Method and device for determining safety risk of elevator system

Abstract

The present disclosure relates to elevator technology, in particular to a method, an apparatus and a non-transitory computer-readable storage medium storing a computer program for implementing the method for determining a security risk of an elevator system. According to one aspect of the present disclosure, there is provided a method for determining a safety risk of an elevator system. According to the method, the device for determining the safety risk generates a feature vector at least based on operating state data of the elevator system, and then determines a probability that the transport object is trapped in a car of the elevator system using a neural network model. In the method, the feature vector and the probability are respectively an input variable and an output variable of a neural network model, and the feature vector comprises a component corresponding to a combination selected from state values of a plurality of first category features. Furthermore, a state value of the first category of features is determined based on the operating state data.

Description

Technical Field

[0001] The present disclosure relates to elevator technology, and particularly to a method, an apparatus, and a non-transitory computer-readable storage medium storing a computer program for implementing the method for determining a safety risk of an elevator system. Background Art

[0002] When a passenger is trapped in a car, he or she can press an emergency call button in the car or use a mobile phone to call for help; in addition, the elevator system can also automatically send an alarm message to the background system in case of a suspected entrapment event. Subsequently, rescue personnel will be dispatched to the scene to carry out rescue activities. However, in actual operation, quite a number of distress events and suspected entrapment events are not caused by elevator system failures or can be automatically eliminated by means of the recovery function of the elevator system, so there is no need for rescue personnel to go to the scene to handle them. For the industry, how to identify real entrapment events to provide rescue assistance while reducing the consumption of rescue resources has always been a challenging task. Summary of the Invention

[0003] According to one aspect of the present disclosure, there is provided a method for determining a safety risk of an elevator system. According to this method, a device for determining a safety risk generates at least a feature vector based on operation state data of the elevator system, and then uses a neural network model to determine the probability that a transported object is trapped in a car of the elevator system. In the above method, the above feature vector and probability are input variables and output variables of the neural network model respectively, and the feature vector includes components corresponding to a combination selected from state values of a plurality of first category features. In addition, the state values of the first category features are determined based on the operation state data.

[0004] Optionally, in the above method, the generation of the feature vector is further based on a safety event. Correspondingly, the feature vector further includes components corresponding to a combination of the state values of the first category features and the state values of the second category features, and the state values of the second category features are determined based on the safety event.

[0005] Optionally, the above method further includes a step of generating an evaluation result regarding the safety risk based on the above probability; or optionally, the above method further includes a step of training the neural network model. In the training step, the operation logs of the elevator system are used to label training samples for training the neural network model.

[0006] Optionally, in the above method, the operating state data describes multiple types of operating states of the elevator system, and the state value of each first-category feature includes one of the following state values of one of the types of operating states: i) the state value before a safety event occurs; ii) the state value when a safety event occurs; iii) the state value after a safety event occurs. Further, each component of the feature vector corresponds to one of the following: i) a combination of state values of first-category features; ii) a combination of state values of first-category features and state values of second-category features. Still further, the step of generating the feature vector includes: generating state values of multiple first-category features from the operating state data and generating state values of second-category features from the safety event, and then generating the feature vector from the state values of the first-category features and the state values of the second-category features.

[0007] Optionally, in the above method, the step of generating the feature vector is performed in response to the occurrence of a safety event.

[0008] Optionally, in the above method, the type of the operating state includes one or more of the following: elevator system operating mode, car movement direction, floor where the car is located, car leveling alignment state, traction machine movement state, car door and landing door opening and closing state.

[0009] Optionally, in the above method, the type of the safety event includes one or more of the following: power failure, sensor failure, door system failure, mechanical component failure, and abnormal elevator system operating mode.

[0010] Optionally, in the above method, the state values of the first-category features and the state values of the second-category features are represented in a one-hot encoding form, and each component of the vector feature is represented in a binary value form.

[0011] Optionally, in the above method, in the step of generating the evaluation result regarding the safety risk, based on a preset mapping relationship, the level of the safety risk is determined from the above probability, where the mapping relationship defines the probability value range corresponding to each level.

[0012] Optionally, the above method is implemented by one of the following: cloud computing hardware device, a control system for controlling multiple elevator systems, and a controller in the elevator system.

