JP6730216B2 - Elevator management system and elevator management method - Google Patents

Description

The present invention relates to an elevator management system and an elevator management method, for example, to an elevator management system that manages a plurality of cars operating between a plurality of floors as a group.

Conventionally, this type of elevator management system reciprocates the cars so that the intervals between the cars in the direction of gravity are even between the bottom floor and the top floor. However, for some cars, if a delay occurs due to the getting on and off of many passengers, it is not possible to operate multiple cars evenly, and for example, multiple cars on the same floor stop at the same time. As a result, there was a situation where the waiting time was long on other floors.

Therefore, Patent Document 1 proposes an invention that controls the elevator management system so that the waiting times of the cars on each floor are equalized in order to improve the operating efficiency of the cars. This invention sets the position and the moving direction of the car after a predetermined time, while assuming that each of the plurality of cars repeats the operation between the lowermost floor and the uppermost floor, so as to match this. First, the car is operated.

Japanese Patent No. 4139819

The conventional elevator management system has a problem that waste is generated in the operation of the car. Therefore, the present invention is intended to propose an elevator management system and an elevator management method for efficiently operating a car.

In order to solve such a problem, in the present invention, in an elevator management system that manages an elevator including a control device that operates a car over a plurality of floors, a management device that manages the control device is provided, and the management device is , A reception circuit for receiving the destination floor designation information and car call information, a memory for accumulating and recording the reception information of the reception circuit, and learning the operation tendency of the car based on the information recorded in the memory A controller and an output circuit that outputs management information to the control device are provided, and the controller predicts destination floor designation information and car call information after a predetermined time from the reception information of the reception circuit, based on the learning result. Then, based on the result of the prediction after the predetermined time, the management information is formed so as to limit the range of the operation floor of the car, and the control device controls the operation of the car based on the management information. did.

Further, in the present invention, in an elevator management method of managing a control device for operating an elevator car over a plurality of floors by a management device, the management device receives destination floor designation information and car call information. Then, the received information of the receiving circuit is accumulated and recorded, the operation tendency of the car is learned based on the received information, the management information is output to the control device as the learning result, and the received information is output from the received information based on the learning result. The destination floor designation information after a predetermined time and the car call information are predicted, and based on the result of the prediction, the management information is formed so as to limit the range of the operating floor of the car, and the management information is displayed in the control device. The operation of the car is controlled based on.

According to the present invention, it is possible to realize an elevator management system and an elevator management method for efficiently operating a car.

It is a block diagram showing composition of an elevator management system by this embodiment. It is an approximate line figure showing the important section of the schematic structure of the elevator device by this embodiment. It is a figure which shows the operation route of the elevator apparatus by this Embodiment. It is a conceptual diagram with which explanation of learning by this Embodiment is offered. It is a conceptual diagram which shows the structure of the operation data table by this Embodiment. It is a conceptual diagram which shows the structure of the learning data table by this Embodiment. It is a flow chart which shows a processing procedure of operation data storage processing. It is a flow chart which shows a processing procedure of operation data learning processing. It is a flow chart which shows a processing procedure of operation route decision processing. It is a flow chart which shows a processing procedure of operation instruction processing. It is a flow chart which shows a processing procedure of operation route amendment processing during operation. It is a flow chart which shows a processing procedure of operation route amendment processing at the time of a door opening. It is a block diagram showing composition of an elevator management system by other embodiments. It is a block diagram showing composition of an elevator management system by other embodiments. It is a block diagram showing composition of an elevator management system by other embodiments. It is a block diagram showing composition of an elevator management system by other embodiments.

(1) Configuration of Elevator Management System According to this Embodiment In FIG. 1, reference numeral 1 denotes an elevator management system according to this embodiment. The elevator management system 1 includes a management server 2 that manages a plurality of elevators 3. The management server 2 and the plurality of elevators 3 are connected via a communication path 19 such as an intranet.

The management server 2 acquires the operation data of each elevator 3 via the receiving circuit, learns the operation status of each elevator 3 from the acquired operation data, and outputs the operation of each elevator 3 to the output circuit. It is a management device that manages by outputting management information via the management information. The management server 2 includes a CPU (Central Processing Unit) 4, an auxiliary storage device 5, and a memory 6.

