_{\mathit{sp}2}s+K_{\mathit{si}2}}{J_M{s}^2+K_{\mathit{sp}2}s+K_{\mathit{si}2}}\frac{K_{\mathit{sp}1}}{J_A+K_{\mathit{sp}1}}{\mathrm{\omega }}_{\mathit{ref}}\\ \end{array}+\frac{s}{J_M{s}^2+K_{\mathit{sp}2}s+K_{\mathit{si}2}}{\mathrm{\tau }}_L$$

  
• When the model inertia J_A (estimated inertia) and the error ΔJ_M of the actual inertia J_M are introduced, the above relationship can be expressed by the following equation (3):
• $$\begin{array}{l}\left[\mathrm{Formula}4\right]\\ J_M=J_A+\mathrm{\Delta }{J}_M\end{array}$$

  
• The velocity deviation E between the model velocity ω_A and actual velocity ω_M is as expressed by the following equation (4):
• $$\begin{array}{l}\left[\mathrm{Formula}5\right]\\ E={\mathrm{\omega }}_M-{\mathrm{\omega }}_A\\ =\frac{\mathrm{\Delta }{J}_M{s}^2}{J_M{s}^2+K_{\mathit{sp}2}s+K_{\mathit{si}2}}\frac{K_{\mathit{sp}1}}{J_{A}s+K_{\mathit{sp}1}}{\mathrm{\omega }}_{\mathit{ref}}+\frac{s}{J_M{s}^2+K_{\mathit{sp}2}s+K_{\mathit{si}2}}{\mathrm{\tau }}_L\end{array}$$

  
(Method of Identifying Inertia Error by First Arithmetic Operation Part 81a)
• Assuming that the input is a constant velocity input (ω_ref = v[m/s]), from the final-value theorem, the integral value of the velocity deviation E is expressed by the following equation (5):
• $$\begin{array}{l}\left[\mathrm{Formula}6\right]\\ \underset{s\to 0}{\begin{array}{c}\lim \end{array}}s\frac{1}{s}E\left(\mathit{s}\right){|}_{{\mathrm{\omega }}_{\mathit{ref}}=\frac{v}{s}}\\ =\underset{s\to 0}{\begin{array}{c}\lim \end{array}}\left(\frac{\mathrm{\Delta }{J}_M{s}^2}{J_M{s}^2+K_{\mathit{sp}2}s+K_{\mathit{si}2}}\frac{K_{\mathit{sp}1}}{J_{A}s+K_{\mathit{sp}1}}\frac{v}{s}+\frac{s}{J_M{s}^2+K_{\mathit{sp}2}s+K_{\mathit{si}2}}{\mathrm{\tau }}_L\right)\\ =\underset{s\to 0}{\begin{array}{c}\lim \end{array}}\frac{s}{K_{\mathit{si}2}}{\mathrm{\tau }}_L\end{array}$$

  
• From equation (5), when the disturbance torque τ_L is a constant value K_L, from "τ_L → K_L/s", the integral value of the velocity deviation E converges to K_L/K_Si2. Hence, the value of the disturbance torque τ_L can be identified by observing the velocity deviation E during constant velocity. From the final-value theorem, the integral of a steady-state deviation, when a constant acceleration input (ω_ref= αt[m/s²]) is applied, is expressed by the following equation (6):
• $$\begin{array}{l}\left[\mathrm{Formula}7\right]\\ \underset{s\to 0}{\begin{array}{c}\lim \end{array}}s\frac{1}{s}E\left(\mathit{s}\right){|}_{{\mathrm{\omega }}_{\mathit{ref}}=\frac{\mathrm{\alpha }}{s^2}}\\ =\underset{s\to 0}{\begin{array}{c}\lim \end{array}}\left(\frac{\mathrm{\Delta }{J}_M{s}^2}{J_M{s}^2+K_{\mathit{sp}2}s+K_{\mathit{si}2}}\frac{K_{\mathit{sp}1}}{J_{A}s+K_{\mathit{sp}1}}\frac{\mathrm{\alpha }}{s^2}+\frac{s}{J_M{s}^2+K_{\mathit{sp}2}s+K_{\mathit{si}2}}{\mathrm{\tau }}_L\right)\\ =\frac{\mathrm{\Delta }{J}_M\mathrm{\alpha }}{K_{\mathit{si}2}}+\underset{s\to 0}{\begin{array}{c}\lim \end{array}}\frac{s}{K_{\mathit{si}2}}{\mathrm{\tau }}_L\end{array}$$

  
• From equation (6), when the disturbance torque τ_L is a constant value K_L, the final value of the integral of the velocity deviation E converges to a constant value, i.e., the value expressed by the following equation (B):
• $$\begin{array}{l}\left[\mathrm{<figure-callout\; id="8"\; label="Formula"\; filenames="imgf0003.png"\; state="\{\{state\}\}">Formula</figure-callout>}8\right]\\ \frac{\mathrm{\Delta }{J}_M\mathrm{\alpha }}{K_{\mathit{si}2}}+\frac{{\mathrm{\tau }}_L}{K_{\mathit{si}2}}\end{array}$$

