Updating method and device of elevator running state prediction model and prediction system
By acquiring elevator operating parameters and performing quantitative index calculations, the system determines whether to update the elevator operating status prediction model. This solves the problem that existing systems cannot accurately predict elevator status, and achieves efficient model updates and improved system reliability.
Patent Information
- Application Number
- CN202411071053.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-06
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-08-06
AI Technical Summary
Existing elevator operation status prediction systems cannot accurately predict elevator status, leading to a decrease in the reliability of the prediction system and a lack of effective model update methods.
By acquiring the operating parameters of the target elevator, data alignment and quantification index calculation are performed to determine whether it falls within the preset threshold range. If so, parameter aggregation is performed to update the elevator operating status prediction model.
This improves the update efficiency of the elevator operation status prediction model and the operational reliability of the system, ensures the accuracy and timeliness of detection, and enhances the ability to respond to abnormal situations.
Smart Images

Figure CN118929362A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of elevator operation status prediction, and in particular to the technical field of an elevator operation status prediction model updating method, device and prediction system. Background Art
[0002] Patent application CN117611135A discloses a visual equipment operation and maintenance control method and system, which specifically discloses: using a neural network game theory model to realize multi-source heterogeneous data collection of equipment and environment, and using federated learning technology to process data under the premise of protecting data privacy; using the RELATION NET algorithm based on graph neural network to establish a dynamic digital twin of the entire life cycle of equipment operation to achieve quasi-real-time mapping of multi-dimensional status of equipment; using the PPO algorithm based on reinforcement learning to train the strategy model for equipment anomaly detection and fault prediction, and using the elastic analysis engine to monitor model performance and safety; using the model editor expert network based on reinforcement transfer learning to achieve online adaptation and upgrading of the model. This application realizes quasi-real-time high-fidelity mapping of equipment status through the collection and deep cross-domain analysis of multi-source heterogeneous data of the power system, providing a basis for equipment health assessment and condition prediction.
[0003] It is understandable that the above technical solution can be used for elevator equipment in smart industrial parks or smart residential communities to achieve health assessment and status prediction of elevator equipment, which is beneficial to improve the stability of daily operation of elevator equipment on the one hand, and reduce the labor cost of elevator equipment operation and maintenance on the other hand. However, as the equipment continues to run, the prediction results of the elevator operation status prediction system cannot accurately predict the state of the elevator, resulting in a decrease in the reliability of the prediction system.
[0004] Therefore, there is an urgent need for an updating method of the elevator operation status prediction model to solve the above problems, and then obtain a reliable elevator operation status prediction system. Summary of the invention
[0005] The technical problem to be solved by the present invention is to provide an updating method, device and prediction system for an elevator operation status prediction model, which can improve the reliability of the operation of the elevator operation status prediction system by efficiently executing the updating operation of the prediction model.
[0006] To this end, the present invention provides a method for updating an elevator operation state prediction model, which comprises:
[0007] S101, obtaining a target operating parameter group of a target elevator during a preset first monitoring time period;
[0008] S102, aligning the parameters in the target operating parameter group to obtain a monitoring data set of the target elevator; the monitoring data set includes the target operating parameters of the target elevator in a preset first monitoring time period and label information corresponding to the target operating parameters indicating a predicted operating state of the target elevator;
[0009] S103, updating the quantitative index based on the current data predicted by the monitoring data set;
[0010] S104, determining whether the current data update quantitative index is within a preset first threshold range, if not, not updating; if yes, executing step S105;
[0011] S105, performing a parameter aggregation operation on the elevator operation state prediction model parameters corresponding to the target elevator to obtain a first result;
[0012] S106. Feedback the first result to the target elevator and update the corresponding elevator operation status prediction model.
[0013] The updating method of the elevator operation status prediction model disclosed in the present invention predicts the current data update quantitative index according to the monitoring data set, and then determines whether to execute the update operation of the elevator operation status prediction model by judging whether the current data update quantitative index falls within the first threshold range, which is conducive to quantitatively evaluating the necessity of executing the update operation, accurately updating the elevator operation status model, improving the efficiency of updating the elevator operation status prediction model, and thus improving the high efficiency of the operation of the elevator operation status prediction system.