The CPU 4 is a processor (controller) that controls the operation of the entire management server 2. The auxiliary storage device 5 is composed of, for example, a large-capacity nonvolatile storage device such as a hard disk device or an SSD (Solid State Drive), and is used to store programs and data for a long period of time. A part of the storage area provided by the auxiliary storage device 5 is used as an operation data table TB10 and a learning data table TB20 described later.

The memory 6 is composed of, for example, a volatile semiconductor memory, is also used as a work memory of the CPU 4, and includes an operation storage module 7, an operation learning module 8, a route determination module 9, and a route instruction module 10. The memory 6 may appropriately accumulate and record the operation data.

As shown in FIG. 2, each elevator 3 operates so as to raise and lower the car 12 in the hoistway installed in the building, for example, between the halls provided on the first floor to the seventh floor. ing.

The car 12 is attached to the other end of the main rope 13 having a counterweight 14 attached to one end. The main rope 13 is wound around the hoisting machine 15. The hoisting machine 15 is an elevating mechanism that drives the car 12 to elevate and lower, and a control device that controls the elevating operation of the car 12 (hereinafter, referred to as an elevator control device) in a machine room provided above the hoistway. It is installed with 11).

The elevator control device 11 (FIG. 1) is a computer device that controls the operation of the car 12, and the car 12 is operated according to the passenger's operation (car call information) on the call button 16 (FIG. 2) provided in the hall. The hoisting machine 15 is controlled so as to move up and down.

Elevator 3 is inefficient because it is difficult to determine which route at which time car 12 does not have to travel, but usually car 12 goes from the top floor to the bottom floor. It is operated as something that can be operated. In the present invention, a learning function is installed in the elevator management system 1 to make this determination.

Next, the learning function implemented in the management server 2 of the elevator management system 1 will be described. Note that the learning function implements, for example, deep learning.

The learning function of the elevator management system 1 is such that the car call information of each floor and/or the operation (destination floor designation information) of the destination floor designation button of each car 12 is received, and then a predetermined time (for example, After 5 minutes, which is a cycle in which the car 12 makes one round trip on the traveling path, the operation state of the call button 16 and the destination floor designation button of each car 12 is predicted to learn the operation tendency.

FIG. 3 shows an example of the learning function. The neurons (circles in FIG. 3) in adjacent layers (columns in FIG. 3) are weighted and the calculation is performed. In this learning function, when the dimensional array corresponding to the number of inputs of the call button 16 and the destination floor designating button of each car 12 in each layer is input from the input layer, the management server 2 in a plurality of hidden layers and output layers. Performs a predetermined calculation. It should be noted that the input layer is one layer in which input is performed, and the output layer is one layer in which output is performed. There are multiple hidden layers between the input layer and the output layer.

Then, as a result of the calculation, the management server 2 outputs the dimensional array corresponding to the number of inputs of the call button 16 and the destination floor designating button of each car 12 in each hierarchy. The predetermined operation in the hidden layer is, for example, an operation using an activation function such as a sigmoid function, a hyperbolic tangent function, or a ramp function. The predetermined operation in the output layer is, for example, an operation using a softmax function or the like.

In FIG. 3, when the management server 2 is called the car 12 that rises on the first floor, and the car 12 that rises and the car 12 that descends on the second floor are called based on this learning function. , Predicts how the car 12 will be called in the next cycle.

In this case, the management server 2 predicts that in the next cycle, the ascending car 12 on the second floor will be called and the descending car 12 on the fourth floor will be called.

In FIG. 3, the case where the button is pressed is indicated by “◯”, and the case where the button is not pressed is indicated by “x”. Also, in each floor, the button for calling the car 12 that rises in the hall is "↑", the button for calling the car 12 that descends in the hall is "↓", and the button for getting off at that floor in the car 12 Is represented by "→". In addition, in FIG. 3, since there is only one car 12 as an example, there is one row of “→” on each floor, but there may be a plurality of rows.