  
• From equation (5), the disturbance torque value τ_L can be identified during a constant state, and other parameters (integral gain K_si2, instruction velocity α) are known. From the foregoing, during acceleration of the elevator car, the first arithmetic operation part 81a can identify the inertia error ΔJ_M by observing the velocity deviation E between the model velocity ω_A of the model arithmetic operation unit 30 and the actual velocity ω_M detected by the velocity detector 40 and calculating the integral value of the velocity deviation E. The value of the constant disturbance τ_L can be fixed by observing the velocity deviation E of constant velocity.
(Method of Predicting Convergence value of Inertia Error by Second Arithmetic Operation Part 82a)
• That the inertia error ΔJ_M calculated by the first arithmetic operation part 81a finally converges to a certain value has been described so far. A method of predicting the convergence value of the inertia from a pre-convergent transient response by the second arithmetic operation part 82a will now be described.
• Assume that ω_ref is a constant acceleration input (acceleration build-up with which when t = t₁, the acceleration is α). Then, the integral value of the velocity deviation E is expressed by the following equation (7):
• $$\begin{array}{l}\left[\begin{array}{c}\mathrm{Formula}9\end{array}\right]\\ \frac{E}{s}{|}_{{\mathrm{\omega }}_{\mathit{ref}}=\frac{\mathrm{<figure-callout\; id="1"\; label="\alpha \; t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}{t_{}}}=\frac{\mathrm{\Delta }{J}_M}{J_M{s}^2+K_{\mathit{sp}2}s+K_{\mathit{si}2}}\frac{K_{\mathit{sp}1}}{J_{A}s+K_{\mathit{sp}1}}\frac{\mathrm{<figure-callout\; id="1"\; label="\alpha \; t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}{t_{}}+\frac{s}{J_M{s}^2+K_{\mathit{sp}2}s+K_{\mathit{si}2}}{\mathrm{\tau }}_L\end{array}$$

  
• Assume that the respective coefficients of the above equation (7) are defined as b₁ to b₅. Then, when the disturbance torque τ_L is a constant value, the time response of the integral value of the velocity deviation E can be obtained as the following equation (8) by subjecting equation (7) to inverse Laplace transform:
• $$\begin{array}{l}\left[\mathrm{Formula}10\right]\\ L^{-1}\frac{1}{s}E{|}_{{\mathrm{\omega }}_{\mathit{ref}}=\frac{\mathrm{<figure-callout\; id="1"\; label="\alpha \; t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}{t_{}}}={\mathrm{<figure-callout\; id="1"\; label="b"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">b</figure-callout>}}_1{e}^{-{\mathrm{<figure-callout\; id="1"\; label="s"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">s</figure-callout>}}_{1}t}+{\mathrm{<figure-callout\; id="2"\; label="b"\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">b</figure-callout>}}_2{e}^{-{\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">s</figure-callout>}}_{2}t}+{\mathrm{<figure-callout\; id="3"\; label="b"\; filenames="imgf0004.png"\; state="\{\{state\}\}">b</figure-callout>}}_3{e}^{-{\mathrm{<figure-callout\; id="3"\; label="s"\; filenames="imgf0004.png"\; state="\{\{state\}\}">s</figure-callout>}}_{3}t}+{\mathrm{<figure-callout\; id="4"\; label="b"\; filenames="imgf0004.png"\; state="\{\{state\}\}">b</figure-callout>}}_{4}t+b_5\end{array}$$

  
• The right side of equation (8) will be substituted by f(t) so that it can be used in a following equation. In equation (8), S₁, S₂, and S₃ are the roots of the transfer function from the velocity instruction input ω_ref to the velocity deviation E. When calculating these roots, the inertia J_M of the actual control target is necessary. If a preset inertia J_A is used, the resultant convergence rate does not differ largely. Thus, calculation can be performer by using J_A in place of J_M. Laplace transform of input of constant acceleration build-up to constant acceleration, which is employed in an elevator, is given by the following equation (9):
• $$\begin{array}{l}\left[\begin{array}{c}\mathrm{Formula}11\end{array}\right]\\ L\left({\mathrm{\omega }}_{\mathit{ref}}\right)=L\left(\mathrm{\int }\frac{\mathrm{<figure-callout\; id="1"\; label="\alpha \; t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}{t_{}}t-\frac{\mathrm{<figure-callout\; id="1"\; label="\alpha \; t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}{t_{}}\left(t-{\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">t</figure-callout>}}_1\right)u\left(t-{\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">t</figure-callout>}}_1\right)\mathit{dt}\right)=\frac{\mathrm{<figure-callout\; id="1"\; label="\alpha \; t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}{t_{}}\left(1-e^{-{\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">t</figure-callout>}}_{1}s}\right)\end{array}$$

  
• From the foregoing, the time response of the integral of the velocity deviation E, when an input of constant acceleration build-up to constant acceleration is applied, can be expressed by the following equation (10):
• $$\begin{array}{l}\left[\begin{array}{c}\mathrm{Formula}12\end{array}\right]\\ L^{-1}\left(\frac{1}{s}E{|}_{{\mathrm{\omega }}_{\mathit{ref}}=\frac{\mathrm{<figure-callout\; id="1"\; label="\alpha \; t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}{t_{}}\left(1-e^{-{\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">t</figure-callout>}}_{1}s}\right)}\right)=f\left(t\right)-u\left(t-{\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">t</figure-callout>}}_1\right)f\left(t-{\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">t</figure-callout>}}_1\right)\\ ={\mathrm{<figure-callout\; id="1"\; label="b"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">b</figure-callout>}}_1\left(1-{\mathrm{<figure-callout\; id="1"\; label="e\; s"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">e</figure-callout>}}^{s_{}}\right)e^{-{\mathrm{<figure-callout\; id="1"\; label="s"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">s</figure-callout>}}_{1}t}+{\mathrm{<figure-callout\; id="2"\; label="b"\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">b</figure-callout>}}_2\left(1-{\mathrm{<figure-callout\; id="2"\; label="e\; s"\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">e</figure-callout>}}^{s_{}}\right)e^{-{\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">s</figure-callout>}}_{2}t}+{\mathrm{<figure-callout\; id="3"\; label="b"\; filenames="imgf0004.png"\; state="\{\{state\}\}">b</figure-callout>}}_3\left(1-{\mathrm{<figure-callout\; id="3"\; label="e\; s"\; filenames="imgf0004.png"\; state="\{\{state\}\}">e</figure-callout>}}^{s_{}}\right)e^{-{\mathrm{<figure-callout\; id="3"\; label="s"\; filenames="imgf0004.png"\; state="\{\{state\}\}">s</figure-callout>}}_{3}t}+{\mathrm{<figure-callout\; id="4"\; label="b"\; filenames="imgf0004.png"\; state="\{\{state\}\}">b</figure-callout>}}_4{\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">t</figure-callout>}}_1\end{array}$$

