JP7157772B2 - Elevator control device and elevator control method - Google Patents

Description

The present invention relates to an elevator control device and an elevator control method.

Elevators such as elevators move up and down (run) by driving a main rope with a motor, which is a hoist.
Conventionally, the hoisting speed of the main rope, which is the running speed, is controlled by PI control (Proportional-Integral control) that combines proportional action and integral action, and the hoist is operated at an appropriate and stable speed.
On the other hand, in recent years, it has been proposed to control the speed of the hoist by non-linear control such as control by AI (artificial intelligence) technology and model predictive control. By controlling the speed of the hoist with non-linear control, it is possible to perform high-performance operation control that realizes high-speed running of the car while minimizing vibrations given to passengers in the car, for example.

Patent Literature 1 describes a technology for an elevator control device having a learning function that stops service on the assumption that an abnormality has occurred in an elevator when fluctuations in travel control parameters such as travel control fall outside the allowable range. ing.

Japanese Patent No. 5289574

In this way, by controlling the speed of the elevator through non-linear control that applies learning functions such as AI technology, high-performance operation control that was impossible with conventional control becomes possible.
However, in nonlinear control using a learning function, there is a problem that stability is not guaranteed because control is performed using calculated values and predicted values based on learning results. In other words, if the calculated values and predicted values based on the learning results are correct, high-performance operation control is performed. Problems such as giving unpleasant vibrations and causing landing errors may occur.

Here, conventionally, as described in Patent Document 1, when the fluctuation of the travel control parameter obtained by the nonlinear control using the learning function is outside the allowable range, that is, when the prediction result is not appropriate, the elevator malfunctions. I was trying to stop the service as it was in the state.
However, it cannot be said that it is preferable for users of buildings and facilities in which the elevators are installed to stop the operation of the elevators, and it is desired that the elevators continue to operate as much as possible.

The present invention provides an elevator control device and an elevator control method that can continue to operate without stopping service even when an elevator malfunctions when non-linear control such as AI control or model predictive control is applied. intended to

In order to solve the above problems, for example, the configurations described in the claims are adopted.
The present application includes a plurality of means for solving the above problems. To give an example, an elevator control device for controlling the speed of an elevator hoisting machine according to a speed command, wherein the car has a vibration sensor. Then, a torque command for the hoist is generated and output, and from the combination of the speed information and the torque command obtained when the torque command is output, the torque command that minimizes the amount of vertical vibration obtained from the vibration sensor is selected. , a first control unit that performs nonlinear control determined by reinforcement learning, a second control unit that generates and outputs a torque command for the hoist by linear control, and a monitoring process that monitors the torque command or the speed of the hoist and the monitoring processing unit, either one of the torque command output by the first control unit and the torque command output by the second control unit is selected based on the monitoring by the unit, and the torque command is output to the drive circuit of the hoist. and a selection unit for supplying the selected speed of the elevator hoist, the selection unit selecting either nonlinear control or linear control.

According to the present invention, when the torque command value becomes an abnormal value for some reason while the hoisting machine is being controlled by nonlinear control, the control is switched to linear control, and the control state of the hoisting machine becomes unstable. can be effectively prevented from becoming
Problems, configurations, and effects other than those described above will be clarified by the following description of the embodiments.

1 is a block diagram showing a configuration example of an elevator control device according to an embodiment of the present invention; FIG. FIG. 4 is a diagram showing an example of information and instructions received by a first control unit in learning processing according to an embodiment of the present invention; 1 is a block diagram showing a hardware configuration example of an elevator control device according to an embodiment of the present invention; FIG. 4 is a flow chart showing the overall flow of control processing according to an embodiment of the present invention; 4 is a flow chart showing the flow of control processing (an example of varying according to stroke) according to an embodiment of the present invention; 4 is a graph showing an example of temporal change in torque command according to an embodiment of the present invention; FIG. 11 is a block diagram showing a configuration example of an elevator control device according to another embodiment of the present invention;

An embodiment of the present invention (hereinafter referred to as "this example") will now be described with reference to FIGS. 1 to 6. FIG.
[Configuration of Elevator Control Device]
FIG. 1 shows the configuration of the elevator control device 10 of this example.
First, before explaining the configuration of the elevator control device 10, the configuration of the elevator mechanism section 20, which is the configuration for running the car 21, will be briefly described.