In the case of the prediction of FIG. 3, the deep learning learns that the management server 2 predicts that the car 12 is not called on the fourth to seventh floors during the time required to make one round trip. The management server 2 instructs the elevator control device 11 to operate the car 12 on the operation route having the fourth floor as the destination floor as shown by the solid line in FIG. Specifically, the management server 2 instructs the elevator control device 11 to reverse the traveling direction of the car 12 after the waiting time for getting on and off on the fourth floor. In addition, the broken line in FIG. 4 has shown the conventional operation route, and in the conventional operation, the car 12 reaches the 7th floor which is the top floor.

As a means for realizing the above learning function, as shown in FIG. 1, the memory 6 of the management server 2 includes an operation storage module 7, an operation learning module 8, a route determination module 9, and a route instruction module 10. The operation data table TB10 and the learning data table TB20 are stored in the auxiliary storage device 5 of the management server 2.

The operation storage module 7 is a program having a function of acquiring operation data from the elevator control device 11 of each elevator 3, for example, every day, and storing the acquired operation data in the operation data table TB10.

The operation learning module 8 performs the learning as shown in FIG. 3 based on the operation data acquired from the operation data table TB10, for example, every year for one year, and learns the weighting value as shown in FIG. This is a program that changes the values required for. Further, the operation learning module 8 records in the learning data table TB20 the state of the call button 16 and the state of the destination floor designating button of each car 12 (learning result) calculated as the learning result. The learning result is calculated for each combination of the state of the call button 16 and the state of the destination floor designating button of each car 12 in each arbitrary hierarchy.

The route determination module 9 is a program that acquires a learning result from the learning data table TB20 and determines an operation route of the car 12. The route determination module 9 derives a route that can be omitted without traveling the car 12 from each learning result, and determines a route that does not pass through the route as a shortened route as an operation route of the car 12.

For example, in the case of the learning result such as the input line and the output line in FIG. 6, it is understood that the route determination module 9 does not need to pass the second floor to the seventh floor in the input line and the output line.

Therefore, the route determination module 9 determines the shortened route that does not pass through the second floor to the seventh floor as the operation route of the car 12. If there is no route that can be omitted, the route determination module 9 determines the normal route that goes back and forth from the bottom floor to the top floor as the operation route of the car 12.

The route instruction module 10 is a program that corrects the operation route determined by the route determination module 9 according to the position of each car, the traveling direction, etc. given from the elevator control device 11 of each elevator 3. For example, while the car 12 is in operation, it still occurs when the call button 16 is pressed or the door of the car 12 is open in the operation route where the car 12 has not yet passed. When there is an unpredictable situation, the operation route of the car 12 is corrected, and the corrected operation route is transmitted to the elevator control device 11 of the elevator 3 as management information.

In addition, when it is not necessary to correct the operation route of the car 12, the route instruction module 10 transmits the operation route determined by the route determination module 9 to the elevator control device 11 of the elevator 3 as it is. Further, the route instruction module 10 determines one or more cars 12 to be operated from the plurality of cars 12 based on the operation route determined by the route determination module 9.

In the operation data table TB10, as shown in FIG. 5, the state of the call button 16 (car call information) and the state of the destination floor designating button of each car 12 (destination floor designating information) in each floor are stored as the operating data. It is stored every 5 minutes (time required for one round trip). Note that “◯”, “×”, “↑”, “↓”, and “→” in FIG. 5 have the same meanings as in FIG.

Similarly, as shown in FIG. 6, 5 minutes (1 minute) when the state of the call button 16 and the state of the destination floor designating button of each car 12 are input to the learning data table TB20 as shown in FIG. The state of the call button 16 and the state of the destination floor designating button of each car 12 in each layer (the time required to make a round trip) are stored as outputs. Note that “◯”, “×”, “↑”, “↓”, and “→” in FIG. 6 have the same meanings as in FIG.

(2) Various processes by the management server Next, various processes executed by the management server 2 described above will be described. Note that, in the following, the processing subject of various processes is described as a “program”, but it goes without saying that the CPU 4 actually executes the process based on the “program”.

FIG. 7 shows a processing procedure of the operation data acquisition processing executed by the operation storage module 7. The operation storage module 7 acquires operation data from the elevator control device 11 of each elevator 3 according to the processing procedure shown in FIG.