[0013] According to another aspect of the present disclosure, there is provided a device for determining the safety risk of an elevator system. The device includes at least one processor, at least one memory, and a computer program stored on the memory. When the computer program runs on the processor, it will cause the following operations: generating a feature vector based at least on the operating state data of the elevator system, and then using a neural network model to determine the probability that a transported object is trapped in the car of the elevator system. The above-mentioned feature vector and probability are the input variable and output variable of the neural network model respectively, and the feature vector includes components corresponding to a combination selected from the state values of a plurality of first-category features, wherein the state values of the first-category features are determined based on the operating state data.

[0014] According to still another aspect of the present disclosure, there is provided a computer-readable storage medium having stored thereon a computer program adapted to be executed on a processor of a terminal device, and the execution of the computer program causes the steps of the method described above to be executed. Description of the Drawings

[0015] The above and / or other aspects and advantages of the present disclosure will become clearer and easier to understand through the following description of various aspects in conjunction with the drawings, in which the same or similar units are denoted by the same reference numerals. The drawings include:

[0016] Figure 1 A view of an exemplary elevator system.

[0017] Figure 2 A flowchart of a method for determining the safety risk of an elevator system according to an embodiment of the present disclosure.

[0018] Figure 3 A schematic diagram of an exemplary neural network model.

[0019] Figure 4 A flowchart of a method for determining the safety risk of an elevator system according to another embodiment of the present disclosure.

[0020] Figure 5 A flowchart of a method for determining the safety risk of an elevator system according to another embodiment of the present disclosure.

[0021] Figure 6 A schematic block diagram of a device for determining the safety risk of an elevator system according to another embodiment of the present disclosure. Detailed Description of the Embodiments

[0022] The present disclosure will be described more fully hereinafter with reference to the accompanying drawings, in which illustrative embodiments of the disclosure are shown. The present disclosure may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. The above-described embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0023] In this specification, terms such as "including" and "comprising" mean that the technical solutions of the present disclosure do not exclude the presence of other elements and steps that are not directly or explicitly recited, in addition to the elements and steps that are directly and explicitly recited in the specification and claims.

[0024] Figure 1 A view of an exemplary elevator system. Figure 1 The illustrated elevator system 101 includes an elevator car 103, a counterweight 105, a tension member 107, a guide rail (or track system) 109, a machine unit (or machine unit system) 111, a position reference system 113, and an electronic elevator controller (controller) 115. The elevator car 103 and the counterweight 105 may be connected to each other by the tension member 107. The tension member 107 may include or be configured as, for example, a rope, a steel cable, and / or a coated steel belt. The counterweight 105 is configured to balance the load of the elevator car 103 and is configured to assist the elevator car 103 in moving in the elevator shaft (or hoistway) 117 and along the guide rail 109 in an opposite direction and simultaneously with respect to the counterweight 105.

[0025] The tension member 107 may engage the machine unit 111, and the machine unit 111 may be part of the overhead structure of the elevator system 101. The machine unit 111 is configured to control the movement between the elevator car 103 and the counterweight 105. The position reference system 113 may be mounted on a fixed part at the top of the elevator shaft 117, such as on a support or a guide rail, and may be configured to provide position signals related to the position of the elevator car 103 in the elevator shaft 117. In other embodiments, the position reference system 113 may be directly mounted to a moving component of the machine unit 111, or may be located in other positions and / or configurations known in the art. The position reference system 113 may be any device or mechanism known in the art for monitoring the position of the elevator car and / or the counterweight. As will be understood by those skilled in the art, the position reference system 113 includes, for example, but is not limited to, an encoder, a sensor, or other systems, and may implement various sensing such as speed sensing, absolute position sensing.

[0026] As shown, the controller 115 is located in the controller room 121 of the elevator shaft 117 and is configured to control the operation of the elevator system 101 (and in particular the elevator car 103). For example, the controller 115 can provide drive signals to the machine unit 111 to control the acceleration, deceleration, leveling, stopping, etc. of the elevator car 103. The controller 115 can also be configured to receive position signals from the position reference system 113 or any other desired position reference device. As it moves up or down along the guide rails 109 within the elevator shaft 117, the elevator car 103 can stop at one or more landings 125 as controlled by the controller 115. Although shown in the controller room 121, those skilled in the art will appreciate that the controller 115 can be located and / or configured elsewhere or at other positions within the elevator system 101. In one embodiment, the controller can be remotely located or in the cloud.

[0027] The machine unit 111 can include a motor or a similar drive mechanism. According to an embodiment of the present disclosure, the machine unit 111 is configured to include an electric drive motor. The power source for the motor can be any power source, including the power grid, and the power source is supplied to the motor in combination with other components. The machine unit 111 can include a traction sheave that transmits force to the tension member 107 to move the elevator car 103 within the elevator shaft 117.