[0014] Furthermore, the label information of the predicted operating state of the target elevator includes one of first label information indicating that the elevator state prediction model predicts that the target elevator is in a first calibrated operating state, second label information indicating that the elevator state prediction model predicts that the target elevator is in a second calibrated operating state, and third label information indicating that the elevator state prediction model predicts that the target elevator is in an uncalibrated operating state, wherein the uncalibrated operating state indicates that the elevator state prediction model predicts that the target elevator is in an operating state other than the first calibrated operating state and the second calibrated operating state.
[0015] Furthermore, the current data update quantitative index is predicted based on the monitoring data set, specifically including:
[0016] Calculate the evaluation index based on the target operating parameters;
[0017] The calculation formula of the evaluation index C is as follows:
[0018]
[0019] Where k is the total number of target operating parameters, C i is the evaluation index component of the i-th target operating parameter; where C i The calculation formula is as follows:
[0020]
[0021] M represents the sample number of the i-th target operating parameter in the preset monitoring time, K represents the total number of samples of the i-th target operating parameter in the preset monitoring time, N M represents the parameter value corresponding to the Mth sample of the i-th target operating parameter in the preset monitoring time, N max N M The maximum value, N min N M The minimum value of N M The mean of
[0022] Calculate a feature index based on the evaluation index and the number of samples of each type of label information;
[0023] The calculation formula of characteristic index l is as follows:
[0024]
[0025] Where C is the evaluation index, M1 represents the number of samples of the target elevator's label information corresponding to the first label information during the preset monitoring time, M2 represents the number of samples of the target elevator's label information corresponding to the second label information during the preset monitoring time, and M3 represents the number of samples of the target elevator's label information corresponding to the third label information during the preset monitoring time;
[0026] Calculate the current update quantization index according to the characteristic index;
[0027] The calculation formula of the data update quantitative index is as follows:
[0028]
[0029] Wherein, l represents the characteristic index corresponding to the current data update quantization index, and e is the natural logarithm.
[0030] Furthermore, the target operating parameter group includes:
[0031] One or more of the target elevator's control box temperature, load capacity, wire rope cross-sectional area, elevator door opening and closing speed, car operation noise and power supply voltage.
[0032] Furthermore, the updating method of the elevator operation state prediction model further includes an updating monitoring step, which specifically includes:
[0033] S201, obtaining the number of executions of the update operation of the elevator operation state prediction model performed by the target elevator within a preset second monitoring time period;
[0034] S202, determining whether the ratio between the number of executions and the preset second monitoring duration is within a preset second threshold range, if so, executing step S203a, if not, executing step S203b;
[0035] S203a, outputting data indicating that an abnormal situation occurs when the target elevator performs an update operation of the elevator operation status prediction model;
[0036] S203b. Output data indicating that there is no abnormality in the update operation of the elevator operation status prediction model executed by the target elevator.
[0037] On the other hand, the present invention also provides an updating device for an elevator operation state prediction model, comprising:
[0038] A parameter acquisition module, used to obtain a target operating parameter group of a target elevator during a preset first monitoring time period;
[0039] A data preprocessing module is used to perform data alignment on the parameters in the target operation parameter group to obtain a monitoring data set of the target elevator; the monitoring data set includes the target operation parameters of the target elevator in a preset first monitoring time period and label information corresponding to the target operation parameters indicating the predicted operation state of the target elevator;
[0040] Data update quantitative index calculation module: used to predict the current data update quantitative index based on the monitoring data set;
[0041] The first determination module is used to determine whether the current data update quantitative index is within the preset first threshold range. If not, no update is performed; if yes, the federation aggregation module is executed;
[0042] A federation aggregation module: configured to perform a parameter aggregation operation on the elevator operation state prediction model parameters according to the elevator operation state prediction model parameters corresponding to the target elevator to obtain a first result;
[0043] Model updating module: Feedback the first result to the target elevator and update the corresponding elevator operation status prediction model. If not, no update is performed.