In practice, the operation storage module 7 starts the operation data acquisition process shown in FIG. 7 at a fixed time every day, for example.

Then, the operation storage module 7 first acquires the operation data for one day from the elevator control device 11 of each elevator 3 (S11). Subsequently, the operation storage module 7 stores the operation data for one day in the operation data table TB10 (S12), and ends the operation data acquisition process.

FIG. 8 shows a processing procedure of operation data learning processing executed by the operation learning module 8. The operation learning module 8 follows the processing procedure shown in FIG. 8 and, based on the operation data acquired from the operation data table TB10, the state of the call button 16 and the state of the destination floor designation button of each car 12 in any given floor. The state of the call button 16 and the state (learning result) of the destination floor designating button of each car 12 in each floor of 5 minutes (the time required to make one round trip) are learned.

In practice, the operation learning module 8 starts the operation data learning processing shown in FIG. 8 at a time determined every year, for example.

Then, the operation learning module 8 first acquires operation data for one year from the operation data table TB10, and learns based on the acquired operation data (S15). Then, the operation learning module 8 stores the learning result as learning data in the learning data table TB20 (S16), and ends the operation data learning process.

FIG. 9 shows a processing procedure of an operation route determination process executed by the route determination module 9. The route determination module 9 determines the operation route of the car 12 according to the processing procedure shown in FIG.

In practice, the route determination module 9 starts the operation route determination process shown in FIG. 9 when the operation data learning process ends.

Then, the route determination module 9 first acquires the learning data from the learning data table TB20 (S21). Subsequently, the route determination module 9 determines whether or not there is a route that can be omitted for each learning result (S22). When the route determination module 9 obtains a negative result in this determination, it transmits to the route instruction module 10 that the normal route is to be the operation route, and ends the operation data learning process.

On the other hand, when there is a route that can be omitted, the route determination module 9 generates a shortened route that omits the route when a positive result is obtained in the determination in step S22 (S23), and transmits it to the route instruction module 10. Then, the operation data learning process ends.

FIG. 10 shows a processing procedure of the operation instruction processing executed by the route instruction module 10. The route instruction module 10 instructs the car 12 about the operation route according to the processing procedure shown in FIG.

Actually, when the route instruction module 10 receives the passenger's operation on the call button 16 from the elevator control device 11 of the elevator 3, the route instruction module 10 starts the traveling route correction process shown in FIG.

Then, the route instruction module 10 first determines the car 12 to be operated (S25). Subsequently, the route instruction module 10 transmits the operation route to the elevator control device 11 that controls the car 12 (S26), and ends the operation instruction process. Then, the elevator control device 11, which has received the operation route, operates the car 12 according to this operation route.

FIG. 11 shows a processing procedure of a traveling route correction process performed by the route instruction module 10. The route instruction module 10 corrects the operation route of the car 12 according to the processing procedure shown in FIG.

In reality, after the operation instruction processing is completed, the route instruction module 10 controls the operation of the passenger on the call button 16 at the hall in the operation route of the car 12 instructed to be operated by the operation instruction processing, by the elevator control device 11 When received from, the operation route correction process during operation shown in FIG. 11 is started.

Then, the route instruction module 10 first acquires the position and traveling direction of the car 12 from each elevator control device 11 and determines whether or not there is a car 12 approaching the hall (S31). The route instruction module 10 has a car 12 in the vicinity of the hall, and if a positive result is obtained in this determination, the route correction process during operation ends. Since the car 12 in the vicinity stops at the landing, the control by the management server 2 becomes unnecessary.

On the other hand, the route instruction module 10 selects the car 12 closest to the landing when a positive result is obtained in the determination in step S31 because there is no car 12 in the vicinity (S32). Subsequently, the route instruction module 10 transmits an instruction to reverse the traveling direction of the selected car 12 to the elevator control device 11 that controls the selected car 12 (S33), and ends the operation route correction process during operation. ..

FIG. 12 shows a processing procedure of door opening operation route correction processing executed by the route instruction module 10. The route instruction module 10 corrects the operation route of the car 12 according to the processing procedure shown in FIG.

