Search

CN-122009925-A - Multi-scene self-adaptive elevator energy-saving group control system and control method

CN122009925ACN 122009925 ACN122009925 ACN 122009925ACN-122009925-A

Abstract

The application relates to a multi-scene self-adaptive elevator energy-saving group control system and a control method, wherein a control unit is arranged in a management platform, the data acquisition unit, an elevator group unit, an elevator taking interaction unit and an energy storage regulation and control unit are connected to the management platform to form a group control system, a scheduling scheme instruction is generated through the management platform according to the acquired data of the elevator group unit and the data acquisition unit, the control unit issues an execution instruction to carry out multi-objective scheduling, and meanwhile, operation execution data is stored in the management platform for subsequent training optimization, and meanwhile, in the multi-objective scheduling process, the energy consumption data of the energy storage regulation and control unit is acquired, the scheduling scheme is iterated by combining with real-time state monitoring, a new scheduling scheme instruction is regenerated, so that cross-group energy storage regulation and control is realized, and in the period, multi-level fault tolerance processing and self-adaptive capacity of a reinforced learning lifting system are carried out by combining with real-time state monitoring data.

Inventors

  • YU JIAXIN
  • LIU DONG
  • CHEN HAIWEN
  • WEN KE

Assignees

  • 日立楼宇技术（广州）有限公司

Dates

Publication Date
20260512
Application Date
20260228

Claims (10)

