CN117985557A - Elevator cluster control method and system - Google Patents

Abstract

The present invention discloses an elevator cluster control method and system, including: the present invention proposes an elevator cluster control method and system, obtaining the physical characteristics of the elevator to be controlled in the target building in the previous cycle and the passenger's up-and-down behavior characteristics, and constructing a nonlinear dynamic mathematical model; obtaining the passenger flow data in the previous cycle of the elevator to be controlled in the target building, and combining the machine learning algorithm and the recursive least squares method to establish a passenger flow prediction model considering the system time lag; through the model prediction controller combined with the nonlinear dynamic mathematical model and the passenger flow prediction model, predicting the state and output of multiple future moments, and combining the preset control objective function and constraint conditions to formulate an optimal control action sequence; according to the calculated optimal control action sequence, sending specific control instructions to each elevator, so that the elevator runs according to the predetermined plan, and completing the update of the elevator cluster control method.

Description

Elevator cluster control method and system

Technical Field

The present invention relates to the technical field of elevator cluster control, and in particular to an elevator cluster control method and system.

Background technique

In modern buildings and urban infrastructure, elevators are an important part of vertical transportation. Their operating efficiency, safety and passenger comfort are key indicators for measuring the performance of elevator systems. With the increasing number of high-rise buildings and the development of intelligent trends, the control of a single elevator can no longer meet the efficient and personalized transportation needs of large commercial complexes, super high-rise residential or office buildings. Therefore, elevator cluster control systems came into being.

Traditional single-unit elevator control systems focus on the dispatch and service of a single elevator, while elevator cluster control integrates multiple elevators and performs unified management and optimized dispatch to maximize overall performance. This technology involves complex data analysis, prediction algorithms, and efficient communication networks. It can obtain real-time status information of each elevator (such as location, load, destination request, etc.) and dynamically generate the optimal dispatch strategy based on this data. It provides a rapid response channel for special groups or emergencies to ensure the safe evacuation capability of the building.

However, existing elevator cluster control technology still faces a series of challenges, including how to process massive real-time data, design more advanced intelligent scheduling algorithms to cope with complex changes in floor traffic flow patterns, and improve system stability and reliability while ensuring service quality.

Summary of the invention

The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

In view of the above existing problems, the present invention is proposed.

Therefore, the present invention provides an elevator cluster control method and system, which can solve the problems mentioned in the background technology.

In order to solve the above technical problems, the present invention provides the following technical solutions, a method for controlling an elevator cluster, comprising:

Obtain the physical characteristics of the elevator to be controlled in the target building in the previous cycle and the characteristics of the passenger's going up and down stairs behavior, and construct a nonlinear dynamic mathematical model, wherein the elevator control method used in the previous cycle of the elevator is any existing control method;

Obtain the passenger flow data of the elevator to be controlled in the target building in the previous cycle, and establish a passenger flow prediction model that takes into account system time lag by combining machine learning algorithm and recursive least squares method;

By combining the nonlinear dynamic mathematical model and the passenger flow prediction model through a model predictive controller, the states and outputs at multiple future moments are predicted, and an optimal control action sequence is formulated in combination with a preset control objective function and constraints;

According to the calculated optimal control action sequence, specific control instructions are sent to each elevator to make the elevator run according to the predetermined plan and complete the update of the elevator cluster control method.

As a preferred solution of the elevator cluster control method of the present invention, wherein: the physical characteristics of the elevator to be controlled in the target building in the previous cycle and the characteristics of the passenger's upstairs and downstairs behavior are obtained to construct a nonlinear dynamic mathematical model, and the elevator control method used in the previous cycle of the elevator is any existing control method including:

The physical characteristics of the elevator include at least the elevator floor z(t), the elevator speed v(t), the passenger capacity, and the passenger car movement C(t), where t represents the elevator operation time;

The passenger going up and down stairs behavior characteristics at least include the passengers entering and leaving each floor;

The nonlinear dynamic mathematical model includes the elevator position change equation and the passenger capacity change equation:

The position change equation of the elevator is expressed as:

Among them, m is the total mass of the elevator and passengers, _Te is the traction force provided by the traction motor, which is related to the elevator position, speed and passenger capacity, _Ff is the friction force related to the elevator speed and passenger capacity, which increases with the increase of speed and passenger capacity, _Fd is the damping force considering the influence of personnel changes, which is related to the elevator speed and personnel distribution, u(t) represents the force of the current control signal acting on the elevator system, Indicates the acceleration of the elevator car;

The passenger capacity variation equation is expressed as:

where _Pin (t,z(t)) and _Pout (t,z(t)) represent the number of passengers entering and leaving the elevator at floor z(t) at time t, respectively.

As a preferred solution of the elevator cluster control method of the present invention, wherein: the physical characteristics of the elevator to be controlled in the target building in the previous cycle and the characteristics of the passenger's upstairs and downstairs behavior are obtained to construct a nonlinear dynamic mathematical model, and the elevator control method used in the previous cycle of the elevator is any existing control method, and further includes: estimating unknown parameters or function forms in the model based on actual data combined with system identification technology;

The traction force Te _is estimated according to the elevator traction characteristics and passenger capacity, and the estimation formula is as follows:

_Te (z(t),v(t),C(t))＝ktz ₍ t)+ _cvv (t)+ _kcC (t)

The friction force _Ff estimation formula is as follows:

F _i (z(t),v(t),C(t))＝μ _i (m ₁ +C(t)k _p )v(t)

The damping force _Fd estimation formula is as follows:

F _d (z(t),v(t),C(t)) = k _d v(t) + k _c ′C(t)v(t)

Wherein, _m1 represents the inherent mass of the elevator, _kt represents the correlation coefficient between the traction force and the elevator position, _cv represents the correlation coefficient between the traction force and the elevator speed, _kc represents the correlation coefficient between the traction force and the passenger capacity, _μi represents the friction coefficient, that is, the resistance characteristics of the internal and external environment of the system, _kp represents the influence coefficient of the unit passenger capacity on the friction force increment, reflecting the additional friction caused by the increase in the number of passengers, _kd represents the damping coefficient linearly related to the elevator speed, and _kc ′ represents the damping coefficient related to the passenger capacity and speed.

