An elevator operation monitoring system and method based on internet of things

Abstract

The application discloses an elevator operation monitoring system and method based on the Internet of Things, which obtains various elevator fault information of each elevator, carries out data preprocessing according to the fault types of the elevator fault information to obtain an elevator fault data set, carries out feature extraction on the elevator fault data set to obtain a training data set, constructs a fault prediction model according to the training data set, receives an instruction sent by a user to an elevator and obtains the motion state of the elevator, determines the distribution decision of the instruction based on the instruction and the motion state of the elevator, constructs an operation evaluation index of the elevator according to the elevator operation mode and operation parameters of the elevator in the running state, completes real-time monitoring of the operation of the elevator based on the operation evaluation index of the elevator and the elevator fault prediction model, and automatically adjusts the optimal operation mode of the elevator in the system according to the elevator fault prediction model and the user elevator information which has not been executed, so that the comfort degree of the user when taking the elevator is improved, the elevator loss is reduced, and the safety of the elevator operation and the user elevator efficiency are improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of elevator monitoring, and particularly relates to an elevator operation monitoring system and method based on Internet of Things. BACKGROUND

[0002] With the rapid development of elevator Internet of Things technology, the elevator field develops rapidly along the direction of intelligentization and interconnection. Intelligent transformation of the elevator Internet of Things gateway can effectively improve the intelligent degree of the elevator gateway, transform the traditional data terminal into a micro data center in the Internet of Things, decentralize to a certain extent, improve the overall risk resistance, and ensure the stable and safe operation of the elevator. However, while elevators bring convenience to people, they also inevitably have safety risks. The use of large-capacity flow and large load is common in elevators, which leads to an increasing elevator failure rate and a significant shortening of the failure interval time. Once a failure occurs, it will cause huge losses. Since the traditional elevator maintenance mode mainly relies on manual inspection, there are problems of complex composition of elevator maintenance personnel and delayed transmission of elevator maintenance information. Therefore, it is urgent to provide a real-time elevator operation monitoring system to reduce the elevator failure rate and ensure the safety of elevator operation. SUMMARY

[0003] Therefore, the application provides an elevator operation monitoring system and method based on Internet of Things, which can monitor the elevator operation state in real time, ensure the safety of elevator operation, and increase the comfort of users taking the elevator, to solve the above technical problems. The following technical solutions are used to realize the application.

[0004] In a first aspect, the application provides an elevator operation monitoring system based on Internet of Things, comprising:

[0005] A data acquisition unit is configured to acquire elevator failure information of each elevator in each type of elevator failure time period, and to obtain an elevator failure data set by data preprocessing according to the failure type of the elevator failure information. The elevator failure types include safety circuit disconnection, door lock failure, and layer change stop failure. The data preprocessing includes abnormal data processing and data enhancement.

[0006] A feature extraction unit is configured to extract features from the elevator failure data set to obtain a training data set, and to construct a failure prediction model according to the training data set. The process of constructing the failure prediction model includes: performing t-step time shift operation on the elevator failure data set to obtain the training data set. The model input is the previous t-step data at the prediction time, and the output is the prediction of the current output using the previous t-step input. The elevator state at the current time under the action of the input information of the previous t-step through the training of the constructed LSTM network is used to realize the elevator failure prediction.

[0007] The decision distribution unit is configured to receive an instruction sent by a user to an elevator, acquire a motion state of the elevator, and determine a distribution decision of the instruction based on the instruction and the motion state of the elevator, wherein the motion state of the elevator includes a running state, a stopping state and an idle state of the elevator;

[0008] The operation monitoring unit is configured to construct an operation evaluation index of the elevator according to an elevator operation mode and an operation parameter of the elevator in the running state, and complete real-time monitoring of the elevator operation based on the operation evaluation index of the elevator and the elevator fault prediction model, wherein the operation evaluation index of the elevator includes an average waiting time, an average riding time, an elevator crowded degree and an elevator operation energy consumption.

[0009] As a further preferred embodiment of the above technical solution, the real-time monitoring of the elevator operation based on the operation evaluation index of the elevator and the elevator fault prediction model includes:

[0010] The average waiting time AWT is a ratio of a total of time consumed in a process from when a user sends an instruction to when the elevator receives the user to a total number n of instructions in a preset time, and an expression corresponding thereto is wherein T s (i) represents a time when the elevator receives the user, T g (i) represents a time when the elevator generates the instruction;

[0011] The average riding time ART is a ratio of a sum of time consumed by each floor of the elevator in a process from when a user starts to ride the elevator to when the user arrives at a target floor to the total number n of instructions in a preset time, and an expression corresponding thereto is wherein T f (i) represents a time when the user gets off the elevator, F s (i) represents a floor where the user rides the elevator, F f (i) represents a floor where the user gets off the elevator;

[0012] The elevator crowded degree D is a ratio of a sum of time integrals of crowded degrees of all floors of the user in a preset time to a total time spent, and when the D is lower than 0.5, the user feels not crowded, and an expression of the elevator crowded degree D is wherein N represents a total number of floors operated by the elevator in a preset time, P p represents a number of people in the elevator, N p represents an upper limit of a number of people that can be carried by the car, t N represents a time when the elevator operates N floors, and t1 represents a time when the elevator operates one floor;

[0013] The elevator operation energy consumption E e can be divided into an uplink operation energy consumption and a downlink operation energy consumption, the uplink operation energy consumption E up is energy consumed by an average load carried by each floor in an uplink process of the elevator, and the downlink operation energy consumption Edn The energy consumed for the average load per floor during the descending process of the elevator is calculated by the expression wherein S f represents the total number of floors operated in the preset period of the elevator, M g represents the weight of the elevator, M e represents the weight of the elevator car, M p represents the weight of each passenger.