[0044] Furthermore, the data update quantitative index calculation module includes:
[0045] Evaluation index calculation submodule: used to calculate the evaluation index according to the target operating parameters;
[0046] The calculation formula of the evaluation index C is as follows:
[0047]
[0048] Where k is the total number of target operating parameters, C i is the evaluation index component of the i-th target operating parameter; where C i The calculation formula is as follows:
[0049]
[0050] M represents the sample number of the i-th target operating parameter in the preset monitoring time, K represents the total number of samples of the i-th target operating parameter in the preset monitoring time, N M represents the parameter value corresponding to the Mth sample of the i-th target operating parameter in the preset monitoring time, N max N M The maximum value, N min N M The minimum value of N M The mean of
[0051] A feature index calculation submodule: used to calculate the feature index according to the evaluation index and the number of samples of each type of label information;
[0052] The calculation formula of characteristic index l is as follows:
[0053]
[0054] Where C is the evaluation index, M1 represents the number of samples of the target elevator's label information corresponding to the first label information during the preset monitoring time, M2 represents the number of samples of the target elevator's label information corresponding to the second label information during the preset monitoring time, and M3 represents the number of samples of the target elevator's label information corresponding to the third label information during the preset monitoring time;
[0055] Update quantization index calculation submodule: used to calculate the current update quantization index according to the characteristic index;
[0056] The calculation formula of the data update quantitative index is as follows:
[0057]
[0058] Wherein, l represents the characteristic index corresponding to the current data update quantization index, and e is the natural logarithm.
[0059] Furthermore, it also includes an update monitoring unit, which includes:
[0060] An execution count module, used to obtain the execution count of the update operation of the elevator operation state prediction model executed by the target elevator within a preset second monitoring time period;
[0061] A second determination module; used to determine whether the ratio between the number of executions and the preset second monitoring duration is within a preset second threshold range, and if so, to start the step of updating the abnormal output module; if not, to start the step of updating the normal output module;
[0062] An update abnormality output module is used to output data indicating that an abnormal condition occurs when the target elevator performs an update operation of the elevator operation status prediction model;
[0063] The update normal output module is used to output data indicating that there is no abnormality in the update operation of the target elevator executing the elevator operation status prediction model.
[0064] On the other hand, the present invention also provides an elevator operation state prediction system, characterized in that it includes:
[0065] A plurality of elevators and an elevator operation state prediction model corresponding to the plurality of elevators, a server, and an updating device as described in any one of the above items;
[0066] The plurality of elevators are connected to the server, and the updating device is connected to the server.
[0067] To sum up, the present invention first predicts the current data update quantitative index based on the monitoring data set, and then determines whether the current data update quantitative index belongs to the first threshold range to determine whether to execute the update operation of the elevator operation status prediction model, which is conducive to quantitatively evaluating the necessity of executing the update operation, accurately updating the elevator operation status model, improving the efficiency of updating the elevator operation status prediction model, thereby improving the efficiency of the elevator operation status prediction system, and then monitoring the operating status of the target elevator through the frequency of the system executing the update operation. On the premise of ensuring accurate detection, the system's response capability to abnormal conditions is further improved, ensuring the accuracy and timeliness of the elevator operation status detection by the elevator operation status prediction system.
[0068] For better understanding and implementation, the present invention is described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0069] Figure 1 is a flowchart of an exemplary method for updating an elevator operation state prediction model provided by the present invention;
[0070] Figure 2 For execution Figure 1 A structural block diagram of an updating device of the updating method shown;
[0071] Figure 3 is a flow chart of an update monitoring step of an exemplary update method provided by the present invention;
[0072] Figure 4 For execution Figure 3 The structural block diagram of the update monitoring unit of the update monitoring step is shown. DETAILED DESCRIPTION
[0073] The present invention is firstly based on an elevator operation state prediction system. The elevator operation state prediction system is based on the technology of federated learning. The elevators of the smart industrial park in a certain area are divided into one elevator group corresponding to one smart industrial park. For example, the elevators of the smart industrial park A in a certain area are divided into one elevator group AZ, the elevators of the smart park B in the area are divided into another elevator group BZ, and the elevators of the smart park C in the area are divided into another elevator group CZ. Then, a server connected to the elevator group AZ, the elevator group BZ and the elevator group CZ can be set in the area. The elevator group AZ, the elevator group BZ and the elevator group CZ can upload the parameters of the elevator operation state prediction model to the server and perform parameter aggregation operation by the server. The server then sends the updated prediction model parameters to the elevator group AZ, the elevator group BZ and the elevator group CZ. After obtaining the updated prediction model parameters, the elevator group AZ, the elevator group BZ and the elevator group CZ can update the parameters of the prediction model of a specific elevator as needed, which is conducive to improving the accuracy of the operation state prediction of the elevator. Optionally, the updating method of the elevator operation state prediction model disclosed in the present invention may be set in the above-mentioned server. Alternatively, it may be set in a prediction system outside the server that can be connected to the server and can be connected to the elevator group. Optionally, the elevator group may be a network composed of several elevators, based on which the elevator group can communicate within the local area network, and data interaction with the external system can also be realized based on the communication connection with the external system.