The car 21 is connected to one end of a main rope 23 wound around a mesh wheel 24, and a weight 22 is connected to the other end of the main rope 23. - 特許庁The net wheel 24 is driven by the hoist 32 to rotate, and the car 21 connected to the main rope 23 runs due to the rotation.
A vibration sensor 25 for detecting vertical vibration of the car 21 is attached to the car 21, and a detection signal of the vertical vibration obtained by the vibration sensor 25 is supplied to the elevator control device 10 side.

The hoisting machine 32 is composed of a three-phase AC motor or the like, and is operated by a driving power supply from a driving circuit 31 composed of an inverter or the like. Generation of drive power by the drive circuit 31 is controlled by a first control section (nonlinear control section) 12 or a second control section (PI control section) 15, which constitutes the elevator control device 10 and will be described later.
In this case, a rotary encoder 33 is attached to the hoisting machine 32 and the rotary encoder 33 detects the rotational speed of the hoisting machine 32 . A rotation speed detection signal detected by the rotary encoder 33 is supplied to the first control unit 12 and the second control unit 15 .

Next, the configuration of the elevator control device 10 that controls driving of the hoisting machine 32 will be described.
The elevator control device 10 includes a first control section (nonlinear control section) 12 , a second control section (PI control section) 15 and a monitoring processing section 13 .
The elevator control device 10 also includes a selection unit 16 , a range setting unit 17 and a database unit 18 .

A speed command for the hoist 32 obtained at the speed command input terminal 11 is supplied to the first control unit 12 . Also, the speed command for the hoist 32 obtained at the speed command input terminal 11 is supplied to the second control section 15 via the subtractor 14 .
The first control unit 12 and the second control unit 15 generate torque commands for the hoisting machine 32 based on the supplied speed commands.

The first control unit 12 determines a torque command through nonlinear control processing based on reinforcement learning to which AI (artificial intelligence) technology is applied. In addition to the speed command obtained at the speed command input terminal 11, the first control unit 12 also has information on the rotation speed detected by the rotary encoder 33 and vibration detected by the vibration sensor 25 attached to the car 21. A detection signal is provided. A configuration in which the first control unit 12 determines the output by reinforcement learning will be described later with reference to FIG. In addition, although the present embodiment shows a case where AI technology is applied as nonlinear control processing, as another example, if control stability is not theoretically guaranteed, such as model predictive control, Either can be applied.

The second control unit 15 determines a torque command by PI control (Proportional-Integral control), which is linear control processing combining proportional action and integral action. The speed command supplied to the second control unit 15 is a differential speed command obtained by subtracting the rotation speed detected by the rotary encoder 33 from the speed command obtained at the speed command input terminal 11 by the subtractor 14 . The second control unit 15 generates a torque command for the hoisting machine 32 from the differential speed command. The linear control performed by the second control unit 15 has conventionally been generally performed as control of an elevator hoist.

The torque command obtained by the first control unit 12 and the torque command obtained by the second control unit 15 are supplied to the selection unit 16, and one of the torque commands is selected and supplied to the drive circuit 31. be. The drive circuit 31 generates drive power corresponding to the supplied command and supplies it to the hoist 32 . As a result, the hoist 32 operates with torque corresponding to the torque command output by the selector 16 .

Selection of the torque command by the selection unit 16 is controlled by the monitoring processing unit 13 .
The monitoring processing unit 13 performs processing for monitoring the torque command output by the first control unit 12 . Then, when the torque command monitored by the monitoring processing unit 13 is appropriate, the selection unit 16 supplies the torque command output by the first control unit 12 to the drive circuit 31 . Further, when the torque command monitored by the monitoring processing unit 13 is not appropriate, the selection unit 16 supplies the torque command output by the second control unit 15 to the drive circuit 31 .