[0014] As a further preferred embodiment of the above technical solution, the operation evaluation index of the elevator is constructed according to the elevator operation mode and operation parameters of the elevator in the running state, comprising:

[0015] When each elevator is in the normal operation mode, the elevator can receive all instructions, and each floor can stop to pick up and drop off users; when each elevator is in the single-double operation mode, part of the elevators receive single-floor stop information, and the other part of the elevators receive double-floor stop information, if the instruction starting floor and the target floor are single and double or double and single respectively, the next floor of the instruction starting floor is boarded; when each elevator is in the up-down operation mode, part of the elevators receive up instructions, and the other part of the elevators receive down instructions.

[0016] The elevator operation parameters include at least one of the number of people that the elevator can carry, the floor height, the time speed of the elevator running one floor, and the time speed of the elevator running multiple floors.

[0017] As a further preferred embodiment of the above technical solution, the instructions sent by the user to the elevator are received, and the motion state of the elevator is obtained, and the distribution decision of the instructions is determined based on the instructions and the motion state of the elevator, comprising:

[0018] When the user issues an instruction to the elevator, the instruction is in an unassigned state, and the distribution decision of the instruction needs to be made, and the instruction is assigned to a target elevator; when the unassigned instruction enters a program, all elevators are queried in turn, and when the program is in a query stage, it is first determined whether the queried elevator meets the conditions for participating in the distribution decision;

[0019] If yes, the index of the distribution decision is calculated, and the next elevator is queried;

[0020] If no, the next elevator is directly queried until all elevators are queried, and the distribution decision of the instruction is made according to the decision index.

[0021] As a further preferred embodiment of the above technical solution, the instructions sent by the user to the elevator are received, and the motion state of the elevator is obtained, comprising:

[0022] The elevator is in a running state: when the elevator is ascending due to an ascending instruction, the elevator receives the ascending instruction, the instruction starting layer is at least two layers higher than the current layer of the elevator and the instruction starting layer is not higher than the target layer of the elevator, the expression of the index ΔT of the elevator participating in the distribution decision is Wherein ΔF = |f s -F c |-1, f s The instruction starting layer is represented by F c The layer where the elevator is located is represented by H e The height of each layer of the elevator is represented by H sN The stopping time of the multi-layer elevator is represented by T maxN The maximum running speed of the multi-layer elevator is represented by V

[0023] When the elevator is ascending from a floor one layer lower than the target layer to the target layer and the instruction starting layer is the same as the target layer of the elevator, the expression of the index ΔT' of the elevator participating in the distribution decision is

[0024] If the user instruction is an ascending instruction and the instruction starting layer is at least two layers higher than the layer where the elevator is located and the target layer of the elevator is at least two layers away from the layer where the elevator is located, the elevator participates in the distribution decision; if the user instruction is a descending instruction and the instruction starting layer is not lower than the target layer of the elevator and the target layer of the elevator is at least two layers away from the layer where the elevator is located, the elevator participates in the distribution decision, at this time the index participating in the distribution decision is ΔT;

[0025] When the elevator is descending due to an ascending instruction, if the user instruction is a descending instruction and the instruction starting layer is at least two layers lower than the layer where the elevator is located and the target layer of the elevator is at least two layers away from the layer where the elevator is located, the elevator participates in the distribution decision; if the user instruction is an ascending instruction and the instruction starting layer is not higher than the target layer of the elevator and the target layer of the elevator is at least two layers away from the layer where the elevator is located, the elevator participates in the distribution decision, at this time the index participating in the distribution decision is ΔT'.

[0026] As a further optimization of the above technical solution, the elevator is in a stationary state: when the elevator is ascending due to an ascending instruction, the instruction starting layer of the ascending instruction is the same as the layer where the elevator is located, the index of the elevator participating in the distribution decision is 0; if the instruction starting layer of the ascending instruction is higher than the layer where the elevator is located, the index participating in the distribution decision is Wherein R represents the elevator stopping time, a random value in the range of 3-7s is taken in simulation, ΔF = |f s -F c |, T st1 The starting time of the elevator running one layer is represented by T sp1 The stopping time of the elevator running one layer is represented by T max1represents the maximum speed of the elevator when running one floor; if the command start floor of the down command is higher than the floor where the elevator is located, the index of the elevator participating in the allocation decision is ΔT1; if the command start floor of the down command is the same as the floor where the elevator is located, the index of the elevator participating in the allocation decision is 0; if the command start floor of the down command is lower than the floor where the elevator is located, the index of the elevator participating in the allocation decision is ΔT1; if the command start floor of the up-down command is lower than the floor where the elevator is located, the index of the elevator participating in the allocation decision is ΔT1;

[0027] When the elevator is running down due to a down command, if the command start floor of the down command is the same as the floor where the elevator is located, the index of the elevator participating in the allocation decision is 0; if the command start floor of the down command is lower than the floor where the elevator is located, the index of the elevator participating in the allocation decision is ΔT1; if the command start floor of the up command is lower than the floor where the elevator is located, the index of the elevator participating in the decision is ΔT1; if the command start floor of the up command is the same as the floor where the elevator is located, the index of the elevator participating in the allocation decision is 0; if the command start floor of the up command is higher than the floor where the elevator is located, the index of the elevator participating in the allocation decision is ΔT1; if the command start floor of the down command is higher than the floor where the elevator is located, the index of the elevator participating in the allocation decision is ΔT1;

[0028] When the elevator is in an idle state: when a user command occurs while the elevator is in an idle state, the expression of the index ΔT2 of the elevator directly participating in the allocation decision is Where ΔF = |F c -f st |.