[0074] At this time, according to the above-mentioned elevator operation status prediction system, the operation status of the elevator equipment in the smart industrial park or smart community can be predicted. For example, when the prediction result indicates that the elevator will enter an abnormal operation state, the relevant maintenance personnel can check the elevator as soon as possible and perform troubleshooting when necessary, which is conducive to ensuring the safety of elevator users and the efficiency of the daily operation of the smart industrial park or smart community. Among them, the elevator operation status prediction system can interact with the elevator network of the smart industrial park or smart community to predict the operation status of each elevator in the elevator network.
[0075] However, the above-mentioned elevator operation status prediction system lacks the ability to update the elevator operation status prediction model. As a result, with the aging of elevator components, operation misalignment, etc., the original prediction model cannot accurately predict the operation status of the target elevator, resulting in inaccurate evaluation of the prediction system, affecting elevator maintenance and user safety.
[0076] Based on this, the present invention provides an updating method and updating device for an elevator operation status prediction model to realize irregular updating of the elevator prediction model in the system to ensure the accuracy of the elevator operation status prediction system. The present invention is explained through the following embodiments.
[0077] Embodiment 1:
[0078] See also Figure 1 and Figure 2 The updating device comprises: a parameter acquisition module 101, a data preprocessing module 102, a data update quantitative index calculation module 103, a first determination module 104, a federation aggregation module 105 and a model updating module 106; the updating device is used to execute the updating method of the elevator operation state prediction model, wherein the specific execution process of each module of the updating device of the elevator operation state prediction model is as follows:
[0079] The parameter acquisition module 101 is used to execute step S101, and acquire a target operating parameter group of a target elevator in a preset first monitoring time period.
[0080] Optionally, the target operating parameter group includes at least one parameter of the control box temperature, load capacity, wire rope cross-sectional area, elevator door opening and closing speed, car operation noise and power supply voltage of the target elevator.
[0081] The target operating parameter may be any parameter related to the elevator state. The present invention uses the main influencing item as the monitoring parameter, and other monitoring items may also be added.
[0082] The data preprocessing module 102 is used to execute step S102, align the parameters in the target operation parameter group, and obtain a monitoring data set of the target elevator. The monitoring data set includes the target operation parameters of the target elevator in the preset first monitoring time and the label information corresponding to the target operation parameters indicating the predicted operation state of the target elevator. The label information is the target operation parameter sample as input, and the prediction result indicating the operation state of the target elevator is obtained after being processed by the elevator operation state prediction model.
[0083] An elevator operation status prediction model requires one or more elevator operation parameters to evaluate the elevator operation status. The required data is input into the elevator operation status prediction model to obtain the prediction result. For example, in one target elevator operation parameter sample, the control box temperature of the target elevator is T1, the load is H1, the wire rope cross-sectional area is S1, the elevator door opening speed is D1, the car operation noise is Y1, and the power supply voltage is V1. After the elevator operation state prediction model processes the above data, the prediction result indicating the operation state of the target elevator is normal operation; in another target elevator operation parameter sample, the control box temperature of the target elevator is T2, the load is H2, the wire rope cross-sectional area is S2, the elevator door opening speed is D2, the car operation noise is Y2, and the power supply voltage is V2. After the elevator operation state prediction model processes the above data, the prediction result indicating the operation state of the target elevator is abnormal operation; in another target elevator operation parameter sample, the control box temperature of the target elevator is T3, the load is H3, the wire rope cross-sectional area is S3, the elevator door opening speed is D3, the car operation noise is Y3, and the power supply voltage is V3. After the elevator operation state prediction model processes the above data, the prediction result indicating the operation state of the target elevator is an uncalibrated operation state.