In practice, the route instruction module 10 opens the door of the car 12 whose operation is instructed by the operation instruction processing after the operation instruction processing is finished (hereinafter, this is referred to as opening of the door). When a push operation of a button that has not yet occurred (which should have occurred according to the preliminary result in the learning result) for the car 12 is received from the elevator control device 11, the door-open operation shown in FIG. 12 is performed. Start route correction processing.

Then, the route instruction module 10 first determines whether or not the elapsed time from the time when the button pressing operation should have occurred to the time when the door is opened is a predetermined value or less (S41). The route instruction module 10 determines that the error on the prediction side in the learning result cannot be corrected because the predetermined time has elapsed, and if a negative result is obtained in this determination, the route-opening operation route correction is performed. The process ends.

On the other hand, the route instruction module 10 obtains a positive result in the determination of step S41 by determining that the predetermined error has not elapsed and the error on the preliminary side in the learning result can be corrected. Then, an instruction to extend the time when the door is opened is transmitted to the elevator control device 11 that controls the car 12 (S42), and the door open operation route correction process is ended.

(3) Effects of this Embodiment As described above, in the elevator management system 1 of this embodiment, the management server 2 predicts from the learning data what operation each elevator 3 will perform in the next cycle. By doing so, an instruction is given to each elevator 3 to omit the route that can be omitted and to operate.

Therefore, according to the elevator management system 1, it becomes possible to apply the operation according to the usage state to the car 12 without modifying the program, and the car 12 can be operated efficiently.

(4) Other Embodiments In the above-described embodiments, the case where the elevator management system 1 to which the present invention is applied is configured as shown in FIG. 1 has been described, but the present invention is not limited to this, and these elevators are not limited thereto. Various other configurations can be widely applied as the configuration of the management system.

For example, as shown in FIG. 13, the elevator management system 20 may be configured to connect the management server 2 and each elevator 3 via a communication network 21 such as the Internet. In this case, the management server 2 is a cloud server or a server device installed in a data center. The management server 2 is connected to the elevator 3 via a communication network 21, communication devices 22 and 23 such as switching hubs and routers, and a communication path 24 such as an intranet.

In the case of the configuration as shown in FIG. 1, the installation location of the management server 2 is assumed to be in the hoistway of the elevator 3 or the machine room, but such a location has a large capacity for storing operation data for one year. It may be difficult to install such a data storage device or a server device for performing deep learning. However, by configuring the elevator management system 20 as shown in FIG. 13, the present invention can be applied even in such a case.

Further, the elevator management system 20 has an external connection such as the Internet, so that the elevator management system 20 is installed in, for example, another building operated at the same working hours or business hours, and the operation data of the elevators 3 performing similar operations. Can be obtained. As a result, the elevator management system 20 can acquire a large amount of operation data for learning, and can improve learning accuracy.

Further, the elevator management system 20 has an external connection such as the Internet and uses weather data and operation information of public transportation as information for learning, so that learning accuracy can be improved.

Further, as illustrated in FIG. 14, the elevator management system 30 includes a cloud server 35 that calculates learning data, and a management server 31 that is a server device via a communication network 21 and communication devices 37 and 39 such as a switching hub and a router. And each elevator 3 may be connected. The cloud server 35 is composed of a cloud server, a data center, and the like.

With the configuration shown in FIG. 14, the high-performance cloud server 35 performs the processing centered on the operation data learning processing that requires heavy operation data transfer and learning processing, and frequently communicates with the elevator 3. The management server 31 can perform processing centered on the operation instruction processing that needs to be performed and communication delay is fatal.

Therefore, the elevator management system 30 can be installed in a place where the installation space is relatively limited, while reducing the influence of communication delay. By installing the management server 31 in the DMZ (demilitarized zone), the learning data acquisition module 34 can acquire the learning data and the operation instruction to the elevator 3 can be promptly performed.

Further, as shown in FIG. 15, in the elevator management system 50, by providing the communication module 54 in the management server 51 which is a server device, communication via the communication device 37 (FIG. 14) such as a switching hub or a router is performed. Is unnecessary. Therefore, the present invention can be applied even when the switching hub, the router, or the like cannot be used because of an environment where the switching hub, the router, or the like is not installed, or for security reasons. In the configuration of FIG. 15, the operation of the conventional elevator 3 can be restored by simply removing the management server 51.