  
• From equation (10), the integral value of the velocity deviation E converges over time to a value of b₄t₁. Naturally, b₄t₁ in equation (10), when calculated, becomes equal to the convergence value (the following equation (C)) in equation (6):
• $$\begin{array}{l}\left[\mathrm{Formula}13\right]\\ \frac{\mathrm{\Delta }{J}_M\mathrm{\alpha }}{K_{\mathit{si}2}}+\frac{{\mathrm{\tau }}_L}{K_{\mathit{si}2}}\end{array}$$

  
• From the foregoing, since the transient response of the inertia error of an elevator at an arbitrary time, when an input of acceleration build-up to acceleration is applied, can be calculated. Hence, the value of the post-convergent inertia error can be obtained by observing the integral value of the pre-convergent velocity deviation. More specifically, the second arithmetic operation part 82a calculates (predicts) the inertia J_M of the elevator by the following equation (11). Assume that the integral value of the velocity deviation from "acceleration build-up" to "constant acceleration ", which is calculated by the first arithmetic operation part 81a is E_s(t), and that the value of disturbance measured during constant velocity and calculated by the first arithmetic operation part 81a is τ_L. Then, the actual inertia J_M is expressed by the following equation (11):
• $$\begin{array}{l}\left[\begin{array}{c}\mathrm{Formula}14\end{array}\right]\\ J_M=\left(\frac{E_s\left(\mathrm{<figure-callout\; id="1"\; label="t\; b"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">t</figure-callout>},\right)}{b_{}}{\mathrm{<figure-callout\; id="4"\; label="b"\; filenames="imgf0004.png"\; state="\{\{state\}\}">b</figure-callout>}}_4{\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">t</figure-callout>}}_1{K}_{\mathit{si}2}-{\mathrm{\tau }}_L\right)\frac{1}{\mathrm{\alpha }}+J_A\end{array}$$