[0084] That is, there are three predicted operating states for the elevator: 1. normal operation, 2. abnormal operation, and 3. unknown state.
[0085] The data update quantitative index calculation module 103 is used to execute step S103 and predict the current data update quantitative index according to the monitoring data set.
[0086] The data update quantitative index calculation module 103 includes: an evaluation index calculation submodule 103.1, a feature index calculation submodule 103.2, and an update quantitative index calculation submodule 103.3.
[0087] The evaluation index calculation submodule 103.1 is used to execute step S103.1: calculating the evaluation index according to the target operating parameters;
[0088] The calculation formula of the evaluation index is as follows:
[0089]
[0090] Where k is the total number of target operating parameters, C i is the evaluation index component of the i-th target operating parameter; where C i The calculation formula is as follows:
[0091]
[0092] M represents the sample number of the i-th target operating parameter in the preset monitoring time, K represents the total number of samples of the i-th target operating parameter in the preset monitoring time, N M represents the parameter value corresponding to the Mth sample of the i-th target operating parameter in the preset monitoring time, N max N M The maximum value, N min N M The minimum value of N M The mean of .
[0093] The feature index calculation submodule 103.2 is used to execute step S103.2: calculate the feature index according to the evaluation index and the number of samples of each type of label information;
[0094] The update quantization index calculation submodule 103.3 is used to execute step S103.3: calculate the current update quantization index according to the characteristic index;
[0095] The specific calculation formula of the data update quantitative index is as follows:
[0096] The control module controls the data processing module to determine the current data update quantitative index based on the data difference between the monitoring data set and the preset first training data set, and the algorithm used is as follows:
[0097]
[0098] Wherein, l represents the characteristic index corresponding to the current data update quantization index, and e is the natural logarithm.
[0099] In a specific embodiment of the present application, the calculation formula of the characteristic index l is as follows:
[0100]
[0101] C represents an evaluation index of the impact of the target operating parameters on the characteristic index, M1 represents the number of samples of the target elevator's label information samples whose corresponding label information is the first label information during the preset monitoring period, M2 represents the number of samples of the target elevator's label information samples whose corresponding label information is the second label information during the preset monitoring period, and M3 represents the number of samples of the target elevator's label information samples whose corresponding label information is the second label information during the preset monitoring period.
[0102] C=C t +C h +C s +C d +C y +C v
[0103] C t Indicates the first sub-item of the evaluation index regarding the influence of the control box temperature of the target elevator on the evaluation index, C h Indicates the second sub-item of the evaluation index regarding the impact of the target elevator's load capacity on the evaluation index, C s The third sub-item of the evaluation index, C, indicates the influence of the cross-sectional area of the steel wire rope of the target elevator on the evaluation index. d The fourth sub-item of the evaluation index, C, indicates the influence of the door opening and closing speed of the target elevator on the evaluation index. y Indicates the fifth sub-item of the evaluation index regarding the impact of the target elevator's car running noise on the evaluation index, C v The sixth sub-item of the evaluation index indicating the influence of the power supply voltage of the target elevator on the evaluation index.