[0029] As a further preferred embodiment of the above technical solution, when the allocation decision index calculation is completed and the elevator meets the condition of participating in the allocation decision, the minimum value ΔT min corresponding to the elevator number and quantity, if the minimum value ΔT min corresponding to the elevator quantity is 1, the command will be allocated to the elevator; if the minimum value ΔT min corresponding to the elevator quantity is greater than 1, the elevator to which the command will be allocated is determined according to whether the minimum value ΔT

[0030] As a further preferred embodiment of the above technical solution, feature extraction is performed on the elevator fault data set to obtain a training data set, and a fault prediction model is constructed according to the training data set, including:

[0031] The ReliefF algorithm is used to screen the elevator fault factors to obtain the key factors in the elevator fault prediction model, all feature weights are set to 0, a sample a is randomly extracted from the sample set A, k neighbor samples of a are taken out, and the weight value W of each feature is updated. The weight W is sorted.

[0032] As a further preferred embodiment of the above technical solution, the elevator fault information of each elevator in each type of elevator fault period is obtained, the elevator fault data set is obtained by data preprocessing according to the fault type of the elevator fault information, including:

[0033] The missing data in the elevator fault information is filled, if the data is numerical data, the mean value of the previous time and the next time is filled; if the data is state data, the data of the previous time is directly filled; the numerical abnormal sample is screened out, the sample does not meet the mapping conversion rule or the numerical value is out of the normal range, and the normal repair cannot be performed, and the deletion operation is performed on the sample;

[0034] The data in the elevator fault information is divided into numerical data and state data according to the type, the state data is converted into numerical data according to the numerical conversion rule, and the expression of the standardization algorithm for standardizing the data is Wherein z i Indicates the sample attribute, μ indicates the sample mean, and σ indicates the sample standard deviation, and the overall dimensionless elevator fault data set is obtained by standardization processing.

[0035] In a second aspect, the application also provides an elevator operation monitoring method based on the Internet of Things, including the following steps:

[0036] The elevator fault information of each elevator in each type of elevator fault period is obtained, and the elevator fault data set is obtained by data preprocessing according to the fault type of the elevator fault information, wherein the elevator fault type includes safety circuit disconnection, door lock failure and layer change stop fault, and the data preprocessing includes abnormal data processing and data enhancement;

[0037] The training data set is obtained by feature extraction of the elevator fault data set, and the fault prediction model is constructed according to the training data set, wherein the process of constructing the fault prediction model includes: the training data set is obtained by t-step time shift operation of the elevator fault data set, the model input is the previous t-step data of the prediction time, the output is the prediction of the current output using the previous t-step input, and the elevator state at the current time under the action of the input information of the previous t-step through the training of the constructed LSTM network to realize the elevator fault prediction;

[0038] The instruction sent by the user to the elevator is received, and the motion state of the elevator is obtained, and the distribution decision of the instruction is determined based on the instruction and the motion state of the elevator, wherein the motion state of the elevator includes that the elevator is in a running state, a stopping state and an idle state;

[0039] The operation evaluation index of the elevator is constructed according to the elevator operation mode and operation parameters when the elevator is in the running state, and the real-time monitoring of the elevator operation is completed based on the operation evaluation index of the elevator and the elevator fault prediction model, wherein the operation evaluation index of the elevator includes the average waiting time, the average boarding time, the elevator congestion degree and the elevator operation energy consumption.

[0040] The application provides an elevator operation monitoring system and method based on Internet of Things, which obtains elevator fault information of each elevator in each type of elevator fault period, performs data preprocessing according to the fault type of the elevator fault information to obtain an elevator fault data set, extracts features from the elevator fault data set to obtain a training data set, constructs a fault prediction model according to the training data set, receives an instruction sent by a user to the elevator, and obtains the motion state of the elevator, determines the distribution decision of the instruction based on the instruction and the motion state of the elevator, constructs an operation evaluation index of the elevator according to the elevator operation mode and operation parameters when the elevator is in the running state, and completes the real-time monitoring of the elevator operation based on the operation evaluation index of the elevator and the elevator fault prediction model, and automatically adjusts the optimal operation mode of the elevator in the system according to the elevator fault prediction model and the user boarding information that has not been executed, which guarantees the normal operation of the elevator monitoring system, improves the user boarding comfort, reduces the elevator loss, and improves the safety of the elevator operation and the user boarding efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some of the embodiments of the application, and therefore should not be considered as a limitation to the scope. For those skilled in the art, other related drawings can also be obtained without creative labor.

[0042] Figure 1 The structural block diagram of the elevator operation monitoring system based on Internet of Things provided by the embodiments of the application is shown in the figure.

[0043] Figure 2 The execution process diagram of the instruction distribution decision provided by the application is shown in the figure.

[0044] Figure 3 The flowchart of the elevator operation monitoring method based on Internet of Things provided by the application is shown in the figure. DETAILED DESCRIPTION

[0045] The embodiments of the application will be described in detail below, and the examples of the embodiments are shown in the drawings, wherein the same or similar reference signs represent the same or similar elements or elements with the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and are only used to explain the application, and cannot be understood as a limitation to the application.

[0046] Referring to Figure 1 The application provides an elevator operation monitoring system based on the Internet of Things, comprising:

[0047] A data acquisition unit is configured to acquire elevator fault information of each elevator in each type of elevator fault period, and to obtain an elevator fault data set by performing data preprocessing according to the fault type of the elevator fault information, wherein the elevator fault type includes safety circuit disconnection, door lock failure and layer switching stoppage failure, and the data preprocessing includes abnormal data processing and data enhancement.

[0048] A feature extraction unit is configured to perform feature extraction on the elevator fault data set to obtain a training data set, and to construct a fault prediction model according to the training data set, wherein the process of constructing the fault prediction model includes: performing t-step time shift operation on the elevator fault data set to obtain the training data set, the model input is the previous t-step data at the prediction time, the output is the prediction of the current output using the previous t-step input, and the elevator state at the current time under the action of the input information of the previous t-step through the training of the constructed LSTM network is used to realize elevator fault prediction.