As a preferred solution of the elevator cluster control method described in the present invention, the step of obtaining the passenger flow data of the elevator to be controlled in the target building in the previous cycle and establishing a passenger flow prediction model considering the system time delay by combining the machine learning algorithm and the recursive least squares method comprises:

First, a preliminary prediction model is established using a machine learning algorithm. The machine learning method using time series data is selected, and its prediction model is expressed as:

_Ppred (t+1)＝LSTM(P(t),P(t-1),...,P(tn))

Where P(t) is the passenger flow at time t, n is the length of the historical window, and LSTM is a trained long short-term memory network model;

Secondly, consider the system time lag factor. If the system has a time lag effect, the time lag variable is introduced to improve the model. The improved model is expressed as:

_Ppred (t+τ)＝F(LSTM(P(t),P(t-1),...,P(tn)),τ)

Among them, τ is the time-delay parameter of the system, and F is the function that corrects the prediction results by combining the time-delay information;

Again, the recursive least squares method is used to update the parameters:

Initialize model parameters θ;

For each time step t, the model parameters θ are updated according to the actual observed passenger flow data P _obs (t+τ):

Finally, the passenger flow prediction model is embedded into the nonlinear dynamic mathematical model.

As a preferred solution of the elevator cluster control method described in the present invention, wherein: the model predictive controller is combined with the nonlinear dynamic mathematical model and the passenger flow prediction model to predict the state and output at multiple future moments, and combined with the preset control objective function and constraint conditions, the optimal control action sequence is formulated, including:

Assuming that there are multiple elevators E ₁ , E ₂ , .. , _EM in the target building to be controlled, serving passenger requests on multiple floors F ₁ , F ₂ , .. , F _N , the preset control objective function is expressed as:

Among them, _w1i is the weight factor of elevator _Ei for the average waiting time of passengers, _w2i is the weight factor of elevator _Ei for energy consumption, represents the average waiting time of passengers when elevator E _i serves floor F _j at the tth moment in the prediction time domain, E _i (t) represents the energy consumed by elevator E _i at the tth moment in the prediction time domain, S _i (k) is the state vector of elevator E _i at the tth moment, and the state vector at least includes current position, speed, and passenger capacity information, U represents a penalty term, λ represents the penalty factor of the penalty term, and k represents the time step.

As a preferred solution of the elevator cluster control method described in the present invention, wherein: the state and output at multiple future moments are predicted by combining the model predictive controller with the nonlinear dynamic mathematical model and the passenger flow prediction model, and the optimal control action sequence is formulated in combination with the preset control objective function and constraint conditions, and further includes:

The constraints include:

To prevent consecutive stops on the same floor, establish a minimum stop floor interval constraint:

t _stop ≤ _ti,j -ti _,j-1 , j = 2, 3, ..., _Ni , i = 1, 2, ..., M

Where t _i,j represents the time when elevator E _i arrives at the jth floor;

Establish a response time constraint for a passenger summon request:

_tresponse ≤tj _,arrival -ti _,call i＝1,2,...,M

Establish system balance constraints:

Wherein, _Nf represents the total number of floors of the building, Pi _,j (k) is the service demand of elevator _Ei on the jth floor at time k, and the service demand includes at least the number of passengers and the number of service requests, _Pavg,j is the average service demand expected to be achieved by all elevators on the jth floor, and B is the set threshold used to control the imbalance degree of the system.

As a preferred solution of the elevator cluster control method of the present invention, the constraint conditions include:

Establish priority constraints: the elevator response priority of the lowest three floors and the highest three floors is higher than the response priority of any other floors;

To prevent multiple elevators from going to the same floor at the same time and causing congestion or empty travel, collaborative scheduling constraints are established:

Where _M1 represents the number of elevator devices participating in the collaborative scheduling, T _i,end (D _d ) is the time point when the ^iath elevator device completes its current task within the scheduling period D d, T _j,start (D _d +1) is the time when the ^jbth device starts a new task in the next scheduling period D _d +1, and _{D min is} the set minimum time interval threshold;

When multiple elevators are running in the same shaft, establish safe distance constraints:

in, Indicates elevator/> At the position at time k, /> Indicates elevator/> At the time k, _M2 represents the number of elevator floors in the same shaft when multiple elevators are running in the same shaft, and _hij (k) represents the safety distance compensation value calculated based on the elevator length and current position.

An elevator cluster control system, characterized by comprising:

The first model building module is used to obtain the physical characteristics of the elevator to be controlled in the target building in the previous cycle and the characteristics of the passenger's going up and down stairs behavior, and build a nonlinear dynamic mathematical model. The elevator control method used in the previous cycle of the elevator is any existing control method;

The second model building module is used to obtain the passenger flow data of the elevator to be controlled in the target building in the previous cycle, and to establish a passenger flow prediction model considering the system time lag by combining the machine learning algorithm and the recursive least squares method;

An optimal sequence acquisition module is used to predict the state and output at multiple future moments by combining the nonlinear dynamic mathematical model and the passenger flow prediction model through a model prediction controller, and formulate an optimal control action sequence in combination with a preset control objective function and constraint conditions;

The operation and update module is used to send specific control instructions to each elevator according to the calculated optimal control action sequence, so that the elevator can run according to the predetermined plan and complete the update of the elevator cluster control method.