[0028] In this specification, the operating state data of the elevator system refers to the parameters or state values used to describe the operating state or working state of various components or units (such as the various units of the elevator system shown above) Figure 1 The types of motion states, for example, include but are not limited to the elevator system operating mode, car motion direction, floor where the car is located, car leveling alignment state, traction machine motion state, and car door and landing door opening and closing states, etc. in some embodiments of the present disclosure.

[0029] In this specification, the transported object refers to the object transported by the elevator system, which includes, for example, people, animals, and equipment, etc.

[0030] In this specification, being trapped in the car or a trapped event refers to an event where the transported object cannot leave the car and intervention measures need to be taken. As for the type of intervention measures taken, it can be defined by the user according to application requirements. In some examples, for example, the intervention measure can be defined as dispatching personnel to the site for rescue.

[0031] It should be noted that in the present disclosure, the fact that the transported object cannot leave the car does not necessarily mean that it is in a trapped state or that a trapped event has necessarily occurred. For example, when the car is moving between two floors, the car door and the landing door will be automatically prohibited from opening, but at this time the elevator system is still in a normal operating state, so this event is not a trapped event; another example is that although the car door cannot be opened when the car stops at a floor, if the elevator system maintenance personnel can eliminate the fault through remote control, then this event may not be regarded as a trapped event either.

[0032] In this specification, a safety event refers to an event related to a trapped event during the operation of the elevator system, and it includes, for example, but is not limited to one or more of the following types: power failure of the elevator system, sensor failure, door system failure, mechanical component failure, and abnormal operation mode of the elevator system, etc.

[0033] Generally speaking, the occurrence of a trapped event will necessarily be reflected in the operating state data or safety events of the elevator system. However, since there are many factors that cause trapped events and there may be complex coupling relationships between these factors, it will be very difficult to directly identify trapped events or determine the occurrence probability of trapped events based on the operating state data or safety events. In some embodiments of the present disclosure, a neural network model is used to establish the correlation or mapping relationship between the operating state data and the occurrence probability of trapped events. In particular, in some embodiments, the components of the feature vector input to the neural network model used are combinations of categorical features (hereinafter also referred to as first categorical features) derived from the operating state data. In other embodiments, in addition to the combinations of categorical features derived from the operating state data (such as combinations of various types of operating state data), the components of the feature vector can also be combinations of the first categorical features and categorical features (hereinafter also referred to as second categorical features) derived from safety events. The forms and generation methods of the first categorical features and the second categorical features will be further described below.

[0034] In some embodiments of the present disclosure, by using combinations of the first categorical features derived from the motion state data instead of directly using the motion state data to construct the components of the feature vector, the motion state data that has a strong correlation with the trapped event can be "aggregated" together, making the correlation between the motion state data and the trapped event more obvious, and providing more interpretability for the output result of the neural network model. In other embodiments of the present disclosure, by further introducing the second categorical features derived from safety events when constructing the components of the feature vector, the above technical effects are further enhanced.

[0035] Figure 2Flowchart of a method for determining the safety risk of an elevator system according to an embodiment of the present disclosure. The methods described below can be implemented by various devices, such as but not limited to cloud computing hardware devices (e.g., cloud server clusters), control systems for controlling multiple elevator systems, and controllers in elevator systems. Hereinafter, these devices are collectively referred to as devices for determining the safety risk of an elevator system or safety risk determination devices.

[0036] Figure 2 The method shown starts at step 201. In this step, the safety risk determination device trains a neural network model for determining the entrapment probability of the transported object. Figure 3 Schematic diagram of an exemplary neural network model, which can be used to determine the entrapment probability P of the transported object. As Figure 3 shown, the neural network model 30 includes an input layer 310, a hidden layer 320, and an output layer 330. The input layer 310 includes a plurality of input nodes 310-1 to 310-n. The number of hidden layers 320 can be one or more. The output layer 330 includes an output node y, and the result output by it can be used as the entrapment probability P of the transported object.

[0037] The operation logs of the elevator system usually contain rich fault information and maintenance operation information (such as the fault types of the elevator system and the corresponding maintenance operation records). Therefore, various operation state data can be labeled by the log data to obtain a training sample set and a test sample set for the neural network model. By reusing the operation logs, not only the training cost of the neural network model is reduced, but also the quality of the training samples is improved.