[0049] A decision allocation unit is configured to receive an instruction sent by a user to an elevator, and to acquire a motion state of the elevator, and to determine an allocation decision of the instruction based on the instruction and the motion state of the elevator, wherein the motion state of the elevator includes that the elevator is in a running state, a stopping state and an idle state.

[0050] An operation monitoring unit is configured to construct an operation evaluation index of the elevator according to an elevator operation mode and an operation parameter when the elevator is in the running state, and to complete real-time monitoring of the elevator operation based on the operation evaluation index of the elevator and the elevator fault prediction model, wherein the operation evaluation index of the elevator includes average waiting time, average riding time, elevator crowding degree and elevator operation energy consumption.

[0051] In this embodiment, the real-time monitoring of the elevator operation based on the operation evaluation index of the elevator and the elevator fault prediction model includes: the average waiting time AWT is the ratio of the total time consumed by the user in the process from issuing an instruction to receiving the instruction by the elevator to the total number n of instructions in a preset time, and the corresponding expression is wherein T s (i) represents the time when the elevator receives the instruction of the user, T g (i) represents the time when the elevator instruction is generated; the average riding time ART is the ratio of the sum of the time consumed by each floor of the elevator in the process from the user starting to ride the elevator to the user reaching the target floor to the total number n of instructions in a preset time, and the corresponding expression is wherein T f (i) represents the time when the user gets off the elevator, F s (i) represents the floor where the user rides the elevator, F f(i) represents the floor where the user gets out of the elevator; the elevator congestion degree D is the ratio of the time integral sum of the user congestion degree of all floors within the preset time and the total time spent, when D is lower than 0.5, the user feels uncongested, and the expression of the elevator congestion degree D is where N represents the total number of floors operated by the elevator within the preset time, P p represents the number of people in the elevator, N p represents the upper limit of the number of people that the car can carry, t N represents the time for which the elevator operates N floors, t1 represents the time for which the elevator operates one floor; the elevator operation energy consumption E e can be divided into uplink operation energy consumption and downlink operation energy consumption, the uplink elevator operation energy consumption E up is the energy consumed by the average load per floor transported in the uplink process of the elevator, the downlink operation energy consumption E dn is the energy consumed by the average load per floor transported in the downlink process of the elevator, and the calculation expression is where S f represents the total number of floors operated by the elevator within the preset period, M g represents the weight of the elevator, M e represents the weight of the elevator car, M p represents the weight of each passenger.

[0052] It should be noted that the elevator operation evaluation index is constructed according to the elevator operation mode and operation parameters of the elevator in the running state, including when each elevator is in the normal operation mode, the elevator can receive all instructions, and each floor can stop up and down users; when each elevator is in the single-double operation mode, part of the elevators receive single-floor stop information, and the other part of the elevators receive double-floor stop information, if the instruction starting floor and the target floor are single and double or double and single respectively, the instruction starting floor next floor takes the elevator; when each elevator is in the up-down operation mode, part of the elevators receive up instruction, and the other part of the elevators receive down instruction; the elevator operation parameters include at least one of the number of people the elevator can carry, the floor height, the time speed of the elevator running one floor, and the time speed of the elevator running multiple floors. The motion state of the elevator includes three cases of the elevator being in the running state, the stopping state and the idle state, and the three cases can be converted according to specific adjustment, two conditions need to be met when the elevator changes from the idle state to the running state, condition one is that the elevator in the idle state is assigned an instruction and the instruction initial floor is not equal to the current floor of the elevator, condition two is that the current floor of the elevator in the idle state is not one floor and the stop time reaches the waiting time, and the elevator will start to go to one floor. The condition for the elevator running state to change to the stopping state is that the elevator reaches the target floor, and the condition for the elevator to change from the static state to the running state is that the elevator receives an instruction and sets off to the target floor. The elevator changes from the stopping state to the idle state, which requires that the elevator has no instruction when stopping and the instruction state reaches a certain time, and the condition for the elevator to change from the idle state to the stopping state is that the elevator is assigned an instruction and the instruction starting floor is equal to the current floor of the elevator.

[0053] It should be understood that the elevator fault prediction refers to judging whether the elevator is likely to fail in the future through changes in the elevator part operation data before the fault occurs, and the intelligent gateway installed on the elevator is used to monitor and collect real-time operation data and historical data of the elevator in real time. The data obtained by the intelligent gateway is rich in types, including system state, running state, door state and other types of data. The elevator fault early warning model processes and extracts features from the data. Due to the differences in data types, numerical ranges and dimensions of various types of information, the data needs to be preprocessed. The prediction model obtains the certainty or probability value of the elevator failure in the future. The safety circuit is disconnected, the door lock fails, and the layer switching stops. The fault type occurs frequently in the actual maintenance process of the elevator, which can provide more data support for encoding and classifying the output type. The abnormal data in the data set is filled, deleted and other operations, and the various types of information in the data set are fitted and converted to realize standardization processing. When the fault information in the data set is insufficient, the data set needs to be enhanced. By obtaining the elevator fault information of each elevator in each type of elevator fault time period, the data is preprocessed according to the fault type of the elevator fault information to obtain the elevator fault data set. The feature extraction of the elevator fault data set obtains the training data set, and the fault prediction model is constructed according to the training data set. The instruction sent by the user to the elevator is received, and the motion state of the elevator is obtained. Based on the instruction and the motion state of the elevator, the allocation decision of the instruction is determined. According to the elevator running mode and the running parameter of the elevator in the running state, the running evaluation index of the elevator is constructed. Based on the running evaluation index of the elevator and the elevator fault prediction model, the real-time monitoring of the elevator running is completed. According to the elevator fault prediction model and the user's unexecuted elevator information, the optimal running mode of the elevator in the system is automatically adjusted, which ensures the normal operation of the elevator monitoring system, improves the user's elevator comfort, reduces the elevator loss, and improves the safety and efficiency of the elevator operation.