When monitoring the torque command output by the first control unit 12 , the monitoring processing unit 13 monitors whether the torque command is within the range between the upper limit value and the lower limit value set by the range setting unit 17 . When the range setting unit 17 sets the range of the torque command, the range setting unit 17 refers to the travel control information obtained from the travel control information input terminal 19 and the past range setting information stored in the database unit 18. to set.
That is, the range setting unit 17 acquires the travel control information obtained from the travel control information input terminal 19 via the monitoring processing unit 13 . The travel control information obtained at the travel control information input terminal 19 includes travel mode information of the car 21, information on the distance from the departure floor to the stop floor, and the like. The running mode information indicates whether the vehicle is accelerating, running at a constant speed, decelerating, or the like. In the database unit 18, for example, the range of the upper limit value and the lower limit value set based on the torque command that can be obtained during normal running of the elevator is registered in advance.

[Configuration in which the first control unit performs reinforcement learning]
FIG. 2 shows an example of a configuration in which the first control unit 12 performs reinforcement learning.
In the case of this example, the first control unit 12 that performs reinforcement learning serves as an agent, and the first control unit 12 sends a torque command to the elevator mechanism unit 20 that is the control target. Then, the vibration sensor 25 of the elevator mechanism section 20 provides the first control section 12 with the amount of vertical vibration as a reward.
Further, measurement information from the elevator mechanism section 20 is supplied to the first control section 12 . The measurement information includes speed information of the hoist 32 detected by the rotary encoder 33, for example.

The first control unit 12 repeatedly learns the torque command that minimizes the amount of vertical vibration of the car 21 as a reward from the combination of the measurement information obtained from the elevator mechanism unit 20 and the torque command. to decide.

[Hardware configuration example of elevator control device]
FIG. 3 shows a hardware configuration example when the elevator control device 10 is configured by a computer device 100 .
A computer device 100 shown in FIG. 3 includes a CPU (Central Processing Unit) 101, a ROM (Read Only Memory) 102, and a RAM (Random Access Memory) 103, which are respectively connected to a bus. Further, computer device 100 includes nonvolatile storage 104 , network interface 105 , and input/output unit 106 .

The CPU 101 is an arithmetic processing unit that reads out from the ROM 102 and executes a software program code for executing processing for generating a torque command.
The RAM 103 is temporarily written with variables, parameters, and the like generated during arithmetic processing.

A large-capacity information storage unit such as a HDD (Hard Disk Drive) or an SSD (Solid State Drive) is used for the nonvolatile storage 104, for example. The nonvolatile storage 104 stores data such as control history. Information stored in the database unit 18 is also stored in the nonvolatile storage 104 . Furthermore, the non-volatile storage 104 may store part or all of the software that executes the process of generating the torque command.

For the network interface 105, for example, a NIC (Network Interface Card) or the like is used, and performs transmission/reception processing of various information with an external device.
The input/output unit 106 performs processing for outputting a torque command to the drive circuit 31 and also performs input processing for outputs of the rotary encoder 33 and the vibration sensor 25 .

It should be noted that the configuration of the elevator control device 10 with the computer shown in FIG. 3 is an example, and may be configured with a device that performs arithmetic processing other than the computer. For example, some or all of the functions performed by the elevator control device 10 may be realized by hardware such as FPGA (Field Programmable Gate Array) or ASIC (Application Specific Integrated Circuit).
Further, although the elevator control device 10 is one computer here, for example, the first control unit 12 and the second control unit 15 may be configured by separate computers (or hardware such as FPGA and ASIC).

[Flow of control processing]
FIG. 4 is a flow chart showing the flow of processing when the elevator control device 10 of this example generates a torque command.
When the car 21 starts running by the hoisting machine 32, the selection unit 16 selects the first control unit 12, performs processing for generating a torque command by control based on reinforcement learning in the first control unit 12, and generates a torque command. is supplied to the drive circuit 31 to drive the hoist 32 (step S11).
Then, the monitoring processing unit 13 executes monitoring processing of the torque command generated in step S11 (step S12). Here, the monitoring processing unit 13 determines whether or not the torque being monitored falls within the range between the upper limit value and the lower limit value set by the range setting unit 17 (step S13).