  
• The operation of the first arithmetic operation part 81a and second arithmetic operation part 82a will be described with reference to Fig. 3. Note that (a) of Fig. 3 shows how the inertia error converges. In (a) of Fig. 3, the axis of abscissa represents the time, and the axis of ordinate represents the inertia error. Note that (b) of Fig. 3 is a velocity diagram. In (b) of Fig. 3, the axis of abscissa represents the time, and the axis of ordinate represents the elevator velocity. As shown in (a) of Fig. 3, the inertia error converges to a certain value over time. The first arithmetic operation part 81a calculates a pre-convergent intermediate inertia error ΔJ_M (pre-convergent). In this case, the pre-convergent intermediate inertia error ΔJ_M (pre-convergent) is the integral value E_s(t) of the velocity deviation from acceleration build-up to constant acceleration. The first arithmetic operation part 81a calculates the "inertia error ΔJ_M (pre-convergent) = integral value E_s(t)" in (a) of Fig. 3. The second arithmetic operation part 82a obtains f(s) (rate-corresponding value) which corresponds to the rate of the intermediate inertia error "ΔJ_M (pre-convergent)" to the to-be-converging inertia error "ΔJ_M (post-convergent)". Note that f(s) is the right side of equation (10). By using the result of equation (10), the second arithmetic operation part 82a calculates (predicts) the inertia error ΔJ_M (post-convergent), being the convergence value, by dividing the "integral value E_s(t) of the velocity deviation" corresponding to the intermediate value of the pre-convergent intermediate inertia error by the rate f(s) with respect to the inertia error ΔJ_M (post-convergent) to which the intermediate value (E_s(t)) should converge, as indicated in equation (11).
• As described above, the second arithmetic operation part 82a, by using inverse Laplace transform, calculates (equation (10)) the rate-corresponding value f(s), corresponding to the rate of the intermediate value to the value of the inertia error ΔJ_M (post-convergent) that should converge as indicated in equation (10), in accordance with a predetermined calculation procedure, and calculates (predicts) the post-convergent inertia error ΔJ_M (post-convergent) by equation (11) based on the intermediate value ΔJ_M (pre-convergent) and the rate-corresponding value f(s).
(Procedure of Calculating Inertia of Control Target)
• The procedure of calculating the inertia of the control target will be described with reference to Fig. 4. Fig. 4 is a graph showing a velocity instruction value 4. The axis of abscissa represents time t, and the axis of ordinate represents velocity V. As shown in Fig. 3, the velocity instruction value has a region 1 which is an acceleration build-up period, a region 2 which is a constant acceleration period, and a region 3 which is a constant velocity period.
1. (1) When the car is operated by the velocity instruction shown in Fig. 4, the inertia of the control target can be obtained from the arithmetic operation results of the first arithmetic operation part 81a and second arithmetic operation part 82a. More specifically, first, the "car" is driven with parameters set by the parameter setting unit 20 in advance, to travel during a time and at a velocity shown in Fig. 4. In this travel, the inertia of the control target can be obtained from the arithmetic operation results of the first arithmetic operation part 81a and second arithmetic operation part 82a.
2. (2) The second arithmetic operation part 82a calculates the inertia error, including a disturbance, from the integral value of the velocity deviation which is received from the first arithmetic operation part 81a during "the constant acceleration build-up region 1 to the constant acceleration region 2 ".
3. (3) The first arithmetic operation part 81a calculates a disturbance-corresponding element during the constant travel region 3. After that, the disturbance-corresponding element is removed from the inertia error including the disturbance, so that the inertia value of the control target is obtained.
• As described above, the elevator control apparatus 110 according to Embodiment 1 is provided with the inertia error prediction unit 80A that predicts the post-convergent inertia error from the intermediate inertia error. Thus, even during a transient state where equation (A) of the background art is not established, an accurate inertia error can be calculated. Hence, the elevator control apparatus 110 with high following capability can be provided.
• As described above, the elevator control apparatus 110 according to Embodiment 1 is provided with the second arithmetic operation part 82a which receives the intermediate inertia error (E_s(t) corresponding to the inertia error) from the first arithmetic operation part 81 a that calculates the pre-convergent intermediate inertia error based on the integral arithmetic operation of the velocity deviation, and which predicts the post-convergent inertia error. Thus, even during a transient state where equation (A) of the background art is not established, an accurate inertia error can be calculated. Hence, the elevator control apparatus 110 with high following capability can be provided.
Embodiment 2
• An elevator control apparatus 120 according to Embodiment 2 will be described with reference to Figs. 5 and 6. The elevator control apparatus 120 of Embodiment 2 is different from the elevator control apparatus 110 of Embodiment 1 in that an inertia error prediction unit 80B has only one first arithmetic operation part 81b and in that the first arithmetic operation part 81b predicts the inertia error in the acceleration build-up region 1 (Fig. 4). Fig. 5 is a block diagram showing the configuration of the elevator control apparatus 120, and corresponds to Fig. 1. Fig. 5 is different from Fig. 1 in that the inertia error prediction unit 80B has only the first arithmetic operation part 81b. Fig. 6 shows the detailed configuration of the elevator control apparatus 120, and corresponds to Fig. 2. Fig. 6 is different from Fig. 2 in that the inertia error prediction unit 80B has only the first arithmetic operation part 81b. The first arithmetic operation part 8 1 b calculates an error between an estimated inertia J_A and an actual inertia J_M based on a model velocity ω_A and an actual velocity ω_M, and ends inertia identification value calculation during the period of constant acceleration build-up.
(Transfer Function from Input of Velocity Instruction ω_ref to Velocity Deviation E)
• The operation of the first arithmetic operation part 81b will be described. The first arithmetic operation part 81b inputs the actual velocity ω_M from a velocity detector 40 and the model velocity ω_A from a model arithmetic operation unit 30, and calculates the inertia error. Referring to Fig. 6, the transfer function from a velocity instruction ω_ref, which the elevator should follow, to the estimated model velocity ω_A and the actual velocity ω_M, of a hoist (power device 95) is as follows. Namely, the estimated model velocity ω_A is identical to that of equation (1) of Embodiment 1. The actual velocity ω_M is identical to that of equation (2) of Embodiment 1. The relational expression between the model velocity ω_A and actual velocity ω_M, when an error ΔJ_M between the model inertia J_A and actual inertia J_M is introduced, is identical to equation (3) of Embodiment 1. A velocity deviation E is identical to that of equation (4) of Embodiment 1.
• From the final-value theorem, the above-mentioned equation (4) yields the following equation (21):
• $$\begin{array}{l}\left[\mathrm{Formula}15\right]\\ \underset{\mathrm{s}\mathrm{\to }\mathrm{0}}{\begin{array}{c}\lim \end{array}}\mathrm{sE}\left(\mathrm{s}\right){\mathrm{|}}_{{\mathrm{\omega }}_{\mathrm{ref}}\mathrm{=}\frac{\mathrm{<figure-callout\; id="1"\; label="\alpha \; t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}{{\mathrm{t}}_{}}}\\ \mathrm{=}\underset{\mathrm{s}\mathrm{\to }\mathrm{0}}{\begin{array}{c}\lim \end{array}}\left(\frac{\mathrm{\Delta }{\mathrm{J}}_{\mathrm{M}}}{{\mathrm{J}}_{\mathrm{M}}{\mathrm{s}}^{\mathrm{2}}\mathrm{+}{\mathrm{K}}_{\mathrm{sp}\mathrm{2}}\mathrm{s}\mathrm{+}{\mathrm{K}}_{\mathrm{si}\mathrm{2}}}\frac{{\mathrm{K}}_{\mathrm{sp}\mathrm{1}}}{{\mathrm{J}}_{\mathrm{A}}\mathrm{s}\mathrm{+}{\mathrm{K}}_{\mathrm{sp}\mathrm{1}}}\frac{\mathrm{<figure-callout\; id="1"\; label="\alpha \; t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}{{\mathrm{t}}_{}}\mathrm{+}\frac{{\mathrm{s}}^{\mathrm{2}}}{{\mathrm{J}}_{\mathrm{M}}{\mathrm{s}}^{\mathrm{2}}\mathrm{+}{\mathrm{K}}_{\mathrm{sp}\mathrm{2}}\mathrm{s}\mathrm{+}{\mathrm{K}}_{\mathrm{si}\mathrm{2}}}{\mathrm{\tau }}_{\mathrm{L}}\right)\\ \mathrm{=}\frac{\mathrm{\Delta }{\mathrm{J}}_{\mathrm{M}}\mathrm{\alpha }}{{\mathrm{K}}_{\mathrm{si}\mathrm{2}}{\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">t</figure-callout>}}_{\mathrm{1}}}\mathrm{+}\underset{\mathrm{s}\mathrm{\to }\mathrm{0}}{\begin{array}{c}\lim \end{array}}\frac{{\mathrm{s}}^{\mathrm{2}}}{{\mathrm{K}}_{\mathrm{si}\mathrm{2}}}{\mathrm{\tau }}_{\mathrm{L}}\end{array}$$