[0054] Optionally, the instruction sent by the user to the elevator is received, and the motion state of the elevator is obtained. Based on the instruction and the motion state of the elevator, the allocation decision of the instruction is determined, comprising:

[0055] S1: When the user issues an instruction to the elevator, the instruction is in an unallocated state, and the instruction needs to be allocated to the target elevator; when the unallocated instruction enters the program, all elevators are queried in turn, and when the program is in the query stage, it is judged whether the queried elevator meets the conditions for participating in the allocation decision;

[0056] S2: If yes, the index of the allocation decision is obtained, and the next elevator is queried;

[0057] S3: If no, the next elevator is directly queried until all elevators are queried, and the instruction is allocated according to the decision index.

[0058] In this embodiment, the user sends instructions to the elevator, and the motion state of the elevator is obtained, including: the elevator is in a running state: when the elevator is ascending due to an ascending instruction, the elevator receives the ascending instruction, the instruction starting layer of the ascending instruction is at least two layers higher than the current layer of the elevator, and the instruction starting layer is not higher than the target layer of the elevator, the expression of the index ΔT of the elevator participating in the distribution decision is wherein ΔF = |f s -F c |-1, f s represents the instruction starting layer, F c represents the layer where the elevator is located, H e represents the height of each layer of the elevator, T sN represents the stopping time of the multi-layer elevator, V maxN represents the maximum running speed of the multi-layer elevator; when the elevator is ascending from a floor one layer lower than the target layer to the target layer and the instruction starting layer is the same as the target layer of the elevator, the expression of the index ΔT' of the elevator participating in the distribution decision is If the user instruction is an ascending instruction, and the instruction starting layer is at least two layers higher than the layer where the elevator is located and the target layer of the elevator is at least two layers apart from the layer where the elevator is located, the elevator participates in the distribution decision; if the user instruction is a descending instruction, and the instruction starting layer is not lower than the target layer of the elevator and the target layer of the elevator is at least two layers apart from the layer where the elevator is located, the elevator participates in the distribution decision, and the index participating in the distribution decision is ΔT; when the elevator is descending due to an ascending instruction, if the user instruction is a descending instruction, and the instruction starting layer is at least two layers lower than the layer where the elevator is located and the target layer of the elevator is at least two layers apart from the layer where the elevator is located, the elevator participates in the distribution decision; if the user instruction is an ascending instruction, and the instruction starting layer is not higher than the target layer of the elevator and the target layer of the elevator is at least two layers apart from the layer where the elevator is located, the elevator participates in the distribution decision, and the index participating in the distribution decision is ΔT'.

[0059] It should be noted that when the elevator is in a stationary state: when the elevator is ascending due to an ascending instruction, the instruction starting layer of the ascending instruction is the same as the layer where the elevator is located, the index of the elevator participating in the distribution decision is 0; if the instruction starting layer of the ascending instruction is higher than the layer where the elevator is located, the index of the elevator participating in the distribution decision is wherein R represents the elevator stopping time, a random value in the range of 3-7s is taken in simulation, ΔF = |f s -F c |, T st1 represents the starting time of the elevator running one layer, T sp1 represents the stopping time of the elevator running one layer, V max1represents the maximum speed of the elevator when running one floor; the indicator of the elevator participating in the allocation decision is ΔΤ1 when the command start floor of the down command is higher than the floor where the elevator is located; the indicator of the elevator participating in the allocation decision is 0 when the command start floor of the down command is the same as the floor where the elevator is located; the indicator of the elevator participating in the allocation decision is ΔΤ1 when the command start floor of the down command is lower than the floor where the elevator is located; the indicator of the elevator participating in the allocation decision is ΔΤ1 when the command start floor of the up-down command is lower than the floor where the elevator is located; the indicator of the elevator participating in the allocation decision is 0 when the elevator is running down due to a down command and the command start floor of the down command is the same as the floor where the elevator is located; the indicator of the elevator participating in the allocation decision is ΔΤ1 when the command start floor of the down command is lower than the floor where the elevator is located; the indicator of the elevator participating in the allocation decision is ΔΤ1 when the command start floor of the up command is lower than the floor where the elevator is located; the indicator of the elevator participating in the allocation decision is 0 when the command start floor of the up command is the same as the floor where the elevator is located; the indicator of the elevator participating in the allocation decision is ΔΤ1 when the command start floor of the up command is higher than the floor where the elevator is located; the indicator of the elevator participating in the allocation decision is ΔΤ1 when the command start floor of the down command is higher than the floor where the elevator is located; the elevator is in an idle state: when a user command appears while the elevator is in an idle state, the expression of the indicator ΔΤ2 of the elevator directly participating in the allocation decision is where ΔF = |F c -f st |.