  1. 1. A multi-scenario adaptive elevator energy-efficient group control system, comprising: The elevator group unit is used for realizing vertical transportation of personnel and feeding back running states including speed, position and door opening and closing states in real time; The data acquisition unit is used for providing multi-source data input, acquiring personnel information, energy consumption data and waiting number of the elevator; the control unit is used for executing scheduling, energy storage, video analysis and fault tolerance and overall instruction issuing; the energy storage regulation and control unit is used for monitoring and recovering braking electric energy, realizing cross-group dynamic allocation and guaranteeing high-efficiency utilization of the electric energy; The management platform comprises a cloud end and/or a local server, a visual interface and a data storage module, wherein the management platform is used for parameter configuration, state monitoring, historical data management and alarm prompt; The energy-saving energy storage control system comprises a management platform, a data acquisition unit, an elevator group unit, a elevator taking interaction unit, an energy storage regulation and control unit, a control unit, an energy storage regulation and control unit, a control scheme instruction, an energy storage scheme instruction, a power consumption data, a real-time state monitoring and control unit, wherein the data acquisition unit, the elevator group unit, the elevator taking interaction unit and the energy storage regulation and control unit are arranged in the management platform, the energy storage regulation and control unit is connected to the management platform to form a group control system, the management platform is used for generating a scheduling scheme instruction according to the acquired data of the elevator group unit and the data acquisition unit, the control unit is used for issuing an execution instruction to carry out multi-target scheduling, meanwhile, the operation execution data is stored in the management platform for subsequent training and optimization, and meanwhile, the energy consumption data of the energy storage regulation and control unit are acquired in the multi-target scheduling process, the energy storage regulation and control unit is used for iterating a scheduling scheme by combining real-time state monitoring, and regenerating a new scheduling scheme instruction, so that energy storage regulation and control is realized, and energy-saving group control is finally realized.
  2. 2. The multi-scenario adaptive elevator energy-saving group control system of claim 1, wherein when multi-objective scheduling is performed, scheduling scheme composite score evaluation is performed using the following formula: (1); And issuing a scheduling instruction to the corresponding elevator and the elevator taking guide module according to the scoring scheme, and determining the stopping floors, the sequence and the door opening time.
  3. 3. The multi-scene adaptive elevator energy-saving group control system according to claim 2, wherein the specific flow of the multi-objective scheduling is as follows: A1, preprocessing data; A2, generating an initial scheme; a3, optimizing dynamic weights; a4, calculating scheme scores; A5, scheme screening and executing; a6, dynamically monitoring and iterating; During data preprocessing, acquiring elevator running states, elevator waiting numbers, energy consumption data and energy storage surplus through edge computing nodes, removing abnormal data by adopting a 3 sigma criterion, carrying out normalization processing on effective data, unifying dimensions, and transmitting the effective data to a main controller; when the initial scheme is generated, a genetic algorithm is adopted to generate an initial scheduling scheme library, an adaptability function is defined, a high-quality scheme is screened by a roulette selection method, a crossover operation and a mutation operation are executed, and a diversified scheme library is generated by iterating 20 times; When optimizing dynamic weight, adopting LSTM model to input history data of about 10 minutes and outputting real-time weight Performing scene adaptation adjustment, weight verification; In the calculation of the scheme score, a base index is calculated for each scheme in the scheme library Substituting the congestion degree penalty C and the start-stop frequency penalty D into a core formula (1) to obtain a comprehensive score f of each scheme; During scheme screening and execution, the longest waiting time is eliminated for >90 seconds, and the number of live persons is > The scheme of (2); sorting according to the ascending order of the scores f, and selecting the scheme with the lowest score; if the same-division scheme exists, the scheme with the highest energy storage utilization rate is preferentially selected, and a scheduling instruction is issued to the corresponding elevator and the elevator taking guide module, so that the floor stop, the sequence and the door opening time length are defined; and (3) during dynamic monitoring and iteration, monitoring the elevator taking state in real time, returning to the step (A2) to regenerate the scheme if the state variation is more than 10%, storing the running data into a historical library after the scheduling is completed, and performing offline training and optimization by using an LSTM model.
  4. 4. The multi-scenario adaptive elevator energy-saving group control system of claim 1, wherein the priority calculation is performed by the following formula when performing cross-group energy storage regulation: (2); loss calculation was performed by the following formula: (3); And performing cross-group energy storage regulation and control according to the priority and the loss.
  5. 5. The multi-scenario adaptive elevator energy-saving group control system of claim 4, wherein the cross-group energy storage regulation and control flow is as follows: b1, data acquisition and demand prediction; B2, calculating priority; B3, judging and matching the energy storage state; B4, calculating transmission loss; B5, electric energy allocation is carried out; B6, updating the state and utilizing the state in a gradient manner; The method comprises the steps of collecting real-time data of energy storage cabinets of all groups, elevator running states and environment temperature t during data collection and demand prediction, inputting energy storage data of approximately 24 hours in the same time period, passenger flow prediction of 1 hour in the future and elevator maintenance plan by a GRU model, and outputting the energy storage demand of 1 hour in the future of all groups by the GRU model Marking a group of demand gaps; In priority calculation, determining base priority of each group The priority of emergency elevator is automatically set up in emergency scene, and the energy storage remaining quantity ratio is calculated Substituting the predicted increase rate q of the passenger flow into a priority formula (2) to obtain the final priority P of each group, and sorting the groups according to descending order of P; When the energy storage states are judged and matched, judging the states of all the energy storage cabinets, wherein the residual electric quantity is more than 80 percent and is in a surplus state, 20 to 80 percent is in a self-sufficient state, and 20 percent is in a shortage state; during transmission loss calculation, the transmission distance between surplus and shortage groups is read, the ambient temperature t is obtained, and a temperature correction factor is calculated Calculating transmission loss L according to the loss correction model, and if L is more than 10% of the transmission electric quantity, replacing a matching object; When the electric energy allocation is executed, determining that the single conveying capacity is not more than 50% of the residual capacity of the surplus energy storage cabinets and the conveying capacity of a gap required by a shortage group is met; And when the capacity of the energy storage cabinet is attenuated to be below 80%, the energy storage cabinet is automatically switched to an emergency standby mode, and the energy storage cabinet is started only in a main energy storage cabinet fault or emergency scene.