A computer device comprises a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the steps of the above method when executing the computer program.

A computer-readable storage medium stores a computer program thereon, wherein the computer program implements the steps of the method described above when executed by a processor.

Beneficial effects of the present invention: The present invention proposes an elevator cluster control method and system, which obtains the physical characteristics of the elevator to be controlled in the target building in the previous cycle and the characteristics of the passenger's going up and down stairs behavior, and constructs a nonlinear dynamic mathematical model, wherein the elevator control method used in the previous cycle of the elevator is any existing control method; the passenger flow data in the previous cycle of the elevator to be controlled in the target building is obtained, and a passenger flow prediction model considering system time lag is established in combination with a machine learning algorithm and a recursive least squares method; the state and output at multiple future moments are predicted by combining the nonlinear dynamic mathematical model and the passenger flow prediction model through a model predictive controller, and an optimal control action sequence is formulated in combination with a preset control objective function and constraints; according to the calculated optimal control action sequence, specific control instructions are sent to each elevator to make the elevator run according to a predetermined plan, and the update of the elevator cluster control method is completed.

The elevator cluster control method and system provided by the present invention have the following advantages:

1. Improve elevator operation efficiency: By constructing a nonlinear dynamic mathematical model and a passenger flow prediction model, the optimal scheduling of elevator clusters can be achieved, the passenger waiting time can be reduced, and the elevator operation efficiency can be improved.

2. Improve passenger comfort: According to passenger flow forecast, reasonably allocate elevator passenger capacity to avoid elevator overload or empty load, and improve passenger comfort.

3. Enhance system stability: Through the model predictive controller, the elevator operation strategy is adjusted in real time to ensure the stable operation of the elevator cluster in complex environments.

4. Simplify operation and maintenance management: Through unified management and control of elevator clusters, the workload of operation and maintenance personnel can be simplified and the level of building management can be improved.

5. Strong adaptability: The elevator cluster control method and system proposed in the present invention can be applied to buildings of different types and sizes and have wide adaptability.

The present invention not only provides a new technical solution for elevator cluster control, but also provides strong support for building automation management and intelligent development. As an intelligent control system, the elevator cluster control method and system have broad application prospects and market potential. With the continuous upgrading of building construction and management in the future, the present invention will bring significant economic and social benefits to the elevator industry and building automation field.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

FIG1 is a method flow chart of an elevator cluster control method and system provided by an embodiment of the present invention;

FIG. 2 is an internal structure diagram of a computer device of an elevator cluster control method and system provided by an embodiment of the present invention.

Detailed ways

In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

Example 1

Referring to FIG. 1-2 , a first embodiment of the present invention is shown. The embodiment provides an elevator cluster control method and system, including an elevator cluster control method and an elevator cluster control system. The elevator cluster control method includes:

S101: Obtain the physical characteristics of the elevator to be controlled in the target building in the previous cycle and the characteristics of the passenger's going up and down stairs behavior, and build a nonlinear dynamic mathematical model. The elevator control method used in the previous cycle of the elevator is any existing control method;

Among them, the physical characteristics of the elevator to be controlled in the target building in the previous cycle and the characteristics of the passenger's upstairs and downstairs behavior are obtained to build a nonlinear dynamic mathematical model. The elevator control method used in the previous cycle of the elevator is any existing control method including:

Specifically, the physical characteristics of the elevator include at least the elevator floor z(t), the elevator speed v(t), the passenger capacity, and the passenger car movement C(t), where t represents the elevator operation time;

In an optional embodiment, the physical characteristics of the elevator may also include the number of elevator operations, the elevator operation time, the elevator load rate, etc.;

Furthermore, the passenger's up-and-down behavior characteristics at least include the entry and exit of passengers on each floor;

In an optional embodiment, the passenger's going up and down stairs behavior characteristics may also include the number of times the passenger goes up and down stairs, the time the passenger goes up and down stairs, the passenger's riding habits, etc.

It should be noted that the previous period is at least greater than one month. In the embodiment of the present application, the period is set to three months;

In the embodiment of the present application, in order to make up for the three-month lead time, a mathematical twin model is established to acquire simulation data. In actual applications, in order to obtain accurate data, the mathematical twin modeling operation can be omitted, and the physical characteristics of the elevator to be controlled in the target building in the previous cycle and the characteristics of the passenger's up and down behavior are directly acquired to construct a nonlinear dynamic mathematical model.

It should be noted that the physical characteristics of the elevator to be controlled in the target building in the previous cycle and the characteristics of the passengers' going-up-and-down-stairs behavior can be obtained by establishing a mathematical twin model under response conditions based on the selected existing control method, and the data of the previous cycle can be obtained through the mathematical twin model. This allows the updated elevator cluster control method to be used in a timely manner after the elevator is established in the target building.

It should be noted that the steps for establishing a mathematical twin model for elevator cluster control can be as follows:

Step 1: Collect elevator physical property data: Collect the operating status information of elevators in at least three buildings of the same type as the target building, with the same personnel distribution and personnel functions, including but not limited to the current floor, speed, direction, load, fault status, maintenance records and other real-time or historical data. For example, if the target building is a shopping mall and the building height plus the basement is eight floors, then collect the operating status information of elevators in at least three shopping malls in the same area with eight floors plus the basement.

Collect passenger behavior data: statistics on the frequency of passengers entering and exiting the elevator, waiting time, destination floor distribution, peak hours, etc. on each floor during different time periods. Even more accurate passenger flow and route selection can be obtained through smart card systems or monitoring systems.