[0060] It should be understood that when the allocation decision indicator calculation is completed and the elevator meets the condition of participating in the allocation decision, the minimum value ΔΤ min corresponding to the elevator number and quantity, if the minimum value ΔΤ min corresponds to the elevator quantity being 1, the command will be allocated to the elevator; if the minimum value ΔΤ min corresponds to the elevator quantity being greater than 1, the elevator to which the command will be allocated is decided according to whether the minimum value ΔΤ min is zero and the state of the elevator. When the elevator is in a running state, the minimum value ΔΤ minWhen the minimum value of the decision-making allocation index is zero, the elevator is in the same floor as the starting floor of the command being allocated. Since the elevator in the stop state is in the process of opening the door for the user to get on or off, the user does not need to wait and can directly get on the elevator, so the elevator in the stop state will be given priority to receive the command. When there is only an elevator in the idle state, the minimum value of the decision-making allocation index is zero, and the command is allocated to the elevator in the idle state. For the case where the minimum value of the decision-making allocation index is not zero, if the allocation decision index of the elevator in the running, stop and idle states is the same as the minimum value at the same time, in order to ensure that the user can definitely take the elevator, the priority order of the command is the elevator in the idle state, the elevator in the running state and the elevator in the stop state. When there are multiple elevators in the same state that meet the allocation decision condition at the same time, the command will be preferentially allocated to the elevator with a smaller number. At the same time that the command is allocated to the elevator, the system will record the command allocation time, the command state and the number of the elevator to which the command is allocated.

[0061] Optionally, feature extraction is performed on the elevator fault data set to obtain a training data set, and a fault prediction model is constructed according to the training data set, including:

[0062] The ReliefF algorithm is used to screen the elevator fault factors to obtain key factors in the elevator fault prediction model, the weight of all features is set to 0, a sample a is randomly extracted from the sample set A, k neighbor samples of a are taken out, and the weight value W of each feature is updated. The weight W is sorted.

[0063] In the embodiment, elevator fault information of time periods of various elevator faults of each elevator is obtained, data preprocessing is performed on the fault types of the elevator fault information to obtain an elevator fault data set, including: filling missing data in the elevator fault information, if the data is numerical data, filling the mean value of the previous time and the next time; if the data is state data, directly filling the data of the previous time; screening out numerical abnormal samples, which do not meet the mapping conversion rule or the value exceeds the normal range and cannot be normally repaired, and performing a deletion operation on the samples; the data in the elevator fault information is divided into numerical data and state data according to the type, the state data is converted into numerical data according to the numerical conversion rule, and a standardization algorithm is used to perform a standardization operation on the data, and the expression is wherein z i represents the sample attribute, μ represents the sample mean, and σ represents the sample standard deviation. The overall dimensionless elevator fault data set is obtained through the standardization processing, thereby improving the overall quality of the data set and enhancing the accuracy of the model prediction.

[0064] Referring to Figure 3 The application further provides an elevator operation monitoring method based on the Internet of Things, including the following steps:

[0065] S10: Obtain elevator fault information of each elevator in each type of elevator fault time period, and perform data preprocessing according to the fault type of the elevator fault information to obtain an elevator fault data set, wherein the elevator fault type includes safety circuit disconnection, door lock failure and layer change stop failure, and the data preprocessing includes abnormal data processing and data enhancement;

[0066] S11: Feature extraction is performed on the elevator fault data set to obtain a training data set, and a fault prediction model is constructed according to the training data set, wherein the process of constructing the fault prediction model includes: performing t-step time shift operation on the elevator fault data set to obtain the training data set, the model input is the previous t-step data of the prediction time, the output is the prediction of the current output using the previous t-step input, and the elevator state at the current time under the action of the input information of the previous t-step through the training of the constructed LSTM network is used to realize the elevator fault prediction;

[0067] S12: Receive the instruction sent by the user to the elevator, and obtain the motion state of the elevator, determine the distribution decision of the instruction based on the instruction and the motion state of the elevator, wherein the motion state of the elevator includes that the elevator is in a running state, a stop state and an idle state;

[0068] S13: Construct an elevator running evaluation index according to the elevator running mode and running parameter of the elevator in the running state, and complete real-time monitoring of the elevator running based on the elevator running evaluation index and the elevator fault prediction model, wherein the elevator running evaluation index includes average waiting time, average riding time, elevator congestion degree and elevator running energy consumption.

[0069] In this embodiment, the reasons for the elevator failure are numerous, including not only the running state of the elevator itself, but also the environment in the elevator control cabinet or other factors, the data types collected by the intelligent gateway are complex, the relationship between the parameters cannot be confirmed, and some parameters may have an adverse effect on subsequent fault prediction. The data set needs to be feature extracted, the data dimension is reduced, the main factors and secondary factors in the parameters are analyzed, and the predictability of the model for future failures is enhanced. By obtaining the elevator fault information of each elevator in each type of elevator fault time period, the data is preprocessed according to the fault type of the elevator fault information to obtain an elevator fault data set, the training data set is obtained by feature extraction on the elevator fault data set, the fault prediction model is constructed according to the training data set, the instructions sent by the user to the elevator are received, and the motion state of the elevator is obtained, the allocation decision of the instructions is determined based on the instructions and the motion state of the elevator, the running evaluation index of the elevator is constructed according to the running mode and the running parameter of the elevator in the running state, the real-time monitoring of the elevator running is completed based on the running evaluation index of the elevator and the elevator fault prediction model, the optimal running mode of the elevator in the system is automatically adjusted according to the elevator fault prediction model and the user's elevator information that has not been executed, the normal operation of the elevator monitoring system is ensured, the user's elevator comfort is improved, and the elevator loss is reduced, thereby improving the safety of the elevator running and the user's elevator efficiency.

[0070] In all examples shown and described herein, any specific values should be interpreted as merely exemplary and not as a limitation, and thus other examples of the example embodiments can have different values.

[0071] It should be noted that like reference numerals and letters refer to like items throughout the attached drawings, and once an item is defined in one drawing, it need not be further defined and explained in subsequent drawings.

[0072] The above-described embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the present application. It should be noted that for ordinary skilled persons in the art, without departing from the concept of the present application, several modifications and improvements can be made, which are within the protection scope of the present application.