  
• From equation (21), if a disturbance torque τ_L is a constant value, the velocity deviation converges to a certain value expressed by the following equation (22):
• $$\begin{array}{l}\left[\mathrm{Formula}16\right]\\ \frac{\mathrm{\Delta }{\mathrm{J}}_{\mathrm{M}}\mathrm{\alpha }}{{\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">t</figure-callout>}}_{\mathrm{1}}{\mathrm{K}}_{\mathrm{si}\mathrm{2}}}\end{array}$$

  
• From equation (22), the inertia error ΔJ_M can be identified by observing the velocity deviation during constant acceleration build-up.
• As described above, the inertia error prediction unit 80B predicts an inertia error which has converged to the convergence value, by applying the final-value theorem to the velocity deviation E of the period of the acceleration build-up state.
(Procedure of Calculating Inertia of Control Target)
• A procedure of calculating the inertia of the control target in case of Embodiment 2 will be described. As to the inertia of the control target, first, the inertia value of the control target is obtained from a velocity deviation which an inertia error arithmetic operation unit receives during a constant acceleration build-up region 1 when the car is caused to travel with parameters preset by a parameter setting unit in advance during a time and at a velocity shown in Fig. 4 (velocity instruction graph). At this time, even if the car is not able to actually travel after a constant acceleration region 2, since the inertia error arithmetic operation unit has already terminated the calculation and identified the actual inertia value, the inertia identification can be performed even in case where the elevator cannot accelerate but stops for some reason. This "procedure of Calculating Inertia of Control Target" also applies to Embodiment 3 to be described later.
• As described above, since the elevator control apparatus 120 of Embodiment 2 is provided with the first arithmetic operation part 81b that identifies the inertia value in the acceleration build-up state, an accurate inertia error can be calculated. Hence, the elevator control apparatus 120 with high following capability can be provided.
Embodiment 3
• An elevator control apparatus 130 according to Embodiment 3 will be described with reference to Figs. 7 and 8. The elevator control apparatus 120 of Embodiment 2 can identify the inertia error in the acceleration build-up region 1 by using the finial value of the velocity deviation E. The inertia error, however, may not converge sufficiently when, e.g., the elevator halts during traveling. For such a case, a second arithmetic operation part 82c which predicts the convergence value of the inertia error is provided separately, so that the inertia can be identified.
• Fig. 7 is a block diagram showing the configuration of the elevator control apparatus 130. The configuration shown in Fig. 7 is identical to that of Fig. 1. However, the processing contents of a first arithmetic operation part 81c and the second arithmetic operation part 82c of the elevator control apparatus 130 are different from the processing contents of the first arithmetic operation part 81a and second arithmetic operation part 82a of the elevator control apparatus 110. Thus, Fig. 7 is provided independently of Fig. 1. Although Fig. 8 corresponds to Fig. 2, it is provided independently of Fig. 2 for the same reason as that for Fig. 7.
• Figs. 7 and 8 show a configuration obtained by adding the second arithmetic operation part to the elevator control apparatus 120 of Figs. 5 and 6, respectively.
• A description will be made with reference to Fig. 8. Referring to Fig. 8, the first arithmetic operation part 81c outputs a difference ΔJ_M between an estimated inertia J_A and an actual inertia J_M based on a velocity deviation E between a model velocity ω_A and an actual velocity ω_M of the control target. The second arithmetic operation part 82c receives a pre-convergent output from the first arithmetic operation part 81 c and calculates the convergence prediction value of the inertia error. Based on the output from the second arithmetic operation part 82c, a parameter correction unit 90 corrects the estimated inertia J_A for a model arithmetic operation unit 30 and the gain of a compensation arithmetic operation unit 50.