  6. 6. The multi-scenario adaptive elevator energy-saving group control system of claim 4, wherein when multi-level fault tolerance and reinforcement learning is performed, the main controller sets three levels of fault tolerance through a state space, an action space and a reward function by means of standby fault- > mode switching, algorithm fault- > standby algorithm, communication fault- > edge takeover: (4) And (5) performing self-adaptive optimization.
  7. 7. The multi-scenario adaptive elevator energy-saving group control system of claim 6, wherein the multi-level fault tolerance and reinforcement learning comprises the steps of: C1, state monitoring and fault identification; C2, primary fault-tolerant processing; c3, secondary fault-tolerant processing; C4, three-level fault-tolerant processing; C5, shielding treatment and verification; the system monitors the equipment state, the system running state and the communication link in real time when the state is monitored and the fault is identified, wherein the equipment does not respond for more than 3 seconds to judge the equipment as a fault; When the camera fails in the primary fault-tolerant processing, the camera is automatically switched to a pressure button and WiFi positioning mode, and the number of people is estimated through the WiFi signal intensity of the mobile phone of the passenger; when the energy storage cabinet fails, isolating the failure energy storage cabinet, adjusting the distribution priority of other energy storage cabinets in the area, and guaranteeing the basic operation of the elevator; When the primary dispatching fails during the secondary fault-tolerant processing, the primary dispatching is automatically switched to the standby dispatching, and idle elevators closest to the primary dispatching are preferentially dispatched in the core rule to preferentially meet the demand of the passenger dense floors; When the communication between the main controller and the management platform is interrupted during the three-level fault-tolerant processing, the edge computing node takes over the local scheduling and executes the processing based on the parameters of the local cache, and after the communication is recovered, the operation data during the interruption are automatically synchronized to the management platform to ensure the data integrity; During shielding processing and verification, a state space S is defined, namely, elevator running state, passenger flow distribution, energy storage residual quantity, temperature and humidity environment parameters are defined, an action space A is defined, namely, scheduling scheme selection, energy storage allocation strategy and weight adjustment are defined, and a reward function is defined: (5) And updating an action value function Q (S, A) by adopting a Q-learning algorithm, wherein the learning rate alpha=0.1, the discount factor gamma=0.9, and optimizing an action strategy through continuous interactive learning so as to improve the self-adaptive capacity of the system.
  8. 8. The multi-scene adaptive elevator energy saving group control system of claim 4 further comprising video analysis fused with passenger number, identifying special crowd using YOLOv lightweight model, calculating passenger number by the following formula, (6); (7); And updating and fusing the final passenger number in real time through Kalman filtering.
  9. 9. The multi-scene adaptive elevator energy saving group control system of claim 8, wherein the video analysis and passenger number fusion comprises the steps of: D1, video acquisition and preprocessing; d2, behavior recognition and personnel counting; D3, multi-source data acquisition; D4, weight updating and fusion calculation; D5, shielding treatment and verification; During video acquisition and preprocessing, video data are synchronously acquired by a plurality of cameras in a hall and a car; the method comprises the steps of carrying out noise reduction and brightness enhancement treatment on video frames, improving recognition accuracy, splicing multi-view video by adopting perspective transformation, and eliminating shooting blind areas; During behavior recognition and personnel counting, a YOLOv lightweight model is loaded, personnel in a video frame are detected, coordinates and categories of personnel frame selection are output, a Hungary algorithm is adopted to track personnel tracks, the same personnel in adjacent frames are associated, repeated counting is avoided, and the total number of personnel in a waiting area is counted Marking the number and the positions of special crowds, and calculating the video identification confidence; When multi-source data are collected, the pressure sensitive buttons collect the passenger pressing data and convert the passenger pressing data into the number of passengers according to rules And calling historical contemporaneous elevator waiting number data, and predicting the current number of passengers by a moving average method ; Initializing Kalman filtering parameters during weight updating and fusion calculation, wherein the estimated error covariance P=0.1, the observation matrix H=1 and the observation noise covariance R=0.05, calculating the filtering gain K, and updating the weight of each data source If the confidence coefficient of a certain data source is less than 60%, reducing the weight of the data source to be less than or equal to 0.1, preferentially distributing the weight to the data with high confidence coefficient, substituting the data into a fusion formula, calculating the final number N of passengers, rounding and outputting; during the shielding processing and the checking, if the shielding area is detected, the shielding area is corrected by track compensation Counting the number of people entering/leaving the waiting area, supplementing the count of the shielding area, comparing the count of the video Counting with pressure button If the deviation is more than 30%, triggering secondary acquisition to perform cross check, and marking the dispatching priority coefficient when the special crowd is identified Synchronous transmissions are sent to the multi-target schedule.
  10. 10. A control method of a multi-scene self-adaptive elevator energy-saving group control system, characterized by being applied to the multi-scene self-adaptive elevator energy-saving group control system as claimed in any one of claims 1 to 9, comprising the following steps: s1, authority and behavior linkage control; s2, taking a ladder to register; s3, an emergency scheduling mechanism; S4, data privacy and security protection; The method comprises the steps of controlling authority and behaviors in a linkage way, supporting four-level authority, automatically matching personnel authority levels through video recognition, and limiting illegal floor access, adjusting scheduling priority for special crowds by combining behavior recognition results, wherein the waiting time weight of the special crowds is reduced by 30%, and the elevator door opening time is prolonged to 8-10 seconds; In the elevator taking registration process, a passenger initiates an elevator taking request through crossing an elevator or manually touching an outbound button; the passenger can correct the destination layer through the touch screen, finish registering after confirming, take advantage of ladder guiding module to display appointed elevator number, stop position and estimated arrival time synchronously, voice broadcast guiding information; In the emergency dispatching mechanism, when an elevator fault is detected, the operation of the fault elevator is automatically paused, fault equipment is isolated, the most recent idle elevator receiving task is dispatched, and meanwhile, fault warning and position information are sent to a management platform; Data privacy and security protection, and strict adherence to data privacy and security specifications in the processes of data acquisition, transmission and processing, wherein specific measures comprise: s41, anonymizing video data, wherein a video analysis module only performs personnel detection and counting at the edge end, does not store or upload original video images, and adopts one-way hash encryption processing for personnel identity information; S42, the WiFi positioning data are desensitized, the WiFi signal intensity data are only used for estimating the number of people and are not related to the MAC address of specific user equipment, and all the positioning data are processed in a memory in real time and destroyed immediately after being completed; S43, authority and access control, the management platform supports multi-level authority management, unauthorized access is forbidden, and all data transmission adopts AES-256 encryption to ensure communication security.