Step 2: Clean the collected data and remove invalid or abnormal data points.

Perform preprocessing operations such as normalization, binning, or time series analysis on the data according to analysis requirements.

Step three: extract key indicators reflecting elevator performance and characteristic variables of passenger behavior patterns from the original data, such as elevator utilization rate, average response time, passenger density distribution, etc.

The time windows and relevant features required to build a passenger demand forecasting model.

Step 4: Establish a physical model of the elevator system to simulate the actual operation process and physical constraints of the elevator.

Design a passenger behavior model, considering the behavioral patterns of passengers in different scenarios and their impact on elevator scheduling.

Step five, understand and combine the selected existing elevator control strategies (such as shortest path priority, minimum number of transfers, load balance, etc.), convert these strategies into mathematical expressions, and incorporate them into the twin model as decision rules.

Step six: Use the real data from the previous cycle to drive the twin model, and repeatedly adjust the parameters and optimize the algorithm through the simulation platform so that the model can accurately reflect the behavior of the actual elevator system under given conditions.

Use machine learning or deep learning technology to further enhance the model’s prediction and decision-making capabilities in unknown situations.

Step seven: Apply the trained and optimized mathematical twin model to the elevator cluster control of the next cycle, update the model parameters in real time, and dynamically adjust the control strategy to adapt to the changing environment and needs.

Step 8: Use the mathematical twin model to simulate the data acquisition of the first three months.

It should be noted that establishing a digital twin model can greatly reduce the time for updating control methods, and can also comprehensively improve the intelligence level and operational efficiency of elevator cluster management.

Furthermore, the nonlinear dynamic mathematical model includes the elevator position change equation and the passenger capacity change equation:

Among them, the elevator position change equation is expressed as:

Among them, m is the total mass of the elevator and passengers, _Te is the traction force provided by the traction motor, which is related to the elevator position, speed and passenger capacity, _Ff is the friction force related to the elevator speed and passenger capacity, which increases with the increase of speed and passenger capacity, _Fd is the damping force considering the influence of personnel changes, which is related to the elevator speed and personnel distribution, u(t) represents the force of the current control signal acting on the elevator system, represents the acceleration of the elevator car;

Furthermore, the passenger capacity change equation is expressed as:

where _Pin (t,z(t)) and _Pout (t,z(t)) represent the number of passengers entering and leaving the elevator at floor z(t) at time t, respectively.

Furthermore, the physical characteristics of the elevator to be controlled in the target building in the previous cycle and the characteristics of the passenger's up-and-down behavior are obtained to construct a nonlinear dynamic mathematical model, and the elevator control method used in the previous cycle of the elevator is any existing control method, which also includes: estimating unknown parameters or function forms in the model based on actual data combined with system identification technology;

The traction force Te _is estimated based on the elevator traction characteristics and passenger capacity. The estimation formula is as follows:

_Te (z(t),v(t),C(t))＝ktz ₍ t)+ _cvv (t)+ _kcC (t)

The friction force F _f is estimated by the following formula:

F _i (z(t),v(t),C(t))＝μ _i (m ₁ +C(t)k _p )v(t)

The damping force _Fd is estimated using the following formula:

F _d (z(t),v(t),C(t)) = k _d v(t) + k _c ′C(t)v(t)

Wherein, _m1 represents the inherent mass of the elevator, _kt represents the correlation coefficient between the traction force and the elevator position, _cv represents the correlation coefficient between the traction force and the elevator speed, _kc represents the correlation coefficient between the traction force and the passenger capacity, _μi represents the friction coefficient, that is, the resistance characteristics of the internal and external environment of the system, _kp represents the influence coefficient of the unit passenger capacity on the friction force increment, reflecting the additional friction caused by the increase in the number of passengers, _kd represents the damping coefficient linearly related to the elevator speed, and _kc ′ represents the damping coefficient related to the passenger capacity and speed.

It should be noted that the physical characteristics of the elevator to be controlled in the target building in the previous cycle and the characteristics of the passengers' up and down behavior are obtained, and a nonlinear dynamic mathematical model is constructed to describe the physical behavior and operation law of the elevator system, providing a basis for the subsequent control strategy formulation. Considering the actual working characteristics of the elevator (such as speed, position, load, etc.) helps to accurately simulate the dynamic response of the elevator. Incorporating existing control methods into the model can evaluate the impact of the current control strategy on the performance of the elevator system and provide a basis for optimizing control.

S102: Obtain the passenger flow data of the elevator to be controlled in the target building in the previous cycle, and establish a passenger flow prediction model considering system time lag by combining machine learning algorithm and recursive least squares method;

Obtain the passenger flow data of the elevator to be controlled in the target building in the previous cycle, and combine the machine learning algorithm and the recursive least squares method to establish a passenger flow prediction model that takes into account the system time lag, including:

First, a preliminary prediction model is established using a machine learning algorithm. The machine learning method using time series data is selected, and its prediction model is expressed as:

_Ppred (t+1)＝LSTM(P(t),P(t-1),...,P(tn))

Where P(t) is the passenger flow at time t, n is the length of the historical window, and LSTM is a trained long short-term memory network model;

Secondly, consider the system time lag factor. If the system has a time lag effect, the time lag variable is introduced to improve the model. The improved model is expressed as:

_Ppred (t+τ)＝F(LSTM(P(t),P(t-1),...,P(tn)),τ)

Among them, τ is the time-delay parameter of the system, and F is the function that corrects the prediction results by combining the time-delay information;

Again, the recursive least squares method is used to update the parameters:

Initialize model parameters θ;

For each time step t, the model parameters θ are updated according to the actual observed passenger flow data P _obs (t+τ):

Finally, the passenger flow prediction model is embedded into the nonlinear dynamic mathematical model.