[0104]
[0105] t1 represents the sample number of the control box temperature of the target elevator during the preset monitoring time, t2 represents the total number of samples of the control box temperature of the target elevator during the preset monitoring time, T t1 represents the control box temperature value corresponding to the t1th sample of the control box temperature of the target elevator during the preset monitoring time, T max for The maximum value, T min for The minimum value of T t1 The mean of
[0106]
[0107] h1 represents the sample number of the target elevator's load capacity during the preset monitoring time, h2 represents the total number of samples of the target elevator's load capacity during the preset monitoring time, Indicates the load corresponding to the h1th sample of the target elevator's load during the preset monitoring time, H max for The maximum value, H min for The minimum value of for The mean of
[0108]
[0109] s1 represents the sample number of the cross-sectional area of the steel wire rope of the target elevator during the preset monitoring time, s2 represents the total number of samples of the cross-sectional area of the steel wire rope of the target elevator during the preset monitoring time, S s1represents the wire rope cross-sectional area corresponding to the s1th sample of the wire rope cross-sectional area of the target elevator during the preset monitoring time, S max for The maximum value, S min for The minimum value of for The mean of
[0110]
[0111] d1 represents the sample sequence number of the elevator door opening and closing speed of the target elevator during the preset monitoring time, d2 represents the total number of samples of the elevator door opening and closing speed of the target elevator during the preset monitoring time, represents the elevator door opening and closing speed corresponding to the d1th sample of the elevator door opening and closing speed of the target elevator during the preset monitoring time, D max for The maximum value, D min for The minimum value of for The mean of
[0112]
[0113] y1 represents the sample number of the target elevator's car running noise during the preset monitoring time, y2 represents the total number of samples of the target elevator's car running noise during the preset monitoring time, and Y y1 represents the volume corresponding to the y1th sample of the target elevator's car running noise during the preset monitoring time, Y max Y y1 The maximum value, Y min Y y1 The minimum value of Y y1 The mean of
[0114]
[0115] v1 represents the sample number of the target elevator's power supply voltage during the preset monitoring time, v2 represents the total number of samples of the target elevator's power supply voltage during the preset monitoring time, Indicates the volume corresponding to the v1th sample of the target elevator's power supply voltage during the preset monitoring time, V max for The maximum value, V min for The minimum value of for The mean of .
[0116] The first determination module 104 is used to execute step S104, determine whether the current data update quantization index is within a preset first threshold range, if not, do not update; if yes, execute step S105.
[0117] The federation aggregation module 105 is used to execute step S105, and perform a parameter aggregation operation according to the elevator operation state prediction model parameters corresponding to the target elevator to obtain a first result.
[0118] The parameter aggregation operation is an integral part of the federation algorithm, which aggregates multiple local model parameters to generate a global model, and finally distributes the updated model parameters to the corresponding devices or systems to complete the update of model parameters.
[0119] The model updating module is used to execute step S106: feeding back the first result to the target elevator to update its elevator operation status prediction model.
[0120] The updating method of the elevator operation status prediction model disclosed in the present invention predicts the current data update quantitative index according to the monitoring data set, and then determines whether to execute the update operation of the elevator operation status prediction model by judging whether the current data update quantitative index falls within the first threshold range, which is conducive to quantitatively evaluating the necessity of executing the update operation, accurately updating the elevator operation status model, improving the efficiency of updating the elevator operation status prediction model, and thus improving the high efficiency of the operation of the elevator operation status prediction system.
[0121] Embodiment 2:
[0122] Taking into account that the frequency with which the system executes the update operation also reflects the operating status of the target elevator, in this embodiment, the frequency (or period) with which the target elevator executes the above-mentioned update operation is used to screen whether there is any abnormality in the update operation. When an abnormality occurs in the update frequency or period, relevant personnel or related systems can respond in a timely manner, which is beneficial to the efficiency of the operation of the elevator operation status prediction system. Therefore, the update device of the present invention also includes an update monitoring unit, and the update method of the elevator operation status prediction model of the present invention also includes an update monitoring step, wherein the update monitoring unit includes: an execution times counting module 201, a second determination module 202, an update abnormal output module 203.a, and an update abnormal normal module 203.b.
[0123] The execution count module 201 executes step S201 to obtain the execution count of the update operation of the elevator operation state prediction model executed by the target elevator within the preset second monitoring time period;
[0124] The second determination module 202 executes step S202 to determine whether the ratio between the number of execution times and the preset second monitoring duration is within a preset second threshold range, and if so, executes step S203a, and if not, executes step S203b.
[0125] The update abnormality output module 203.a executes step S203a and outputs data indicating that an abnormal condition occurs when the target elevator performs an update operation of the elevator operation status prediction model.
[0126] The update abnormality normal module 203.b executes step S203b and outputs data indicating that there is no abnormality in the update operation of the elevator operation status prediction model executed by the target elevator.
[0127] In addition, it should be known that it is also possible to screen whether there is an abnormality in the update operation based on the frequency (or period) at which the above-mentioned update operation of the target elevator is not triggered. At this time, when the ratio between the number of non-execution times and a third monitoring time exceeds a preset third threshold range, the output is that there is no abnormality in the update operation of the target elevator to execute the elevator operation status prediction model, otherwise the output is abnormal.