In step S13, if the monitored torque is within the set range (Yes in step S13), the process returns to step S11 to continue generation of the torque command in the first control unit 12. do.
On the other hand, in step S13, when the monitored torque exceeds the set range (No in step S13), the selector 16 changes the output torque command from the output of the first controller 12 to the second control Switching to the output of the unit 15, the control by the second control unit 15 is executed (step S14). By performing this switching, the torque command supplied to the drive circuit 31 is switched from a command based on nonlinear control by reinforcement learning to a command based on linear control by PI control.

Then, the selector 16 determines whether or not the car 21 has stopped based on the speed command or the like obtained from the speed command input terminal 11 (step S15). If the car 21 is not stopped in step S15 (No in step S15), the control by the second control unit 15 in step S14 is continued.
Then, in step S15, when it is determined that the car 21 has stopped (Yes in step S15), the selection unit 16 returns the first control unit 12 to the selected state, returns to the process of step S11, and returns to the first control unit 12. A torque command is generated by the control unit 12 .

FIG. 5 is a flowchart showing range setting processing in the range setting unit 17 when the monitoring processing unit 13 monitors the torque command.
First, the range setting unit 17 acquires distance information from the departure floor to the stop floor of the car 21 from the travel control information obtained from the travel control information input terminal 19 (step S21). Based on this distance information, the range setting unit 17 determines the travel speed (maximum speed) at which the car 21 travels. For example, when the distance (number of floors) from the departure floor to the stop floor is long, the range setting unit 17 determines that high-speed travel is to be performed. Decide to run.

Further, the range setting unit 17 determines the traveling state in the accelerating traveling mode, the traveling state in the constant speed traveling mode, and the traveling state in the decelerating traveling mode (step S22).
Then, the range setting unit 17 sets a range determined by the upper limit value and the lower limit value of the torque command value in each travel mode suitable for the distance and speed determined in step S21 (step S23). When setting the range of the torque command value, reference is made to the range set under similar circumstances in the past, which is stored in the database unit 18, for example.

[Example of relationship between torque command, upper limit, and lower limit during running]
FIG. 6 shows an example of the relationship between the torque command for the hoist and the upper and lower limits when the car is running. The vertical axis in FIG. 6 is the torque command value, and the horizontal axis is time. The torque command T1 shown in FIG. 6 is a torque command output by the first control unit 12, which is nonlinear control to which reinforcement learning is applied.
As shown in FIG. 6, there is an acceleration section ta in which the car accelerates first, then a constant speed section tb, and finally a deceleration section tc, where the car stops. The length (time) of the constant speed section tb changes depending on the distance traveled by the car, and the length of the acceleration section ta and the deceleration section tc also change depending on the speed during constant speed operation.

A line TH shown in FIG. 6 indicates the upper limit of values that can normally be taken as the torque command. A line TL indicates the lower limit of values that can be taken as the torque command.
These upper limit value TH and lower limit value TL are set by the range setting unit 17 by referring to the information of the running mode and the past data accumulated in the database unit 18, etc., and setting values considered appropriate for the current running. . In other words, when it is within the range of the upper limit and the lower limit, the elevator is in a permissible state for running. It is a state in which there is a possibility that a landing error may occur, or that a significant overshoot or stop short of the destination floor may occur.
As shown in FIG. 6, the upper limit TH and the lower limit TL are values that change depending on the acceleration section ta, the constant speed section tb, and the deceleration section tc. The acceleration section ta and the deceleration section tc also change within the section. .

The monitoring processing unit 13 monitors whether or not the torque command T1 output by the first control unit 12 is within the range indicated by the upper limit value TH and the lower limit value TL.
The x mark in FIG. 6 indicates the case where the elevator deviates from the assumed normal running, and indicates that the torque command T1 becomes lower than the lower limit value TL in the middle of the constant speed section tb.
When such a state occurs, in the case of this example, the monitoring processing unit 13 executes processing for switching the selection unit 16 to the second control unit 15 side. As a result, the hoist 32 is driven by the torque command output by the second control unit 15, which is linear control whose control stability is theoretically guaranteed.