It should be noted that obtaining the passenger flow data of the elevator to be controlled in the target building in the previous cycle and combining the machine learning algorithm and the recursive least squares method to establish a passenger flow prediction model that takes into account the system time lag can predict the passenger entry and exit of each floor in the future time period, make elevator scheduling plans in advance, and reduce passenger waiting time. By combining the machine learning algorithm and the recursive least squares method to deal with the time lag problem, the accuracy and adaptability of the prediction model are improved. It has good prediction capabilities for traffic peaks and valleys in different time periods, and can effectively respond to passenger flow demand fluctuations in different time periods.

S103: using a model predictive controller in combination with a nonlinear dynamic mathematical model and a passenger flow prediction model, predicting the states and outputs at multiple future moments, and formulating an optimal control action sequence in combination with a preset control objective function and constraints;

The model predictive controller combines the nonlinear dynamic mathematical model and the passenger flow prediction model to predict the state and output at multiple future moments, and formulates the optimal control action sequence in combination with the preset control objective function and constraints, including:

Assuming that there are multiple elevators E ₁ , E ₂ , .. , _EM in the target building to be controlled, serving passenger requests on multiple floors F ₁ , F ₂ , .. , F _N , the preset control objective function is expressed as:

Among them, _w1i is the weight factor of elevator _Ei for the average waiting time of passengers, _w2i is the weight factor of elevator _Ei for energy consumption, represents the average waiting time of passengers when elevator E _i serves floor F _j at the tth moment in the prediction time domain, E _i (t) represents the energy consumed by elevator E _i at the tth moment in the prediction time domain, S _i (k) is the state vector of elevator E _i at the tth moment, and the state vector at least includes the current position, speed, and passenger capacity information, U represents the penalty term, λ represents the penalty factor of the penalty term, and k represents the time step.

The model predictive controller combines the nonlinear dynamic mathematical model and the passenger flow prediction model to predict the state and output at multiple future moments, and combines the preset control objective function and constraints to formulate the optimal control action sequence, which also includes:

The constraints include:

To prevent consecutive stops on the same floor, establish a minimum stop floor interval constraint:

t _stop ≤ _ti,j -ti _,j-1 , j = 2, 3, ..., _Ni , i = 1, 2, ..., M

Where t _i,j represents the time when elevator E _i arrives at the jth floor;

Establish a response time constraint for a passenger summon request:

_tresponse ≤tj _,arrival -ti _,call i＝1,2,...,M

Establish system balance constraints:

Where _Nf represents the total number of floors of the building, Pi _,j (k) is the service demand of elevator _Ei on the jth floor at time k, and the service demand includes at least the number of passengers and the number of service requests, _Pavg,j is the average service demand expected to be achieved by all elevators on the jth floor, and B is the set threshold used to control the imbalance of the system.

Constraints include:

Establish priority constraints: the elevator response priority of the lowest three floors and the highest three floors is higher than the response priority of any other floors;

To prevent multiple elevators from going to the same floor at the same time and causing congestion or empty travel, collaborative scheduling constraints are established:

Where _M1 represents the number of elevator devices participating in the collaborative scheduling, T _i,end (D _d ) is the time point when the ^iath elevator device completes its current task within the scheduling period D d, T _j,start (D _d +1) is the time when the ^jbth device starts a new task in the next scheduling period D _d +1, and _{D min is} the set minimum time interval threshold;

When multiple elevators are running in the same shaft, establish safe distance constraints:

in, Indicates elevator/> At the position at time k, /> Indicates elevator/> At the time k, _M2 represents the number of elevator floors in the same shaft when multiple elevators are running in the same shaft, and _hij (k) represents the safety distance compensation value calculated based on the elevator length and current position.

In an optional embodiment, the following constraints may also be considered:

Speed constraint: The maximum and minimum operating speeds of the elevator are limited:

v _min ≤v _i (k) ≤v _max , i＝1,2,...,m

Wherein, m represents the number of elevators, _vi (k) represents the speed of the i-th elevator at time k, _vmin represents the minimum speed threshold, and _vmax represents the maximum speed threshold.

Acceleration constraints: The maximum and minimum acceleration of the elevator are limited:

Wherein, m represents the number of elevators, a _min represents the minimum value of the acceleration threshold, and a _max represents the maximum value of the acceleration threshold.

Passenger capacity limit, a single elevator cannot be overloaded:

C _min ≤C _i (k) ≤C _max , i＝1,2,…,m

Where _Ci (k) represents the passenger capacity of the ith elevator at time k, C _min represents the minimum passenger capacity, and C _max represents the maximum passenger capacity;

Emergency handling restrictions:

Assume that there is a Boolean variable EB(t) to indicate whether an emergency occurs. If an emergency occurs, EB(t) = true = 1. In an emergency, the elevator's target floor is forcibly set to the safe floor G _safe .

in, represents the target floor of the i-th elevator at time t,/> Indicates the target floor calculated under the normal scheduling strategy.

Speed and acceleration limits. In an emergency, the maximum permissible speed and acceleration of the elevator can be expressed as:

in, and/> are the maximum speed and acceleration of the i-th elevator at time t, ν _emergency and a _emergency are the limit values under emergency conditions, and v _normal and a _normal are the limit values under normal conditions.

It should be noted that the model predictive controller combines the nonlinear dynamic model with the passenger flow prediction model to achieve the prediction of the state and output at multiple moments in the future. Combined with the preset control objective function and constraints, the optimal control action sequence is formulated. The overall operating efficiency and service quality of the cluster elevator system are improved, so that the elevator can meet the real-time passenger flow needs while taking into account energy saving and equipment utilization.