[0038] It should be noted that in Figure 2 the method shown, step 201 is an optional step. For example, when the neural network model for determining the entrapment probability of the transported object can be obtained by other means (such as commercial purchase), step 201 will no longer be required.

[0039] Continuing to refer to Figure 2 in step 202, the safety risk determination device generates a feature vector X or [χ1, χ2…χ n . The generation method of the feature vector will be further described below.

[0040] Subsequently, it enters step 203. The safety risk determination device inputs the feature vector X into the neural network model to obtain the probability P that the transported object is trapped in the car of the elevator system as the output variable of the model. Exemplarily, each component χ n of the feature vector [χ1, χ2…χ i(i = 1...n) is input into a corresponding one of the multiple input nodes 310-1 to 310-n. In some specific implementations, each component of the vector feature is represented in the form of a binary value, and the neural network model can be a classifier model, and the classification result output by it includes '1' (for example, indicating that a trapped event has occurred in the elevator system) or '0' (for example, indicating that a trapped event has not occurred in the elevator system).

[0041] After completing step 203, Figure 2 The shown process enters step 204. In step 204, the safety risk determination device generates an evaluation result of the safety risk of the elevator system based on the probability P obtained in step 203. In some specific implementations, the value range of the probability P can be divided into multiple intervals, and each value interval corresponds to one of the safety risk levels, thereby obtaining the mapping relationship between the probability value interval and the safety risk level. In one example, the value range of the probability P is [0, 1], which is divided into three intervals: [0, 0.3), [0.3, 0.6), and [0.6, 1], corresponding to the three levels of low risk, medium risk, and high risk respectively. In another example, the probability P is the discrete values 1 and 0, corresponding to the occurrence and non-occurrence of a trapped event respectively.

[0042] Correspondingly, in step 204, the safety risk determination device can determine the corresponding safety risk level based on the above mapping relationship and the probability P obtained in step 202.

[0043] Figure 4 A flowchart of a method for determining the safety risk of an elevator system according to another embodiment of the present disclosure. In this embodiment, it is assumed that the safety risk determination device is a cloud computing hardware device or a control system for controlling multiple elevator systems. The following is with the help of Figure 4 For Figure 2 The implementation of step 202 in is further described.

[0044] Figure 4 The shown process starts from step 401, and this step can be continued, for example, after Figure 2 Step 201 in. In step 401, the safety risk determination device determines whether it has received a report FAILURE_REPORT on the occurrence of a safety event from the gateway connected to the elevator system. If the report FAILURE_REPORT is received, it enters step 402, otherwise, it continues to monitor whether the report FAILURE_REPORT is received from the gateway.

[0045] In some specific implementations, in response to the occurrence of a security event (such as various types of security events described above), the controller of the elevator system sends a report FAILURE_REPORT about the occurrence of the security event and the operating status data DATA associated with the security event of the elevator system to the cloud via the gateway. As described above, the operating status data can describe various types of operating statuses of the elevator system. In some specific implementations, the operating status data DATA contains data samples S1 to S of m (m≥1) moments m , and each data sample may contain the status values of various types of operating statuses at one moment, which can be represented in the following form, for example:

[0046]

[0047] As shown in the above representation, each data sample contains multiple fields (each string separated by a space within the square brackets, and the number below each string represents the corresponding field number), and their meanings are shown in Table 1:

[0048] Table 1

[0049] Field Number Value Example Usage 1 True / False Indicates Data Integrity 2 A / B / C / D Defines the Connection Method between the Gateway and the Elevator System 3 u / U / d / D Car Travel Direction 4 Floor Number Car 5 NOR / NAV / IDL / COR Operation Mode 6 FR / SR / ST / ID Traction Machine Motion Status 7 "<>><" / "><][" / "><<>" / "><><" / "><DD" Car Door Status 8-10 DZ1LV2LV / dz1lv2lv Status of Car Level Alignment 11 1668651867 Timestamp

[0050] In step 402, the security risk determination device generates the status value of the first category feature from the received operating status data DATA, and generates the status value of the second category feature from the security event. The first category feature and the second category feature are usually variables that change over time, and hereinafter, the value of these variables at a certain moment is referred to as the status value or the current instance. Exemplarily, k (k≥2) types of first category features are denoted as {F1(j)} (j = 1...k), and the second category feature is denoted as F2. It should be noted that although the number of types of the second category feature is 1 in the above example, this is only exemplary, and the number of types of the second category feature can also be multiple.