Claims

1. An Internet of Things based elevator operation monitoring system, characterized by, The method comprises the following steps: a data acquisition unit is configured to acquire elevator fault information of each elevator in each type of elevator fault period, and perform data preprocessing according to the fault type of the elevator fault information to obtain an elevator fault data set, wherein the elevator fault type includes safety circuit disconnection, door lock failure and layer switching stop failure, and the data preprocessing includes abnormal data processing and data enhancement; a feature extraction unit is configured to perform feature extraction on the elevator fault data set to obtain a training data set, and construct a fault prediction model according to the training data set, wherein the process of constructing the fault prediction model includes: performing t-step time shift operation on the elevator fault data set to obtain the training data set, the model input is the previous t-step data of the prediction time, the output is the prediction of the current output using the previous t-step input, and the elevator state at the current time under the action of the input information of the previous t-step through the training of the LSTM network is used to realize the elevator fault prediction; a decision allocation unit is configured to receive an instruction sent by a user to an elevator, acquire a motion state of the elevator, and determine an allocation decision of the instruction based on the instruction and the motion state of the elevator, wherein the motion state of the elevator includes that the elevator is in a running state, a stopping state and an idle state; a running monitoring unit is configured to construct an elevator running evaluation index according to an elevator running mode and a running parameter of the elevator in the running state, and complete real-time monitoring of the elevator running based on the elevator running evaluation index and the elevator fault prediction model, wherein the elevator running evaluation index includes average waiting time, average boarding time, elevator crowding degree and elevator running energy consumption; receiving an instruction sent by a user to an elevator, and acquiring a motion state of the elevator, includes: The elevator is in the running state: when the elevator goes up due to the up command, the elevator receives the up command, the elevator command starting layer is at least two layers higher than the current layer of the elevator and the command starting layer is not higher than the target layer of the elevator, the indicators of the elevator participating in the distribution decision The expression is , , represents the command starting layer, represents the layer where the elevator is located, represents the height of the elevator at each layer, represents the stopping time of the multi-layer elevator, represents the maximum running speed of the multi-layer elevator; Indicators of the elevator participating in the allocation decision when the elevator is ascending from a floor one level below the target level to the target level and the command starting level is the same as the elevator target level The expression for ; If the user instruction is an up instruction, and the instruction starting floor is at least two floors higher than the floor where the elevator is located and the target floor of the elevator is at least two floors away from the floor where the elevator is located, the elevator participates in the allocation decision; if the user instruction is a down instruction, and the instruction starting floor is not lower than the target floor of the elevator and the target floor of the elevator is at least two floors away from the floor where the elevator is located, the elevator participates in the allocation decision, and the index participating in the allocation decision is . When the elevator is going down due to an up direction instruction, if the user is a down direction instruction, the instruction starting floor is at least two floors lower than the floor where the elevator is located, and the elevator target floor is at least two floors away from the floor where the elevator is located, the elevator participates in the allocation decision; if the user instruction is an up direction instruction, the instruction starting floor is not higher than the elevator target floor, and the elevator target floor is at least two floors away from the floor where the elevator is located, the elevator participates in the allocation decision, at this time the index participating in the allocation decision is .

2. The Internet of Things based elevator operation monitoring system as claimed in claim 1 wherein, completing real-time monitoring of the elevator running based on the elevator running evaluation index and the elevator fault prediction model, includes: The average waiting time AWT is the ratio of the sum of the time consumed in the process from the user's instruction to the elevator's response to the user within a preset time to the total number of instructions n in the period, and the corresponding expression is wherein represents the time when the elevator responds to the user, represents the time when the elevator instruction is generated; The average riding time ART is the ratio of the sum of the time consumed by the elevator running each floor during the process from when the user starts to ride the elevator to when the user reaches the target floor to the total number of instructions n in the preset time, and the corresponding expression is wherein represents the time when the user gets out of the elevator, represents the floor where the user rides the elevator, represents the floor where the user gets out of the elevator; The elevator congestion degree D is the ratio of the time integral sum of the congestion degrees of all floor users within a preset time and the total time spent, and when D is lower than 0.5, the user feels uncongested, and the expression of the elevator congestion degree D is wherein N represents the total number of floors operated by the elevator within a preset time, represents the number of people in the elevator, represents the upper limit of the number of people that the car can carry, represents the time during which the elevator has operated N floors, represents the time during which the elevator has operated one floor; Elevator operation energy consumption The elevator operation energy consumption can be divided into upgoing operation energy consumption and downgoing operation energy consumption, the upgoing elevator operation energy consumption is the energy consumed for the elevator to do work for carrying the average load of each floor in the upgoing process, the downgoing operation energy consumption is the energy consumed for the elevator to do work for carrying the average load of each floor in the downgoing process, then the calculation expression is , wherein represents the total number of floors operated in the preset period of the elevator, represents the weight of the elevator, represents the weight of the elevator car, represents the weight of each passenger.

3. The Internet of Things based elevator operation monitoring system as claimed in claim 1 wherein, constructing an elevator running evaluation index according to an elevator running mode and a running parameter of the elevator in the running state, includes: when each elevator is in a normal running mode, the elevator can receive all instructions, and each floor can stop to pick up and drop off users; when each elevator is in a single-double running mode, part of the elevators receive single-floor stop information, and the other part of the elevators receive double-floor stop information, if the instruction starting floor and the target floor are single and double or double and single respectively, the next floor of the instruction starting floor boards; when each elevator is in an up-down running mode, part of the elevators receive up instructions, and the other part of the elevators receive down instructions; the elevator running parameters include at least one of the number of people that the elevator can carry, the floor height, the time speed of the elevator running one floor, and the time speed of the elevator running multiple floors.