S104: According to the calculated optimal control action sequence, specific control instructions are sent to each elevator to make the elevator run according to the predetermined plan and complete the update of the elevator cluster control method.

Assume that the optimal control action sequence obtained by model predictive control (MPC) is:

{u ^* (k|t),u ^* (k+1|t),...,u ^* (k+N _p |t)}

Among them, k represents the current time, t represents the planned start time, N _ρ is the number of steps in the prediction time domain, and u is the control vector, which may contain control variables such as elevator speed, acceleration, and target floor.

For each future time k+i (i=0, 1, ..., N _p -1), the corresponding elevator command is extracted from the optimal control action:

Target floor instruction: If u ^* (k+i|t) contains the target floor information f _target , a command is sent to the elevator to go to that floor.

Operation mode instructions: If the model takes into account the elevator operation mode such as acceleration, deceleration, stop waiting, etc., instructions are issued according to the corresponding speed and acceleration setting values.

According to the real time t and the preset time interval, the corresponding control instructions are sent to each elevator control system at the correct time point. For example, for the instruction of step i, it will be sent at any t+i·Δt, where Δt is the prediction step length.

After receiving the command, the actual elevator system will execute it and provide status feedback. The subsequent control strategy is adjusted according to the feedback results, thus completing the online update of the elevator cluster control method.

In an optional embodiment, the condition for completing the online update can be determined as satisfying any two or more conditions as follows:

1. Command execution confirmation: A sign of update completion is that all elevators have received and started to execute the commands in the corresponding optimal control action sequence.

2. System state convergence: The effectiveness of the update can be determined by monitoring whether system performance indicators (such as passenger waiting time, elevator utilization, etc.) have reached the preset target or are gradually approaching the optimized ideal value. For example, if the average waiting time of the system is continuously lower than a certain threshold for several consecutive time periods, the update can be considered complete.

3. Real-time feedback and iterative adjustment: In practical applications, due to environmental changes and model prediction errors, the elevator cluster control system needs to continuously receive real-time feedback and make fine adjustments based on the feedback results. Therefore, the update completion condition may be to set a cycle number N or a certain performance improvement indicator to reach a stable standard, as shown below:

|P(t)-P*|<, t＝k,…,k+N

Among them, P(t) is the actual performance indicator at time t (such as the average waiting time of passengers), and P ^* is the optimization target value, which is a very small positive number and represents the allowable deviation range.

4. No constraint violations occur: Ensure that during the entire update process, all elevator operations do not violate the pre-set physical constraints, such as speed limits, passenger capacity limits, etc.

In an optional embodiment, the actual operating status of the elevator and the passenger flow are monitored in real time, actual data is collected and compared with the predicted results, and the parameters of the prediction model and optimization algorithm are continuously corrected according to the deviation, so that the scheduling strategy can be continuously improved and adapted to the actual situation, thereby improving the accuracy and robustness of the scheduling decision.

It should be noted that according to the calculated optimal control action sequence, specific instructions are sent to each elevator to ensure that the elevator runs efficiently and orderly according to the plan. The real-time optimization and adjustment of the elevator group control system is realized, and the control strategy is dynamically updated according to the actual operation situation and prediction results, which enhances the robustness and adaptability of the system. In the process of continuous cycle execution and feedback, the elevator cluster control system always maintains the best working state and continuously improves the service level of vertical transportation inside the building.

In summary, the present invention proposes an elevator cluster control method, which obtains the physical characteristics of the elevator to be controlled in the target building in the previous cycle and the characteristics of the passenger's going up and down stairs behavior, and constructs a nonlinear dynamic mathematical model, wherein the elevator control method used in the previous cycle of the elevator is any existing control method; the passenger flow data of the elevator to be controlled in the target building in the previous cycle is obtained, and a passenger flow prediction model considering the system time lag is established in combination with a machine learning algorithm and a recursive least squares method; the state and output at multiple future moments are predicted by combining the nonlinear dynamic mathematical model and the passenger flow prediction model through a model predictive controller, and an optimal control action sequence is formulated in combination with a preset control objective function and constraints; according to the calculated optimal control action sequence, specific control instructions are sent to each elevator to make the elevator run according to a predetermined plan, and the update of the elevator cluster control method is completed.

The elevator cluster control method and system provided by the present invention have the following advantages:

1. Improve elevator operation efficiency: By constructing a nonlinear dynamic mathematical model and a passenger flow prediction model, the optimal scheduling of elevator clusters can be achieved, the passenger waiting time can be reduced, and the elevator operation efficiency can be improved.

2. Improve passenger comfort: According to passenger flow forecast, reasonably allocate elevator passenger capacity to avoid elevator overload or empty load, and improve passenger comfort.

3. Enhance system stability: Through the model predictive controller, the elevator operation strategy is adjusted in real time to ensure the stable operation of the elevator cluster in complex environments.

4. Simplify operation and maintenance management: Through unified management and control of elevator clusters, the workload of operation and maintenance personnel can be simplified and the level of building management can be improved.

5. Strong adaptability: The elevator cluster control method and system proposed in the present invention can be applied to buildings of different types and sizes and have wide adaptability.

The present invention not only provides a new technical solution for elevator cluster control, but also provides strong support for building automation management and intelligent development. As an intelligent control system, the elevator cluster control method and system have broad application prospects and market potential. With the continuous upgrading of building construction and management in the future, the present invention will bring significant economic and social benefits to the elevator industry and building automation field.