[0051] Table 2 shows an example of the types of the first category feature and the second category feature. Among them, the first row represents the types of category features, and "..." in the column corresponding to each type of category feature represents the possible values of that type of category feature.

[0052] Table 2

[0053] Opmode Event Pre_Opmode Last_Opmode Motion_Before_Change DoorZone floor … … … … … … … … … … … … … … … … … … … … … … … … … … … …

[0054] In Table 2, the Event column represents the type of the security event, the Opmode column represents the possible abnormal operating modes of the elevator system when various types of security events occur (hereinafter denoted as moment t0), and the Pre_Opmode column represents a certain moment before various types of security events occur (such as moment (t0 - Δ P), where Δ P is a preset time interval) possible operating modes of the elevator system. The Last_Opmode column represents the operating mode of the elevator system at a certain moment (e.g., at time (t0 + Δ L ), where Δ L is a preset time interval) after various types of safety events occur. The Motion_Before_Change column represents the motion state of the traction machine at a certain moment (e.g., at time (t0 - Δ M ), where Δ M is a preset time interval). The DoorZone column represents the state of the car leveling alignment when various types of safety events occur. The floor column represents the floor where the car is located when various types of safety events occur.

[0055] Table 3 shows example state values of the first category of features and the second category of features or the current instance {F1(j)} (j = 1…k), F2, where the state values can be determined by the operating state data DATA and safety events. Specifically, the state values of the first category of features such as Opmode, Pre_Opmode, Last_Opmode, Motion_Before_Change, DoorZone, open_time, and floor can be determined by the operating state data DATA. For example, the state values of the category features Opmode, Pre_Opmode, and Last_Opmode can be determined by fields 5 and 11 in the data samples S1 to S m . The state value of the category feature Motion_Before_Change can be determined by fields 6 and 11, and the state value of the category feature DoorZone can be determined by fields 8 - 11; on the other hand, the state value of the second category of feature Event in Table 3 can be determined based on the type of safety event in the reported FAILURE_REPORT.

[0056] In Table 3, "NAV" indicates that the elevator system is in an operating mode with an unknown reason for stopping, "INS" indicates that the elevator system is in a maintenance operating mode, and "NOR" indicates that the elevator system is in a normal operating mode.

[0057] Table 3

[0058] Opmode Event Pre_Opmode Last_Opmode Motion_Before_Change DoorZone floor NAV 0218 INS NOR SR 0 18

[0059] After completing step 402, Figure 4The process shown enters step 403. In step 403, the security risk determination device performs an encoding process on the status values of the first category features and the second category features or the current instances {F1(j)} (j = 1...k), F2. In some specific implementation manners, the one-hot encoding method is adopted to convert each of the status values of the first category features and the second category features into corresponding vectors. For example, for a category feature with three states or values (assumed to be state A, state B, and state C), the one-hot encoding will create three new columns for these three states. If the current instance of the category feature is in state A, the value of the first column is '1' and the values of the second and third columns are '0'. If the current instance of the category feature is in state B, the values of the first and third columns are '0' and the value of the second column is '1'. If the current instance of the category feature is in state C, the values of the first and second columns are '0' and the value of the third column is '1'. Thus, each state of a category feature will be represented as a binary vector.

[0060] Table 4 shows an example of the status values of the first and second category features and the corresponding encoded values, where the first row represents the status values of the category features, and the subsequent rows are the vectorized representations of the status values, that is, the numbers in each category feature column form the respective components of the corresponding vector.

[0061] Table 4

[0062] Opmode_NAV Event_0218 Pre_NOR Last_INS Motion_Before_Change_ID DoorZone floor 0 1 1 0 1 1 0 1 0 1 0 0 0 1 … … … … … … …

[0063] It should be noted that the encoding process is actually to convert the categorical feature variables into a format that is easy to be processed by some machine learning algorithms. Therefore, whether to perform the encoding process depends on the neural network model or machine learning algorithm adopted. That is to say, in the Figure 4 method shown, step 403 is an optional step.

[0064] After completing step 403, Figure 4 the process shown enters step 404. In step 404, the security risk determination device generates a feature vector (such as the feature vector [χ1, χ2... χ n ) from the status values {F1(j)} (j = 1...k) of the first category features and the status value F2 of the second category features.

[0065] Various combinations of the status values of the first category features can be constructed by selecting one or more status values from the status values {F1(j)} (j = 1...k) of the first category features. In some specific implementation manners, the component χ of the feature vector i(i = 1...n) can take values corresponding to one of various combinations formed by the state values of the first category of features. Specifically, for a component χ with m (m ≥ 2) possible values i , corresponding state value combinations can be defined for each possible value.