As described above, according to the elevator control device 10 of the present embodiment, high-precision running control with little vibration of the car 21 is performed by the torque command output by the first control unit 12, which normally performs non-linear control processing. possible. In the event of a non-linear control malfunction that cannot be ruled out, even though it is rare, the system switches to the second control unit 15 where linear control processing is performed so that passengers can arrive at their destination floors without experiencing unpleasant vibrations or the like. become. After the car 21 stops, the control is switched to the control of the first control unit 12 again. Therefore, if the cause of the erroneous torque command is based on a temporary erroneous calculation result, then the original high-precision traveling is resumed. Control can be returned.

[Another configuration example of the elevator control device]
In the configuration shown in FIG. 1, the monitoring processing unit 13 monitors whether or not the torque command output by the first control unit 12 is within a set range. Regarding this, the monitoring processing unit 13 may monitor whether the rotation state of the hoist 32 is within a set range.
FIG. 7 shows a configuration for monitoring whether the rotation state of the hoisting machine 32 is within a set range. In FIG. 7, the same reference numerals are given to the same portions as the configuration explained in FIG.

A monitoring processing unit 13 ′ shown in FIG. 7 acquires information on the rotation speed of the hoist 32 detected by the rotary encoder 33 . Then, it monitors whether or not the acquired rotation speed of the hoist 32 is within the set range. As the upper limit value and the lower limit value of the rotation speed to be monitored by the monitoring processing unit 13', the upper limit value and the lower limit value set by the range setting unit 17 are used as in the example of FIG.
Other configurations in FIG. 7 are configured in the same manner as the elevator control device 10 shown in FIG.

In this way, the monitoring processing unit 13′ acquires the rotation speed information as a result of operating the hoisting machine 32 with the torque command output by the first control unit 12, and monitors whether it is within the set range. By doing so, it is possible to monitor that there is an abnormality in the torque command output by the first control unit 12 . Therefore, in the case of the configuration shown in FIG. 7, the rotation speed information exceeds the range when there is an abnormality, and the same effect as the above-described embodiment can be obtained.

[Other Modifications]
The present invention is not limited to each embodiment described above, and includes various modifications. For example, when the range setting unit 17 sets the range of torque values (or rotational speed), not only the travel distance (stroke) of the car 21 but also information on the load is acquired, and A suitable range may be calculated. In addition, the range setting unit 17 is provided with information such as the inertia of each mechanism constituting the elevator mechanism unit 20, and based on the configuration of the elevator mechanism unit 20, a more accurate torque value or upper limit value of rotation speed and A lower limit value may be calculated.

Further, the information about the driving mode may be determined by dividing the acceleration section into more sections (driving modes) such as an acceleration first half section and an acceleration second half section.
Further, the information of the upper limit value and the lower limit value stored in the database unit 18 may be sequentially updated based on the upper limit value and the lower limit value obtained by calculation by the range setting unit 17 in actual operation. Conversely, if the range setting unit 17 always acquires information such as the stroke and the load and calculates the upper limit value and the lower limit value, the database unit 18 may be omitted.

Further, the above-described embodiment has been described in detail for easy understanding of the present invention, and is not necessarily limited to those having all the described configurations. In addition, in the configuration diagrams shown in FIGS. 1 and 7, only the control lines and information lines that are considered necessary for explanation are shown, and not all the control lines and information lines are necessarily shown on the product. . In practice, it may be considered that almost all configurations are interconnected. In addition, in the flowcharts shown in FIGS. 4 and 5, the execution order of some processing steps may be changed or some processing steps may be executed at the same time as long as the processing results of the embodiment are not affected. can be

Further, each of the above configurations, functions, processing units, processing means, and the like may be realized by hardware, for example, by designing a part or all of them using an integrated circuit. Moreover, each of the above configurations, functions, etc. may be realized by software by a processor interpreting and executing a program for realizing each function. Information such as programs, tables, and files that implement each function can be placed in a recording device such as a memory, a hard disk, or an SSD, or a recording medium such as an IC card, an SD card, or an optical disc.