[0128] In summary, the present invention first predicts the current data update quantitative index based on the monitoring data set, and then determines whether to execute the update operation of the elevator operation status prediction model by determining whether the current data update quantitative index belongs to the first threshold range, which is conducive to quantitatively evaluating the necessity of executing the update operation, accurately updating the elevator operation status model, improving the efficiency of updating the elevator operation status prediction model, thereby improving the efficiency of the elevator operation status prediction system, and then monitoring the operating status of the target elevator through the frequency of the system executing the update operation. On the premise of ensuring accurate detection, the system's response capability to abnormal conditions is further improved, ensuring the accuracy and timeliness of elevator operation status detection.
[0129] Finally, it should be noted that: in the updating method of an elevator operation status prediction model disclosed in the embodiment of the present invention, what is disclosed is only a preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the aforementioned embodiments, ordinary technicians in the field should understand that it is still possible to modify the technical solution recorded in the aforementioned embodiments, or to replace some of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solution deviate from the spirit and scope of the technical solution of the embodiment of the present invention.
Claims
1. A method for updating an elevator operation state prediction model, characterized in that: S101, obtaining a target operating parameter group of a target elevator during a preset first monitoring time period; S102, performing data alignment on the parameters in the target operating parameter group to obtain a monitoring data set of the target elevator; The monitoring data set includes a target operating parameter of the target elevator during a preset first monitoring period and label information indicating a predicted operating state of the target elevator corresponding to the target operating parameter; S103, updating the quantitative index based on the current data predicted by the monitoring data set; S104, determining whether the current data update quantitative index is within a preset first threshold range, if not, not updating; if yes, executing step S105; S105, performing a parameter aggregation operation on the elevator operation state prediction model parameters corresponding to the target elevator to obtain a first result; S106. Feedback the first result to the target elevator and update the corresponding elevator operation status prediction model.
2. The updating method of the elevator operation state prediction model according to claim 1, characterized in that: The label information of the predicted operating state of the target elevator includes one of first label information indicating that the elevator state prediction model predicts that the target elevator is in a first calibrated operating state, second label information indicating that the elevator state prediction model predicts that the target elevator is in a second calibrated operating state, and third label information indicating that the elevator state prediction model predicts that the target elevator is in an uncalibrated operating state, wherein the uncalibrated operating state indicates that the elevator state prediction model predicts that the target elevator is in an operating state other than the first calibrated operating state and the second calibrated operating state.
3. The updating method of the elevator operation state prediction model according to claim 2, characterized in that: The specific quantitative indexes for updating the current data based on the monitoring data set include: Calculate the evaluation index based on the target operating parameters; The calculation formula of the evaluation index C is as follows: Where k is the total number of target operating parameters, C i is the evaluation index component of the i-th target operating parameter; where C i The calculation formula is as follows: M represents the sample number of the i-th target operating parameter in the preset monitoring time, K represents the total number of samples of the i-th target operating parameter in the preset monitoring time, N M represents the parameter value corresponding to the Mth sample of the i-th target operating parameter in the preset monitoring time, N max N M The maximum value, N min N M The minimum value of N M The mean of Calculate a feature index based on the evaluation index and the number of samples of each type of label information; The calculation formula of characteristic index l is as follows: Where C is the evaluation index, M1 represents the number of samples of the target elevator's label information corresponding to the first label information during the preset monitoring time, M2 represents the number of samples of the target elevator's label information corresponding to the second label information during the preset monitoring time, and M3 represents the number of samples of the target elevator's label information corresponding to the third label information during the preset monitoring time; Calculate the current update quantization index according to the characteristic index; The calculation formula of the data update quantitative index is as follows: Wherein, l represents the characteristic index corresponding to the current data update quantization index, and e is the natural logarithm.
4. The updating method of the elevator operation state prediction model according to claim 3 is characterized in that: The target operating parameter group includes: One or more of the target elevator's control box temperature, load capacity, wire rope cross-sectional area, elevator door opening and closing speed, car operation noise and power supply voltage.