(Transient Response of Velocity Deviation)
• The operation will be described hereinafter. From equation (4) indicated in Embodiment 1, the transient response of the velocity deviation E during an input of constant acceleration build-up (acceleration build-up = α/t1[m/s^3]) is expressed as the following equation (31):
• $$\begin{array}{l}\left[\mathrm{Formula}17\right]\\ \mathrm{E}{\mathrm{|}}_{{\mathrm{\omega }}_{\mathrm{ref}}\mathrm{=}\frac{\mathrm{\alpha }}{{\mathrm{t}}_{\mathrm{1}}{\mathrm{s}}^3}}=\frac{\mathrm{\Delta }{\mathrm{J}}_{\mathrm{M}}}{{\mathrm{J}}_{\mathrm{M}}{\mathrm{s}}^{\mathrm{2}}\mathrm{+}{\mathrm{K}}_{\mathrm{sp}\mathrm{2}}\mathrm{s}\mathrm{+}{\mathrm{K}}_{\mathrm{si}\mathrm{2}}}\frac{{\mathrm{K}}_{\mathrm{sp}\mathrm{1}}}{{\mathrm{J}}_{\mathrm{A}}\mathrm{s}\mathrm{+}{\mathrm{K}}_{\mathrm{sp}\mathrm{1}}}\frac{\mathrm{\alpha }}{{\mathrm{t}}_{\mathrm{1}}s}\mathrm{+}\frac{s}{{\mathrm{J}}_{\mathrm{M}}{\mathrm{s}}^{\mathrm{2}}\mathrm{+}{\mathrm{K}}_{\mathrm{sp}\mathrm{2}}\mathrm{s}\mathrm{+}{\mathrm{K}}_{\mathrm{si}\mathrm{2}}}\mathrm{\tau }\\ \mathrm{=}\frac{\mathrm{\Delta }{\mathrm{J}}_{\mathrm{M}}}{{\mathrm{J}}_{\mathrm{M}}}\frac{1}{\left(s+s_1\right)}\frac{1}{\left(s+s_2\right)}\frac{{\mathrm{K}}_{\mathrm{sp}\mathrm{1}}}{{\mathrm{J}}_{\mathrm{A}}\left(s+s_3\right)}\frac{\mathrm{\alpha }}{{\mathrm{t}}_{\mathrm{1}}s}+\frac{s}{{\mathrm{J}}_{\mathrm{M}}}\frac{1}{\left(s+s_1\right)}\frac{1}{\left(s+s_2\right)}\mathrm{\tau }\\ \mathrm{=}\frac{\mathrm{\Delta }{\mathrm{J}}_{\mathrm{M}}{\mathrm{K}}_{\mathrm{sp}\mathrm{1}}\mathrm{\alpha }}{{\mathrm{J}}_{\mathrm{M}}{\mathrm{J}}_{\mathrm{A}}{\mathrm{t}}_{\mathrm{1}}}\frac{1}{\left(s+s_1\right)}{\left[\frac{1}{\left(s+s_2\right)}\frac{1}{\left(s+s_3\right)}\frac{1}{s}\right]}_{s=-s_1}\\ +\frac{\mathrm{\Delta }{\mathrm{J}}_{\mathrm{M}}{\mathrm{K}}_{\mathrm{sp}\mathrm{1}}\mathrm{\alpha }}{{\mathrm{J}}_{\mathrm{M}}{\mathrm{J}}_{\mathrm{A}}{\mathrm{t}}_{\mathrm{1}}}\frac{1}{\left(s+s_2\right)}{\left[\frac{1}{\left(s+s_1\right)}\frac{1}{\left(s+s_3\right)}\frac{1}{s}\right]}_{s=-s_2}\\ +\frac{\mathrm{\Delta }{\mathrm{J}}_{\mathrm{M}}{\mathrm{K}}_{\mathrm{sp}\mathrm{1}}\mathrm{\alpha }}{{\mathrm{J}}_{\mathrm{M}}{\mathrm{J}}_{\mathrm{A}}{\mathrm{t}}_{\mathrm{1}}}\frac{1}{\left(s+s_3\right)}{\left[\frac{1}{\left(s+s_1\right)}\frac{1}{\left(s+s_2\right)}\frac{1}{s}\right]}_{s=-s_3}\\ +\frac{\mathrm{\Delta }{\mathrm{J}}_{\mathrm{M}}{\mathrm{K}}_{\mathrm{sp}\mathrm{1}}\mathrm{\alpha }}{{\mathrm{J}}_{\mathrm{M}}{\mathrm{J}}_{\mathrm{A}}{\mathrm{t}}_{\mathrm{1}}}\frac{1}{s}{\left[\frac{1}{\left(s+s_1\right)}\frac{1}{\left(s+s_2\right)}\frac{1}{\left(s+s_3\right)}\right]}_{s=0}\\ +\frac{s}{{\mathrm{J}}_{\mathrm{M}}}\frac{1}{\left(s+s_1\right)}\mathrm{\tau }{\left[\frac{1}{\left(s+s_2\right)}\right]}_{s=-s_1}+\frac{s}{{\mathrm{J}}_{\mathrm{M}}}\frac{1}{\left(s+s_2\right)}\mathrm{\tau }{\left[\frac{1}{\left(s+s_1\right)}\right]}_{s=-s_2}\end{array}$$

  
• Note that s₁ to s₃ are the poles of the control system. The next equation (32) shows s₁ to s₃.
• $$\begin{array}{l}\left[\mathrm{Formula}\mathrm{18}\right]\\ {\mathrm{<figure-callout\; id="1"\; label="s"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">s</figure-callout>}}_{\mathrm{1}}\mathrm{,}{\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">s</figure-callout>}}_{\mathrm{2}}\mathrm{,}{\mathrm{<figure-callout\; id="3"\; label="s"\; filenames="imgf0004.png"\; state="\{\{state\}\}">s</figure-callout>}}_{\mathrm{3}}\mathrm{=}\frac{\mathrm{-}{\mathrm{K}}_{\mathrm{sp}\mathrm{2}}\mathrm{\pm }\sqrt{{{\mathrm{K}}_{\mathrm{sp}\mathrm{2}}}^{\mathrm{2}}\mathrm{-}\mathrm{4}{\mathrm{J}}_{\mathrm{M}}{\mathrm{K}}_{\mathrm{si}\mathrm{2}}}}{\mathrm{2}{\mathrm{J}}_{\mathrm{M}}}\mathrm{,}-\frac{{\mathrm{K}}_{\mathrm{sp}\mathrm{1}}}{{\mathrm{J}}_{\mathrm{A}}}\end{array}$$