4. The Internet of Things based elevator operation monitoring system as claimed in claim 1 wherein, receiving an instruction sent by a user to an elevator, and acquiring a motion state of the elevator, determining an allocation decision of the instruction based on the instruction and the motion state of the elevator, includes: when the user issues an instruction to the elevator, the instruction is in an unallocated state, and the instruction needs to be allocated to a target elevator; when the unallocated instruction enters the program, all elevators are queried in turn, and when the program is in the query stage, it is judged whether the queried elevator meets the condition of participating in the allocation decision; if yes, the index of the allocation decision is calculated, and the next elevator is queried; If not satisfied, directly query the next elevator, until all elevator query is completed, according to the decision index to command allocation decision.

5. The Internet of Things based elevator operation monitoring system as claimed in claim 1 wherein, Also includes: The elevator is in a static state: when the elevator is ascending due to an up command, the command starting layer of the up command is the same as the layer where the elevator is located, and the index of the elevator participating in the distribution decision is 0; if the command starting layer of the up command is higher than the layer where the elevator is located, the index of the elevator participating in the distribution decision is , where R represents the elevator stop layer time, a random value in the range of 3-7s in simulation, , , represents the start time of the elevator running one layer, , represents the stop time of the elevator running one layer, , represents the maximum speed of the elevator running one layer; if the command starting layer of the down command is higher than the layer where the elevator is located, the index of the elevator participating in the distribution decision is ; if the command starting layer of the down command is the same as the layer where the elevator is located, the index of the elevator participating in the distribution decision is 0; if the command starting layer of the down command is lower than the layer where the elevator is located, the index of the elevator participating in the distribution decision is ; if the command starting layer of the up-down command is lower than the layer where the elevator is located, the index of the elevator participating in the distribution decision is ; When the elevator is going down due to a down command, if the command start floor of the down command is the same as the current floor of the elevator, the index of the elevator participating in the distribution decision is 0; if the command start floor of the down command is lower than the current floor of the elevator, the index of the elevator participating in the distribution decision is ; if the command start floor of the up command is lower than the current floor of the elevator, the index of the elevator participating in the decision is ; if the command start floor of the up command is the same as the current floor of the elevator, the index of the elevator participating in the distribution decision is 0; if the command start floor of the up command is higher than the current floor of the elevator, the index of the elevator participating in the distribution decision is ; if the command start floor of the down command is higher than the current floor of the elevator, the index of the elevator participating in the distribution decision is ; Elevator in idle state: When a user instruction occurs while the elevator is in idle state, the indicator of the elevator directly participating in the allocation decision The expression for where .

6. The Internet of Things based elevator operation monitoring system as claimed in claim 5, wherein, When the allocation decision index calculation is completed and the elevator meets the condition of participating in the allocation decision, the minimum value of the allocation decision index is found The corresponding elevator number and quantity, if the decision index minimum value The corresponding elevator quantity is 1, and the instruction will be allocated to the elevator; if the decision criterion minimum value corresponding to the number of elevators is greater than one, the decision criterion minimum value is zero and the state of motion of the elevators, the elevator to which the command is to be assigned is decided.

7. The Internet of Things based elevator operation monitoring system as claimed in claim 1 wherein, Feature extraction is carried out on the elevator fault data set to obtain the training data set, and a fault prediction model is constructed according to the training data set, including: The ReliefF algorithm is used to screen the elevator fault factors to obtain the key factors in the elevator fault prediction model, the weight of all features is set to 0, a sample a is randomly extracted in the sample set A, k neighbor samples of a are taken out, and the weight value W of each feature is updated, and the weight W is sorted.

8. The Internet of Things based elevator operation monitoring system as claimed in claim 1 wherein, Obtain the elevator fault information of each elevator in each type of elevator fault time period, and perform data preprocessing on the elevator fault information according to the fault type to obtain the elevator fault data set, including: The missing data in the elevator fault information is filled, if the data is numerical data, the mean value of the previous time and the next time is filled; If the data is state data, the data of the previous time is directly filled; The samples of numerical anomalies are screened out, which do not meet the mapping conversion rule or the numerical value is out of the normal range and cannot be repaired normally, and the delete operation is performed on the samples; The data in the elevator fault information is divided into numerical data and state data according to the type, the state data is converted into numerical data according to the numerical conversion rule, and the expression of the standardization algorithm for the data is , wherein represents the sample attribute, represents the sample mean, represents the sample standard deviation, and the overall dimensionless elevator fault data set is obtained through the standardization processing.

9. An Internet of Things-based elevator operation monitoring method of an Internet of Things-based elevator operation monitoring system according to any one of claims 1 to 8, characterized by, Including the following steps: Obtain the elevator fault information of each elevator in each type of elevator fault time period, and perform data preprocessing on the elevator fault information according to the fault type to obtain the elevator fault data set, wherein the elevator fault type includes safety circuit disconnection, door lock failure and layer change stop failure, and the data preprocessing includes abnormal data processing and data enhancement; Feature extraction is carried out on the elevator fault data set to obtain the training data set, and a fault prediction model is constructed according to the training data set, wherein the process of constructing the fault prediction model includes: t-step time shift operation is carried out on the elevator fault data set to obtain the training data set, the model input is the previous t-step data of the prediction time, the output is the prediction of the current output using the previous t-step input, and the elevator state at the current time under the action of the input information of the previous t-step through the training of the LSTM network is used to realize the elevator fault prediction; Receive the instruction sent by the user to the elevator, and obtain the motion state of the elevator, determine the allocation decision of the instruction based on the instruction and the motion state of the elevator, wherein the motion state of the elevator includes the running state, the stopping state and the idle state of the elevator; According to the elevator running mode and running parameter of the elevator in the running state, the running evaluation index of the elevator is constructed, and the real-time monitoring of the elevator running is completed based on the running evaluation index of the elevator and the elevator fault prediction model, wherein the running evaluation index of the elevator includes the average waiting time, the average riding time, the elevator congestion degree and the elevator running energy consumption.

Images