5. The method for updating the elevator operation state prediction model according to any one of claims 1 to 4, characterized in that: It also includes an update monitoring step, which specifically includes: S201, obtaining the number of executions of the update operation of the elevator operation state prediction model performed by the target elevator within a preset second monitoring time period; S202, determining whether the ratio between the number of executions and the preset second monitoring duration is within a preset second threshold range, if so, executing step S203a, if not, executing step S203b; S203a, outputting data indicating that an abnormal situation occurs when the target elevator performs an update operation of the elevator operation status prediction model; S203b. Output data indicating that there is no abnormality in the update operation of the elevator operation status prediction model executed by the target elevator.
6. An updating device for an elevator operation state prediction model, characterized in that include: An updating unit, the updating unit comprising: A parameter acquisition module, used to obtain a target operating parameter group of a target elevator during a preset first monitoring time period; A data preprocessing module is used to perform data alignment on the parameters in the target operation parameter group to obtain a monitoring data set of the target elevator; the monitoring data set includes the target operation parameters of the target elevator in a preset first monitoring time period and label information corresponding to the target operation parameters indicating the predicted operation state of the target elevator; Data update quantitative index calculation module: used to predict the current data update quantitative index based on the monitoring data set; The first determination module is used to determine whether the current data update quantitative index is within the preset first threshold range. If not, no update is performed; if yes, the federation aggregation module is executed; A federation aggregation module: configured to perform a parameter aggregation operation on the elevator operation state prediction model parameters according to the elevator operation state prediction model parameters corresponding to the target elevator to obtain a first result; Model updating module: Feedback the first result to the target elevator and update the corresponding elevator operation status prediction model. If not, no update is performed.
7. The updating device for an elevator operation state prediction model according to claim 6, characterized in that: The data update quantitative index calculation module includes: Evaluation index calculation submodule: used to calculate the evaluation index according to the target operating parameters; The calculation formula of the evaluation index C is as follows: Where k is the total number of target operating parameters, C i is the evaluation index component of the i-th target operating parameter; where C i The calculation formula is as follows: M represents the sample number of the i-th target operating parameter in the preset monitoring time, K represents the total number of samples of the i-th target operating parameter in the preset monitoring time, N M represents the parameter value corresponding to the Mth sample of the i-th target operating parameter in the preset monitoring time, N max N M The maximum value, N min N M The minimum value of N M The mean of A feature index calculation submodule: used to calculate the feature index according to the evaluation index and the number of samples of each type of label information; The calculation formula of characteristic index l is as follows: Where C is the evaluation index, M1 represents the number of samples of the target elevator's label information corresponding to the first label information during the preset monitoring time, M2 represents the number of samples of the target elevator's label information corresponding to the second label information during the preset monitoring time, and M3 represents the number of samples of the target elevator's label information corresponding to the third label information during the preset monitoring time; Update quantization index calculation submodule: used to calculate the current update quantization index according to the characteristic index; The calculation formula of the data update quantitative index is as follows: Wherein, l represents the characteristic index corresponding to the current data update quantization index, and e is the natural logarithm.
8. The updating device for an elevator operation state prediction model according to claim 7, characterized in that: Also included is an update monitoring unit, which includes: An execution count module, used to obtain the execution count of the update operation of the elevator operation state prediction model executed by the target elevator within a preset second monitoring time period; A second determination module; used to determine whether the ratio between the number of executions and the preset second monitoring duration is within a preset second threshold range, and if so, to start the step of updating the abnormal output module; if not, to start the step of updating the normal output module; An update abnormality output module is used to output data indicating that an abnormal condition occurs when the target elevator performs an update operation of the elevator operation status prediction model; The update normal output module is used to output data indicating that there is no abnormality in the update operation of the target elevator executing the elevator operation status prediction model.
9. An elevator operation status prediction system, characterized in that: include: A plurality of elevators and an elevator operation status prediction model corresponding to the plurality of elevators, a server, and an updating device according to any one of claims 6 to 8; The plurality of elevators are connected to the server, and the updating device is connected to the server.
Citation Information
Patent Citations
Visual equipment operation and maintenance management and control method and system
CN117611135A
Optimization method and optimization device for elevator carrying efficiency
CN113307111A
Floor population size prediction method based on elevator operation data, and related equipment
CN113666209A
State detection model training method and device, state detection method and device and storage medium
CN117076926A
Elevator operation condition evaluation method, elevator operation condition maintenance method and application
CN118125252A