Further, as shown in FIG. 16, when it is difficult to download the learning data via the communication network, the elevator management system 60 may be connected to an auxiliary storage device 63 such as an SD card in which the learning data TB30 is recorded. In the auxiliary storage device 63, a maintenance staff who regularly performs maintenance and inspection of the elevator 3 updates the learning data TB30 when carrying and performing maintenance and inspection work. The learning data TB30 is created by copying the learning data TB20. In the configuration of FIG. 16, the operation of the conventional elevator 3 can be restored simply by removing the management server 61, which is a server device.

Further, in the above-described embodiment, the case where deep learning is used as the learning means has been described, but the present invention is not limited to this, and a statistical method such as regression analysis may be used, or a machine other than deep learning may be used. You may use learning.

Further, in the above-described embodiment, the call button 16 and the destination floor designation of each car 12 in each floor after one round trip from only the pressed state of the call button 16 and the destination floor designation button of each car 12 in each floor. Although the case of predicting the pressed state of the button has been described, the present invention is not limited to this, and information about seasons such as spring, summer, autumn, winter, year information such as the year in which the Olympics are held and leap years, and time zone information such as morning and night. Etc. may be reflected.

Furthermore, in the above-described embodiment, the case where the local information in the building is not taken into consideration is described, but the present invention is not limited to this, and local information such as usage information of the conference room in the building is acquired from the communication path 19. It may be reflected in the prediction result.

1, 20, 30, 50, 60... Elevator management system, 2, 31, 51, 61... Management server, 3... Elevator, 4, 32, 52, 62... CPU, 5, 63... Auxiliary storage Device, 6, 33, 36, 53, 64... Memory, 7... Operation storage module, 8... Operation learning module, 9... Route determination module, 10... Route instruction module, 11... Elevator control device, 12... Car, 13... Main rope, 14... Balancing weight, 15... Hoisting machine, 16... Call button, 19, 24, 38, 40... Communication path, 21... Communication network, 22, 23, 37, 39... Communication equipment, 35... Cloud server.

Claims (6)

An elevator management system for managing an elevator including a control device for operating a car over a plurality of floors,
A management device for managing the control device,
The management device is
A receiving circuit for receiving the destination floor designation information and the car call information,
A memory for accumulating and recording received information of the receiving circuit;
A controller that learns the operation tendency of the car based on the information recorded in the memory,
An output circuit for outputting management information to the control device,
Equipped with
The controller is
Based on the result of the learning, the destination floor designation information after a predetermined time from the reception information of the receiving circuit, and predict the car call information,
Based on the result of the prediction after the predetermined time, the management information is formed to limit the range of the operating floor of the car,
An elevator management system in which the control device controls the operation of the car based on the management information. The control device is
The elevator management system according to claim 1, wherein a floor to be reached is determined based on predictions of the car call information and the destination floor designation information, and when the floor is reached, the traveling direction of the car is reversed. Predicting the car call information and the destination floor designation information for each of the cars from the present time to the time required for one operation of each of the cars, and determining the operation route of each of the cars. The elevator management system according to claim 1. After the prediction of the car call information and the destination floor designation information, if a car call or destination floor designation other than prediction occurs,
The elevator management system according to claim 1, wherein the operation route is corrected based on the current position of the car and the moving direction of the car. Based on the prediction of the car call information and the destination floor designation information, based on the difference between the time when the car call and the destination floor are designated and the actual time when the car call and the destination floor are designated, the door of the car is opened. Adjust time,
The elevator management system according to claim 1. A method for managing an elevator, wherein a management device manages a control device for operating an elevator car over a plurality of floors,
The management device is
Receives destination floor designation information and car call information,
The received information received is accumulated and recorded,
Based on the received information, learn the operation tendency of the car,
Outputs management information to the control device as a learning result,
Based on the learning result, the destination floor designation information after a predetermined time from the received information, and predict the car call information,
Based on the result of the prediction, the management information is formed to limit the range of the operating floor of the car,
Cause the control device to control the operation of the car based on the management information
Elevator management method.

Images