[0066] In addition, by selecting one or more state values from the state values {F1(j)} (j = 1...k) of the first category of features and adding the state value F2 of the second category of features, various combinations formed by the state values of the first category of features and the state values of the second category of features can be constructed. In some other specific implementation manners, the value of the component χ i of the feature vector can also correspond to one of various combinations formed by the first category of features and the second category of features, that is, corresponding state value combinations are defined for each possible value of the component χ i .

[0067] In the selection of the state value combination, it is possible to consider incorporating the state values of category features (such as Opmode, Pre_Opmode, and Last_Opmode) that belong to the same operating state type and have a time sequence relationship into the combination. In one example, assuming that the component χ i contains two values '1' and '0' and the operating mode of the elevator system changes over time and is correlated with the identity of the person entering the car (ordinary passenger or maintenance personnel), then it is possible to consider determining the value of the component χ i based on the combination of the state values of the first category of features Opmode and Pre_Opmode. For example, when any of the following multiple state value combinations occurs, the value of the component χ i will be determined to be '1' (which indicates a high probability that a maintenance personnel appears in the car), otherwise the value of the component χ i will be determined to be '0' (which indicates a low probability that a maintenance personnel appears in the car):

[0068] Combination - 1: The state value of Opmode is "NAV" and the state value of Pre_Opmode is "INS";

[0069] Combination - 2: The state value of Opmode is "HAD" (which indicates that the elevator system is in the operating mode of hoistway access detection) and the state value of Pre_Opmode is "INS";

[0070] Combination - 3: The state value of Opmode is "HAD" and the state value of Pre_Opmode is "ESB" (which indicates that the elevator system is in the operating mode where the emergency stop button is pressed).

[0071] In another example, assuming that the component χ i+1If it contains two values '1' and '0' and the operating mode of the elevator system changes over time and is correlated with whether the people in the car have the ability to self-rescue, then it can be considered to determine the component χ based on the combination of the state values of the first category features Opmode and Last_Opmode. i+1 For example, when the following combination of state values occurs, the value of the component χ i+1 will be determined to be '1' (which indicates a high probability of door machine jamming), otherwise the value of the component χ i+1 will be determined to be '0' (which indicates a low probability of door machine jamming): the state value of Opmode is "NOR" and the state value of Last_Opmode is "DTO" (which indicates that the elevator system is in the operating mode of door opening timeout protection).

[0072] It should be noted that in step 404, the feature vector can be generated only based on the state values {F1(j)} (j = 1...k) of the first category features, or the feature vector can be generated based on the state values {F1(j)} (j = 1...k) of the first category features and the state value F2 of the second category features.

[0073] After completing step 404, Figure 4 the process shown goes to Figure 2 step 203 in

[0074] Figure 5 FIG. is a flowchart of a method for determining the safety risk of an elevator system according to another embodiment of the present disclosure. In this embodiment, it is assumed that the safety risk determination device is a controller in the elevator system, that is, in the Figure 5 method shown, the safety risk is determined locally. Figure 5 The method shown can also implement Figure 2 step 202 in Figure 5 To avoid repetition, the differences between the embodiment shown below Figure 4 and the embodiment shown in

[0075] Figure 5 The process shown starts from step 501, which can be continued, for example, after Figure 2 step 201 in Figure 5 In step 501, the safety risk determination device generates the state values of the first category features from the operating state data DATA, and when a safety event occurs, it also generates the state values of the second category features based on the safety event. In Figure 5In the illustrated embodiment, the first category of features and the second category of features may have the various features described above, which will not be elaborated herein.

[0076] After completing step 501, Figure 5 the illustrated process proceeds to step 503. In step 503, the security risk determination device performs encoding processing on the status values of the first category of features and the second category of features or the current instances {F1(j)} (j = 1... k), F2. Similar to Figure 4 the illustrated embodiment, a one-hot encoding method may also be used to convert each of the status values of the first category of features and the second category of features into corresponding vectors.

[0077] Subsequently, Figure 5 the illustrated process proceeds to step 503. In step 503, the security risk determination device generates a feature vector (such as the feature vector [χ1, χ2... χ n ) from the status value {F1(j)} (j = 1... k) of the first category of features and the status value F2 of the second category of features.