Description

Multi-scene self-adaptive elevator energy-saving group control system and control method Technical Field The application relates to the technical field of energy conservation, in particular to a multi-scene self-adaptive elevator energy-saving group control system and a control method. Background With the popularization of green building concepts, the energy-saving transformation requirements of elevators serving as building core energy-using equipment are increasingly urgent. The prior art realizes a certain energy saving effect through means such as clean energy substitution, simple energy storage device application, basic AI video analysis and the like, but has the obvious defects that a dynamic balance mechanism of energy saving and user comfort is lacked, a scheduling strategy is solidified to cause poor scene adaptability, an energy storage system and elevator group control are not in cooperation, the electric energy utilization rate is low, the cross-group allocation logic is fuzzy, an effective fault tolerance scheme does not exist when equipment fails, the system stability is not enough, a key algorithm and technical parameters are lacked, the landing performance and the replicability are poor, and finally the energy saving transformation investment return rate is low, and the effect is not expected. Disclosure of Invention In order to solve or partially solve the problems in the related art, the application provides a multi-scene self-adaptive elevator energy-saving group control system and a control method, which are suitable for multi-type building scenes such as office buildings, houses, hotels and the like with the number of elevators being more than or equal to 2, and realize the dynamic balance of energy-saving effect and user comfort. The application discloses a multi-scene self-adaptive elevator energy-saving group control system, which comprises: The elevator group unit comprises not less than two elevators, a load sensor and an operation state monitoring module, wherein the operation state monitoring module is used for realizing vertical transportation of personnel and feeding back operation states including speed, position and door opening and closing states in real time; The data acquisition unit comprises a high-definition camera, an electric energy meter and a pressure sensitive button, and is used for providing multi-source data input, acquiring personnel information, energy consumption data and waiting number of people; the elevator taking interaction unit comprises a touch screen and a voice broadcasting module and is used for providing destination floor registration, preregistration correction and elevator stopping guidance; The control unit comprises an industrial-grade main controller and is used for executing scheduling, energy storage, video analysis and fault tolerance and overall instruction issuing; The energy storage regulation and control unit comprises an energy storage cabinet, an energy consumption monitoring sensor and a copper bar connecting wire, and is used for recovering braking electric energy, realizing cross-group dynamic allocation and guaranteeing high-efficiency utilization of the electric energy; The management platform comprises a cloud end and/or a local server, a visual interface and a data storage module, wherein the management platform is used for parameter configuration, state monitoring, historical data management and alarm prompt; The energy-saving energy storage control system comprises a management platform, a data acquisition unit, an elevator group unit, a elevator taking interaction unit, an energy storage regulation and control unit, a control unit, an energy storage regulation and control unit, a control scheme instruction, an energy storage scheme instruction, a power consumption data, a real-time state monitoring and control unit, wherein the data acquisition unit, the elevator group unit, the elevator taking interaction unit and the energy storage regulation and control unit are arranged in the management platform, the energy storage regulation and control unit is connected to the management platform to form a group control system, the management platform is used for generating a scheduling scheme instruction according to the acquired data of the elevator group unit and the data acquisition unit, the control unit is used for issuing an execution instruction to carry out multi-target scheduling, meanwhile, the operation execution data is stored in the management platform for subsequent training and optimization, and meanwhile, the energy consumption data of the energy storage regulation and control unit are acquired in the multi-target scheduling process, the energy storage regulation and control unit is used for iterating a scheduling scheme by combining real-time state monitoring, and regenerating a new scheduling scheme instruction, so that energy storage regulation and control is realized, and energy-saving group control is finally rea