DESCRIPTION OF SYMBOLS 10... Elevator control apparatus 11... Speed command input terminal 12... First control part 13, 13'... Monitoring process part 14... Subtractor 15... Second control part 16... Selection part 17... Range setting Section 18 Database section 19 Running control information input terminal 20 Elevator mechanism section 21 Car 22 Weight 23 Main rope 24 Mesh wheel 25 Vibration sensor 31 Drive circuit 32... Winding machine, 33... Rotary encoder, 100... Computer device, 101... Central processing unit (CPU), 102... ROM, 103... RAM, 104... Non-volatile storage, 105... Network interface, 106... Input/output unit

Claims (10)

In an elevator control device that controls the speed of an elevator hoist according to a speed command,
the car has a vibration sensor,
A torque command for the hoisting machine is generated and output, and a torque command that minimizes the amount of vertical vibration obtained from the vibration sensor is determined from a combination of the torque command and the speed information obtained when the torque command is output. , a first control unit that performs nonlinear control determined by reinforcement learning ;
a second control unit that generates and outputs a torque command for the hoist by linear control;
a monitoring processing unit that monitors the torque command or the speed of the hoist;
one of the torque command output by the first control unit and the torque command output by the second control unit is selected based on the monitoring by the monitoring processing unit, and a drive circuit for the hoisting machine is selected; a selection unit for supplying to
The elevator control device, wherein the selector selects either the nonlinear control or the linear control for speed control of the hoisting machine of the elevator. After the selection unit selects the torque command from the first control unit, the monitoring processing unit determines that the torque command output by the first control unit is based on the torque command that can be taken during normal running of the elevator. Judging whether it is within the range of the set upper limit value and lower limit value,
The elevator control device according to claim 1, wherein when the torque command exceeds the range of the upper limit value or the lower limit value, the selection unit performs processing to switch from the torque command of the first control unit to the torque command of the second control unit. . 2. The elevator control device according to claim 1, wherein when the car that has traveled due to the drive of the hoisting machine stops, the selection unit switches to the torque command output by the first control unit. The elevator control device according to claim 2, wherein the range of the upper limit value and the lower limit value is set according to a running mode of the car that runs by the hoisting machine. The elevator control device according to claim 4, wherein the running modes include at least a running mode during acceleration, a running mode at constant speed, and a running mode during deceleration. 3. The elevator control device according to claim 2, wherein the range of the upper limit value and the lower limit value set based on the torque command that can be taken during normal running of the elevator is registered in a database in advance. The monitoring processing unit acquires process information for traveling the car, and variably sets the upper limit value and the lower limit value according to the distance from the departure floor to the stop floor obtained from the process information. Item 3. The elevator control device according to Item 2. The monitoring processing unit acquires a detection signal of the rotation speed of the hoisting machine, and when the rotation speed of the hoisting machine exceeds a set range, the selection unit causes the first control unit to 2. The elevator control device according to claim 1, wherein a process of switching to the torque command of the second control section is performed. The first control unit is supplied with the output of a vibration sensor installed in the car, performs nonlinear control by learning processing based on the output of the vibration sensor, and generates a torque command for the hoist. Item 9. The elevator control device according to any one of items 1 to 8. In an elevator control method for controlling the speed of an elevator hoist according to a speed command,
A torque command for the hoisting machine is generated and output, and from the combination of the speed information and the torque command obtained when the torque command is output, the vertical vibration amount obtained from the vibration sensor installed in the car is the largest. Non-linear control processing that determines the torque command to be reduced by reinforcement learning ;
a linear control process for generating and outputting a torque command for the hoist by linear control;
a monitoring process for monitoring the torque command or the speed of the hoist;
Based on the monitoring by the monitoring process, either one of the torque command output by the non-linear control process and the torque command output by the linear control process is selected and supplied to the drive circuit of the hoisting machine. a selection process;
An elevator control method, wherein the speed control of the hoisting machine of the elevator can be performed by both the nonlinear control and the linear control.

Images