[0078] After completing step 503, Figure 5 the illustrated process proceeds to Figure 2 step 203 in

[0079] The specific implementation manners of steps 502 and 503 have been described above and will not be elaborated herein.

[0080] Figure 6 FIG. is a schematic block diagram of a device for determining the security risk of an elevator system according to some other embodiments of the present disclosure. Figure 6 The illustrated device may be one of a cloud computing hardware device, a control system for controlling multiple elevator systems, and a controller in an elevator system.

[0081] As Figure 6 shown, the device 60 includes a communication unit 610, one or more memories 620 (such as non-volatile memories such as flash memory, ROM, hard disk drive, magnetic disk, optical disk), one or more processors 630, and a computer program 640.

[0082] The communication unit 610, as a communication interface, is configured to establish a communication connection between the device 60 and an external device (such as various sensors of the elevator system, etc.) or a network (such as the Internet and wireless local area network, etc.).

[0083] The memory 620 stores the computer program 640 executable by the processor 630. In addition, the memory 620 may also store data generated by the processor 630 when executing the computer program 640 and data received from an external device via the communication unit 610 (such as the operating status data of the elevator system, etc.).

[0084] The processor 630 is configured to run the computer program 640 stored on the memory 620 and perform access operations on the memory 620.

[0085] The computer program 640 may include computer instructions for implementing the Figures 2 - 5 described method such that the corresponding method can be implemented when the computer program 640 is run on the processor 630.

[0086] Those skilled in the art will understand that the various illustrative logical blocks, modules, circuits, and algorithmic steps described herein can be implemented as electronic hardware, computer software, or a combination of both.

[0087] To demonstrate the interchangeability between hardware and software, the various illustrative components, blocks, modules, circuits, and steps have been described generally above in terms of their functionality. Whether such functionality is implemented in hardware or software form depends on the particular application and the design constraints imposed on the overall system. Those skilled in the art can implement the described functionality in varying ways for a particular specific application, but such implementation decisions should not be construed as causing a departure from the scope of the present disclosure.

[0088] Although only some specific embodiments of the present disclosure have been described, those of ordinary skill in the art should understand that the present disclosure can be implemented in many other forms without departing from its gist and scope. Therefore, the examples and embodiments shown are regarded as illustrative rather than restrictive, and the present disclosure may cover various modifications and substitutions without departing from the spirit and scope of the present disclosure as defined by the appended claims.

[0089] The embodiments and examples presented herein are provided so as to best illustrate the embodiments in accordance with the present technology and its specific applications, and thereby enable those skilled in the art to implement and use the present disclosure. However, those skilled in the art will know that the above description and examples are provided only for the sake of illustration and exemplification. The description presented is not intended to cover every aspect of the present disclosure or to limit the present disclosure to the precise form disclosed.

Claims

1. A method for determining the safety risk of an elevator system, comprising: A. generating a feature vector based at least on the operating state data of the elevator system, wherein the feature vector includes components corresponding to a combination selected from the state values of a plurality of first - category features, and the state values of the first - category features are determined based on the operating state data; and B. using a neural network model to determine the probability that a transported object is trapped in the car of the elevator system, wherein the feature vector and the probability are the input variable and the output variable of the neural network model respectively.

2. The method according to claim 1, wherein, In step A, the generation of the feature vector is further based on a safety event, and the feature vector further includes components corresponding to a combination of the state values of the first - category features and the state values of second - category features, and the state values of the second - category features are determined based on the safety event.

3. The method according to claim 1, wherein, Further comprising: C. generating an evaluation result regarding the safety risk based on the probability.

4. The method according to claim 1, wherein Further comprising: D. training the neural network model, wherein the operating log of the elevator system is used to label the training samples for training the neural network model.

5. The method according to claim 2, wherein, The operating state data describes various types of operating states of the elevator system, and the state value of each first - category feature includes one of the following state values of one type of the operating states: i) the state value before the occurrence of the safety event; ii) the state value at the time of the occurrence of the safety event; iii) the state value after the occurrence of the safety event.

6. The method according to claim 5, wherein, Each component of the feature vector corresponds to one of the following: i) a combination of the state values of the first - category features; ii) a combination of the state values of the first - category features and the state values of the second - category features.

7. The method according to claim 6, wherein, Step A includes: A1. generating the state values of the plurality of first - category features from the operating state data and generating the state values of the second - category features from the safety event; and A2. generating the feature vector from the state values of the first - category features and the state values of the second - category features.