In a preferred embodiment, an elevator cluster control system includes:

The first model building module is used to obtain the physical characteristics of the elevator to be controlled in the target building in the previous cycle and the characteristics of the passenger's upstairs and downstairs behavior, and to build a nonlinear dynamic mathematical model. The elevator control method used in the previous cycle of the elevator is any existing control method;

The second model building module is used to obtain the passenger flow data of the elevator to be controlled in the target building in the previous cycle, and to establish a passenger flow prediction model considering the system time lag by combining the machine learning algorithm and the recursive least squares method;

The optimal sequence acquisition module is used to predict the state and output at multiple future moments by combining the model predictive controller with the nonlinear dynamic mathematical model and the passenger flow prediction model, and formulate the optimal control action sequence in combination with the preset control objective function and constraint conditions;

The operation and update module is used to send specific control instructions to each elevator according to the calculated optimal control action sequence, so that the elevator can run according to the predetermined plan and complete the update of the elevator cluster control method.

The above-mentioned unit modules may be embedded in or independent of the processor in the computer device in the form of hardware, or may be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be shown in FIG2. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be implemented through WIFI, an operator network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, an elevator cluster control method is implemented. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

The physical characteristics of the elevator to be controlled in the target building in the previous cycle and the characteristics of the passengers' upstairs and downstairs behavior are obtained to construct a nonlinear dynamic mathematical model. The elevator control method used in the previous cycle of the elevator is any existing control method.

Obtain the passenger flow data of the elevator to be controlled in the target building in the previous cycle, and establish a passenger flow prediction model that takes into account system time lag by combining machine learning algorithm and recursive least squares method;

The model predictive controller combines the nonlinear dynamic mathematical model and the passenger flow prediction model to predict the state and output at multiple future moments, and formulates the optimal control action sequence in combination with the preset control objective function and constraints;

According to the calculated optimal control action sequence, specific control instructions are sent to each elevator to make the elevator run according to the predetermined plan and complete the update of the elevator cluster control method.

It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal translation scripting language JavaScript, etc.

The present application is described with reference to the flowchart and/or block diagram of the method, device (system) and computer program product according to the embodiment of the present application. It should be understood that each process and/or box in the flowchart and/or block diagram, and the combination of the process and/or box in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for realizing the function specified in one process or multiple processes in the flowchart and/or one box or multiple boxes in the block diagram.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device that implements the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims (10)

1. An elevator cluster control method, comprising:

Acquiring the physical characteristics of an elevator in a previous period of the elevator to be controlled in a target building and the behavior characteristics of passengers going upstairs and downstairs, and constructing a nonlinear dynamic mathematical model, wherein the elevator control mode used in the previous period of the elevator is any existing control method;

acquiring passenger flow data in a previous period of an elevator to be controlled in a target building, and establishing a passenger flow prediction model considering system time lag by combining a machine learning algorithm and a recursive least square method;

predicting states and outputs of a plurality of future moments by combining the nonlinear dynamic mathematical model and the passenger flow prediction model through a model prediction controller, and formulating an optimal control action sequence by combining a preset control objective function and constraint conditions;

and according to the calculated optimal control action sequence, a specific control instruction is sent to each elevator, so that the elevator operates according to a preset plan, and the updating of the elevator cluster control method is completed.

2. The elevator cluster control method according to claim 1, wherein the acquiring the physical characteristics of the elevator and the behavior characteristics of passengers going up and down the elevator in the previous period of the elevator to be controlled in the target building constructs a nonlinear dynamic mathematical model, and the elevator control mode used in the previous period of the elevator is any existing control method comprising:

the elevator physical characteristics at least comprise an elevator floor z (t), an elevator speed v (t), a passenger capacity and a passenger car movement condition C (t), wherein t represents the running time of the elevator;

The behavior characteristics of passengers going upstairs and downstairs at least comprise the passengers going in and out of each floor;

The nonlinear dynamic mathematical model comprises a position change equation and a passenger capacity change equation of the elevator:

The equation of change of position of the elevator is expressed as:

Where m is the total mass of the elevator and passengers, T _e is the traction force provided by the traction motor, which is related to the elevator position, speed and passenger capacity, F _f is the friction force related to the elevator speed and passenger capacity, which increases with increasing speed and passenger capacity, F _d is the damping force taking into account the influence of personnel variations, which is related to the elevator speed and personnel distribution, u (T) represents the force of the current control signal acting on the elevator system, Indicating the acceleration of the elevator car;

the passenger capacity change equation is expressed as:

Wherein P _in (t, z (t)) and P _out (t, z (t)) represent the number of passengers entering and leaving the elevator at floor z (t) at time t, respectively.

3. The elevator cluster control method according to claim 2, wherein the acquiring the physical characteristics of the elevator and the behavior characteristics of passengers going up and down the elevator in the previous period of the elevator to be controlled in the target building constructs a nonlinear dynamic mathematical model, and the elevator control mode used in the previous period of the elevator is any existing control method, and further comprises: estimating unknown parameters or function forms in the model according to actual data and a system identification technology;

the traction force T _e is estimated according to the traction characteristic and the passenger capacity of the elevator, and the estimation formula is as follows:

T_e(z(t),v(t),C(t))＝k_tz(t)+c_vv(t)+k_cC(t)

the friction force F _f is estimated as follows:

F_i(z(t),v(t),C(t))＝μ_i(m₁+C(t)k_p)v(t)

The damping force F _d is estimated as follows:

F_d(z(t),v(t),C(t))＝k_dv(t)+k_c′C(t)v(t)

wherein m ₁ represents the intrinsic mass of the elevator, k _t represents the coefficient of correlation of the traction force and the elevator position, c _v represents the coefficient of correlation of the traction force and the elevator speed, k _c represents the coefficient of correlation of the traction force and the passenger capacity, mu _i represents the coefficient of friction, namely the resistance characteristics of the inside and the outside environment of the system, k _p represents the coefficient of influence of the unit passenger capacity on the increase of the friction force, the additional friction force caused by the increase of the number of passengers is reflected, k _d represents the damping coefficient linearly related to the elevator speed, and k _c' represents the damping coefficient related to the passenger capacity and the speed.