  
• Inverse Laplace transform of equation (31) into a time region equation yields the following equation (33):
• $$\begin{array}{c}\left[\mathrm{Formula}19\right]\\ \begin{array}{c}{\mathrm{L}}^{\mathrm{-}\mathrm{1}}\mathrm{E}{\mathrm{|}}_{{\mathrm{\omega }}_{\mathrm{ref}}\mathrm{=}\frac{\mathrm{<figure-callout\; id="1"\; label="\alpha \; t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}{{\mathrm{t}}_{}}}\mathrm{=}{\mathrm{a}}_{\mathrm{1}}{\mathrm{e}}^{\mathrm{-}{\mathrm{<figure-callout\; id="1"\; label="s"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">s</figure-callout>}}_{\mathrm{1}}\mathrm{t}}\mathrm{+}{\mathrm{a}}_{\mathrm{2}}{\mathrm{e}}^{\mathrm{-}{\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">s</figure-callout>}}_{\mathrm{2}}\mathrm{t}}\mathrm{+}{\mathrm{a}}_{\mathrm{3}}{\mathrm{e}}^{\mathrm{-}{\mathrm{<figure-callout\; id="3"\; label="s"\; filenames="imgf0004.png"\; state="\{\{state\}\}">s</figure-callout>}}_{\mathrm{3}}\mathrm{t}}\mathrm{+}{\mathrm{a}}_{\mathrm{4}}\mathrm{u}\left(\mathrm{t}\right)\mathrm{+}\frac{\mathrm{d}}{\mathrm{dt}}\left({\mathrm{<figure-callout\; id="1"\; label="b"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">b</figure-callout>}}_{\mathrm{1}}{\mathrm{e}}^{\mathrm{-}{\mathrm{<figure-callout\; id="1"\; label="s"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">s</figure-callout>}}_{\mathrm{1}}\mathrm{t}}\mathrm{\tau }\left(\mathrm{t}\right)\right)\mathrm{+}\frac{\mathrm{d}}{\mathrm{dt}}\left({\mathrm{<figure-callout\; id="2"\; label="b"\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">b</figure-callout>}}_{\mathrm{2}}{\mathrm{e}}^{\mathrm{-}{\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="imgf0003.png,imgf0004.png"\; state="\{\{state\}\}">s</figure-callout>}}_{\mathrm{2}}\mathrm{t}}\mathrm{\tau }\left(\mathrm{t}\right)\right)\end{array}\end{array}$$

  
• Note that in equation (33), a₁ to a₄ and b₁ and b₂ are coefficients of equation (31), and u(t) is a unit step function.
• To obtain s₁ two s₃, an actual inertia value J_M is required. When obtaining a convergence rate, use of an estimated inertia value J_A, however, does not lead to a large difference. Therefore, the convergence rate of equation (33) can be calculated by using J_A in place of J_M. Also, a₁ to a₄ are known, and the influence of a disturbance τ_L, as far as it is a constant value, converges to 0 with almost the same convergence rate. Hence, a convergence rate ζ of the time response of the velocity deviation E can be calculated by using equation (33).
  The convergence value is given by the following equation (34):
• $$\begin{array}{l}\left[\mathrm{Formula}\mathrm{20}\right]\\ \mathrm{a}\mathrm{4}\mathrm{=}\frac{\mathrm{\Delta }{\mathrm{J}}_{\mathrm{M}}\mathrm{\alpha }}{{\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">t</figure-callout>}}_{\mathrm{1}}{\mathrm{K}}_{\mathrm{si}\mathrm{2}}}\end{array}$$

  
• This demonstrates that if the velocity deviation of a certain time is E(t) and the convergence rate is ζ(t), the inertia error can be calculated by the following equation (35):
• $$\begin{array}{l}\left[\mathrm{Formula}\mathrm{21}\right]\\ \mathrm{\Delta }{\mathrm{J}}_{\mathrm{M}}\mathrm{=}\mathrm{E}\left(\mathrm{t}\right)\frac{{\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="imgf0002.png,imgf0003.png"\; state="\{\{state\}\}">t</figure-callout>}}_{\mathrm{1}}{\mathrm{K}}_{\mathrm{si}\mathrm{2}}}{\mathrm{\alpha }}\frac{\mathrm{1}}{\mathrm{\zeta }\left(\mathrm{t}\right)}\end{array}$$