8. The method according to claim 2, wherein, Step A is executed in response to the occurrence of the safety event.

9. The method according to claim 5, wherein, The types of the operating states include one or more of the following: elevator system operating mode, car movement direction, floor where the car is located, car leveling alignment state, traction machine movement state, car door and landing door opening and closing state.

10. The method according to claim 2, wherein, The types of the safety events include one or more of the following: power failure, sensor failure, door system failure, mechanical component failure, and abnormal elevator system operating mode.

11. The method according to claim 5, wherein, The state values of the first - category features and the state values of the second - category features are represented in a one - hot encoding form, and each component of the vector feature is represented in a binary value form.

12. The method according to claim 1, wherein Step C includes determining the level of the safety risk from the probability based on a pre - set mapping relationship, wherein the mapping relationship defines the probability value range corresponding to each level.

13. The method according to claim 1, wherein The method is implemented by one of the following: cloud computing hardware device, a control system for controlling multiple elevator systems, and a controller in the elevator system.

14. A device for determining the safety risk of an elevator system, comprising: at least one processor; At least one memory; And A computer program stored on the memory, which when run on the processor causes the following operations: A. Generating a feature vector based at least on the operating state data of the elevator system, wherein the feature vector includes components corresponding to a combination selected from state values of a plurality of first-category features, and the state values of the first-category features are determined based on the operating state data; and B. Using a neural network model to determine the probability that a transported object is trapped in the car of the elevator system, wherein the feature vector and the probability are input variables and output variables of the neural network model respectively.

15. The device for determining the safety risk of an elevator system according to claim 14, wherein, In operation A, the generation of the feature vector is further based on a safety event, and the feature vector further includes components corresponding to a combination of the state values of the first-category features and the state values of second-category features, and the state values of the second-category features are determined based on the safety event.

16. The device for determining a safety risk of an elevator system according to claim 14, wherein, The running of the computer program on the processor further causes the following operation: C. Generating an evaluation result regarding the safety risk based on the probability.

17. The apparatus for determining a safety risk of an elevator system according to claim 14, wherein, The running of the computer program on the processor further causes the following operation: D. Training the neural network model, wherein the running log of the elevator system is used to label training samples for training the neural network model.

18. The device for determining the safety risk of an elevator system according to claim 15, wherein, The operating state data describes various types of operating states of the elevator system, and the state value of each first-category feature includes one of the following state values of one type of the operating states: i) the state value before the occurrence of the safety event; ii) the state value at the time of the occurrence of the safety event; iii) the state value after the occurrence of the safety event.

19. The apparatus for determining a safety risk of an elevator system according to claim 18, wherein, Each component of the feature vector corresponds to one of the following: i) a combination of the state values of the first-category features; ii) a combination of the state values of the first-category features and the state values of the second-category features.

20. The apparatus for determining a safety risk of an elevator system according to claim 19, wherein, Operation A includes: A1. Generating the state values of the plurality of first-category features from the operating state data and generating the state values of the second-category features from the safety event; and A2. Generating the feature vector from the state values of the first-category features and the state values of the second-category features.

21. The device for determining a safety risk of an elevator system according to claim 15, wherein, The running of the computer program on the processor causes operation A to be executed in response to the occurrence of the safety event.

22. The apparatus for determining a safety risk of an elevator system according to claim 18, wherein, The types of the operating states include one or more of the following: elevator system operating mode, car movement direction, floor where the car is located, car leveling alignment state, traction machine movement state, opening and closing states of the car door and landing door.

23. The apparatus for determining a safety risk of an elevator system according to claim 18, wherein, The types of the safety events include one or more of the following: power failure, sensor failure, door system failure, mechanical component failure, and abnormal elevator system operating mode.

24. The device for determining the safety risk of an elevator system according to claim 15, wherein, The state values of the first-category features and the state values of the second-category features are represented in a one-hot encoding form, and each component of the vector feature is represented in a binary value form.

25. The device for determining the safety risk of an elevator system according to claim 14, wherein, The execution manner of operation C includes determining the level of the safety risk from the probability based on a preset mapping relationship, wherein the mapping relationship defines the probability value range corresponding to each level.

26. The apparatus for determining a safety risk of an elevator system according to claim 14, wherein, The device for determining the safety risk of the elevator system is one of the following: cloud computing hardware device, control system for controlling multiple elevator systems, and controller in the elevator system.

27. A non-transitory computer-readable storage medium storing instructions therein, characterized in that, The method according to any one of claims 1-13 is implemented by executing the instructions by a processor.

Images