4. The elevator cluster control method of claim 3, wherein the acquiring the passenger flow data in the previous cycle of the elevator to be controlled in the target building and building the passenger flow prediction model considering the system time lag in combination with the machine learning algorithm and the recursive least square method comprises:

First, a preliminary prediction model is established using a machine learning algorithm, and a machine learning method using time series data is selected, wherein the prediction model is expressed as:

P_pred(t+1)＝LSTM(P(t),P(t-1),...,P(t-n))

Wherein P (t) is the passenger flow at time t, n is the history window length, LSTM is a trained long-short term memory network model;

secondly, considering a system time lag factor, if a time lag effect exists in the system, introducing a time lag variable improvement model, wherein the improved model is expressed as:

P_pred(t+τ)＝F(LSTM(P(t),P(t-1),...,P(t-n)),τ)

Wherein τ is a time lag parameter of the system, and F is a function of correcting a prediction result by combining time lag information;

Again, parameter updates are performed using the recursive least squares method:

Initializing a model parameter theta;

For each time step t, the model parameters θ are updated according to the actual observed passenger flow data P _obs (t+τ):

finally, the passenger flow prediction model is embedded into a nonlinear dynamic mathematical model.

5. The elevator cluster control method of claim 4, wherein the predicting states and outputs at a plurality of future times by the model predictive controller in combination with the nonlinear dynamic mathematical model and the passenger flow predictive model, and formulating an optimal control action sequence in combination with a preset control objective function and constraint conditions comprises:

Assuming that there are a plurality of elevators E ₁,E₂,..,E_M in the target building to be controlled in which the passenger requests to serve a plurality of floors F ₁,F₂,...,F_N, the preset control objective function is expressed as:

Where w _1i is the weighting factor of elevator E _i for the average waiting time of the passengers, w _2i is the weighting factor of elevator E _i for the energy consumption, Representing the average waiting time of passengers at elevator E _i to serve floor F _j in the prediction horizon, E _i (t) representing the energy consumed by elevator E _i at time t in the prediction horizon, S _i (k) being the state vector of elevator E _i at time t, said state vector comprising at least the current position, speed, passenger capacity information, U representing a penalty term, λ representing a penalty factor of the penalty term, k representing a time step.

6. The elevator cluster control method of claim 5, wherein the predicting states and outputs at a plurality of future times by the model predictive controller in combination with the nonlinear dynamic mathematical model and the passenger flow predictive model, and formulating an optimal control action sequence in combination with a preset control objective function and constraints, further comprises:

The constraint conditions include:

preventing continuous stops at the same floor, and establishing minimum stop floor interval constraint:

t_stop≤t_i,j-t_i,j-1,j＝2,3,...,N_i，i＝1,2,..,M

Wherein t _i,j represents the time when elevator E _i reaches the j-th floor;

establishing a response time constraint for the passenger summoning request:

t_response≤t_j,arrival-t_i,call i＝1,2,...,M

establishing a system balance constraint:

Where N _f denotes the total floor number of the building, P _i,j (k) is the service demand of the elevator E _i at the j-th floor at time k, said service demand comprising at least the number of passengers, the number of service requests, P _avg,j is the average service demand all elevators are expected to reach at the j-th floor, and B is the set threshold for controlling the degree of imbalance of the system.

7. The elevator cluster control method of claim 6, wherein the constraints include further comprising:

establishing a priority constraint: the response priority of the elevator of the lowest three floors and the highest three floors is higher than the response priority of any other floors;

The method comprises the steps of preventing a plurality of elevators from simultaneously traveling to the same floor to cause congestion or idle traveling, and establishing cooperative scheduling constraint:

Wherein M ₁ represents the number of elevator devices participating in cooperative scheduling, T _i,end(D_d) is the time point when the i ^a th elevator device completes its current task in the scheduling period D _d, T _j,start(D_d +1) is the time when the j ^b th device starts a new task in the following scheduling period D _d +1, and D _min is the set minimum time interval threshold;

When multiple elevators are running in the same shaft, a safe distance constraint is established:

Wherein, Representing elevator/>Location at time k,/>Representing elevator/>At the position of time k, M ₂ represents the number of elevator floors in the same shaft when a plurality of elevators are traveling in the same shaft, and h _ij (k) represents the safety distance compensation value calculated from the elevator length and the current position.

8. An elevator cluster control system, comprising:

The first model building module is used for obtaining the physical characteristics of the elevator and the behavior characteristics of passengers going upstairs and downstairs in the previous period of the elevator to be controlled in the target building and building a nonlinear dynamic mathematical model, wherein the elevator control mode used in the previous period of the elevator is any existing control method;

The second model building module is used for acquiring passenger flow data in a previous period of the elevator to be controlled in the target building, and building a passenger flow prediction model considering system time lag by combining a machine learning algorithm and a recursive least square method;

the optimal sequence acquisition module is used for predicting states and outputs of a plurality of future moments by combining the nonlinear dynamic mathematical model and the passenger flow prediction model through the model prediction controller, and formulating an optimal control action sequence by combining a preset control objective function and constraint conditions;

And the running and updating module is used for sending specific control instructions to each elevator according to the calculated optimal control action sequence, so that the elevator runs according to a preset plan and the updating of the elevator cluster control method is completed.

9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 7 when the computer program is executed.

10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 7.