  
• As described above, an inertia error prediction unit 80C predicts the post-convergent inertia error by applying inverse Laplace transform to the velocity deviation of the period of an acceleration build-up state.
Brief Description of Drawings
• [Fig. 1] is a block diagram showing the configuration of the elevator control apparatus 110 according to Embodiment 1;
• [Fig. 2] is a diagram showing the detailed configuration of the elevator control apparatus 110 according to Embodiment 1;
• [Fig. 3] shows graphs explaining the operations of the first arithmetic operation part 81 a and second arithmetic operation part 82a according to Embodiment 1;
• [Fig. 4] is a graph showing the velocity instruction value of the elevator according to Embodiment 1;
• [Fig. 5] is a block diagram showing the configuration of the elevator control apparatus 120 according to Embodiment 2;
• [Fig. 6] is a diagram showing the detailed configuration of the elevator control apparatus 120 according to Embodiment 2;
• [Fig. 7] is a block diagram showing the configuration of the elevator control apparatus 130 according to Embodiment 3;
• [Fig. 8] is a diagram showing the detailed configuration of the elevator control apparatus 130 according to Embodiment 3; and
• [Fig. 9] is a diagram showing a prior art.
Reference Signs List
• E velocity deviation, ω_M actual velocity, ω_A model velocity, τ_L disturbance torque, q_A model torque, q_c error compensation torque, q_r torque instruction, 1 acceleration build-up, 2 acceleration, 3 constant velocity, 4 velocity instruction value, 10 velocity instruction input unit, 20 parameter setting unit, 30 model arithmetic operation unit, 40 velocity detector, 50 compensation arithmetic operation unit, 60 torque instruction calculation unit, 70 torque controller, 80 inertia error prediction unit 80, 81 first arithmetic operation part, 82 second arithmetic operation part, 90 parameter correction unit, 95 power device, 110, 120, 130 elevator control apparatus, 200 control target, 301 hoist, 302 velocity detector, 303 actual velocity, 310 commercial power supply, 320 converter, 330 inverter, 340 controller, 350 car, 360 counterweight

Claims (6)

1. An elevator control apparatus which controls an elevator as a control target, comprising:

   a model arithmetic operation unit to which a velocity instruction for an electric motor provided to the elevator is input and which obtains a model velocity and a model torque predicted for the control target by arithmetic operation using a preset inertia such that the model velocity follows the velocity instruction;

   a velocity detector which detects an actual velocity being an actual velocity of rotation of the electric motor;

   a compensation arithmetic operation unit which calculates an error compensation torque by using a plurality of predetermined parameters and a velocity deviation between the model velocity calculated by the model arithmetic operation unit and the actual velocity detected by the velocity detector;

   a torque instruction calculation unit which calculates a torque instruction from the model torque calculated by the model arithmetic operation unit and the error compensation torque calculated by the compensation arithmetic operation unit;

   a torque controller which controls and drives the electric motor such that a torque generated by the electric motor coincides with the torque instruction calculated by the torque instruction calculation unit;

   an inertia error prediction unit, which calculates an intermediate value representing a pre-convergent inertia error, being an error of the preset inertia value with respect to an actual inertia value, based on the velocity deviation between the model velocity calculated by the model arithmetic operation unit and the actual velocity detected by the velocity detector, and which predicts a post-convergent inertia error, being the convergence value, based on the intermediate value calculated; and

   a parameter correction unit which corrects the preset inertia value to be used by the model arithmetic operation unit, by using the post-convergent inertia error predicted by the inertia error prediction unit.
2. The elevator control apparatus according to claim 1,
   wherein the inertia error prediction unit comprises
   a first arithmetic operation part which calculates the intermediate value based on integral arithmetic operation of the velocity deviation of a period where an acceleration build-up state and a constant acceleration state of a car of the elevator are consecutive, and
   a second arithmetic operation part which predicts the post-convergent inertia error based on the intermediate value calculated by the first arithmetic operation part.
3. The elevator control apparatus according to claim 2,
   wherein the second arithmetic operation part, by using inverse Laplace transform, calculates a rate-corresponding value, corresponding to a rate of the intermediate value calculated by the first arithmetic operation part to the value of the inertia error that should converge, and predicts the post-convergent inertia error based on the intermediate value and the rate-corresponding value.
4. An elevator control apparatus which controls an elevator as a control target, comprising:

   a model arithmetic operation unit to which a velocity instruction for an electric motor provided to the elevator is input and which obtains a model velocity and a model torque predicted for the control target by arithmetic operation using a preset inertia such that the model velocity follows the velocity instruction;

   a velocity detector which detects an actual velocity being an actual velocity of rotation of the electric motor;

   a compensation arithmetic operation unit which calculates an error compensation torque by using a plurality of predetermined parameters and a velocity deviation between the model velocity calculated by the model arithmetic operation unit and the actual velocity detected by the velocity detector;

   a torque instruction calculation unit which calculates a torque instruction from the model torque calculated by the model arithmetic operation unit and the error compensation torque calculated by the compensation arithmetic operation unit;

   a torque controller which controls and drives the electric motor such that a torque generated by the electric motor coincides with the torque instruction calculated by the torque instruction calculation unit;

   an inertia error prediction unit which predicts a post-convergent inertia error, being an error of the preset inertia value with respect to an actual inertia value, based on a velocity deviation, being a velocity deviation between the model velocity calculated by the model arithmetic operation unit and the actual velocity detected by the velocity detector, of a period of an acceleration build-up state of a car of the elevator; and

   a parameter correction unit which corrects the preset inertia value to be used by the model arithmetic operation unit, by using the post-convergent inertia error predicted by the inertia error prediction unit.
5. The elevator control apparatus according to claim 4,
   wherein the inertia error prediction unit predicts the post-convergent inertia error by applying the final-value theorem to the velocity deviation of the period of the acceleration build-up state.
6. The elevator control apparatus according to claim 4,
   wherein the inertia error prediction unit predicts the post-convergent inertia error by applying inverse Laplace transform to the velocity deviation of the period of the acceleration build-up state.

Images