Identification of an elevator emergency operation

Abstract

A method for identifying an elevator emergency operation among a set of elevator emergency operations is provided, the method, performed by a computing apparatus (100), comprises: receiving (710) data descriptive of an operation of at least one entity of an elevator system (800); inputting (720) data derivable from the received data to a machine-learning model (124); and generating (730) an indication descriptive of the elevator emergency operation identified through the classification operation. Also a computing apparatus (100), a computer program, a computer-implemented method for training a machine-learning model and an elevator system are provided.

Description

[0001] IDENTIFICATION OF AN ELEVATOR EMERGENCY OPERATION

[0002] TECHNICAL FIELD

[0003] The invention concerns in general the technical field of elevators. More particularly, the invention concerns safety solution for elevators.

[0004] BACKGROUND

[0005] Elevators are usually provided with a safety system for ensuring safe elevator operation. The safety system may comprise one or more safety devices e.g. in the form of sensors and actuators in order to perform safety management in the elevator. In case a safety-related problem is detected, such as an unaccepted movement of an elevator car or otherwise abnormal elevator operation, the safety system may be configured to carry out an emergency operation to bring the elevator to a safe state. For example, the safety system may execute an emergency stop of the elevator car by activating one or more mechanical hoisting machinery brakes and / or an elevator safety gear.

[0006] Commonly known safety system in elevators is based on so-called safety chain wherein safety switches are used for activating / inactivating various entities of the elevator in accordance with an operation of the elevator.

[0007] Generally speaking it is important to ensure that the safety system is functioning correctly and any elevator emergency operations do not remain unnoticed. Therefore, there is room for introducing novel and sophisticated approaches in order to improve safety of the elevators.

[0008] SUMMARY

[0009] The following presents a simplified summary in order to provide basic understanding of some aspects of various invention embodiments. The summary is not an extensive overview of the invention. It is neither intended to identify key or critical elements of the invention nor to delineate the scope of the invention. The following summary merely presents some concepts of the invention in a simplified form as a prelude to a more detailed description of exemplifying embodiments of the invention.

[0010] An object of the invention is to present a method, a computing apparatus, a computer program and an elevator system for identifying an elevator emergency operation. Another object of the invention is to present a computer-implemented method for training a machine-learning model for identifying an elevator emergency operation.

[0011] The objects of the invention are reached by a method, a computing apparatus, a computer program, an elevator system and a computer-implemented method for training a machine-learning model as defined by the respective independent claims.

[0012] According to a first aspect, a method for identifying an elevator emergency operation among a set of elevator emergency operations is provided, the method, performed by a computing apparatus, comprises: receiving data descriptive of an operation of at least one entity of an elevator system; inputting data derivable from the data descriptive of the operation of at least one entity of the elevator system to a machine-learning model, the machine-learning model is trained to identify the elevator emergency operation among the set of elevator emergency operations through a classification operation; generating an indication descriptive of the elevator emergency operation identified through the classification operation.

[0013] The data descriptive of the operation of the at least one entity of the elevator system may be received from at least one sensor associated to an elevator car, the at least one sensor is configured to detect at least one physical quantity descriptive on a movement of the elevator car. The method may further comprise: converting the data expressed with first physical quantity to a second data expressed with another physical quantity to perform the classification operation by: classifying the data expressed with the first physical quantity to a plurality of categories based on the second data, and identifying the elevator emergency operation based on data ended up to each category.

[0014] Moreover, the physical quantity descriptive of the movement of the elevator car may be at least one of the following: an acceleration, a speed.

[0015] A position of the elevator car may be applied as the physical quantity in setting the data expressed with first physical quantity to the plurality of categories.

[0016] Further, the data descriptive of the operation of the at least one entity of the elevator system may be generated with a testing procedure of the elevator system.

[0017] Alternatively, the data descriptive of the operation of the at least one entity of the elevator system may be generated during normal elevator operation.

[0018] In accordance with the method, the set of elevator emergency operations may comprise at least one of: a safety gear operation; an emergency stop operation; an operation to prevent unintended move of the elevator car (810); an operation to detect an elevator car buffer run.

[0019] The method may further comprise, in response to the identification of the elevator emergency operation, a generation of a value of at least one operational parameter based on an analysis of the data descriptive of the operation of at least one entity of the elevator system.

[0020] According to a second aspect, a computing apparatus for identifying an elevator emergency operation among a set of elevator emergency operations is provided, the computing apparatus is configured to: receive data descriptive of an operation of at least one entity of an elevator system; input data derivable from the data descriptive of the operation of at least one entity of the elevator system to a machine-learning model, the machine-learning model is trained to identify the elevator emergency operation among the set of elevator emergency operations through a classification operation; generate an indication descriptive of the elevator emergency operation identified through the classification operation.

[0021] The computing apparatus may e.g. be configured to receive the data descriptive of the operation of the at least one entity of the elevator system from at least one sensor associated to an elevator car, the at least one sensor is configured to detect at least one physical quantity descriptive on a movement of the elevator car.

[0022] The computing apparatus may further be configured to: convert the data expressed with first physical quantity to a second data expressed with another physical quantity to perform the classification operation by: classifying the data expressed with the first physical quantity to a plurality of categories based on the second data, and identifying the elevator emergency operation based on data ended up to each category.

[0023] The computing apparatus may be configured to use as the physical quantity descriptive of the movement of the elevator car at least one of the following: an acceleration, a speed.

[0024] Further, the computing apparatus may be configured to apply a position of the elevator car as the physical quantity in setting the data expressed with first physical quantity to the plurality of categories.

[0025] The computing apparatus may also be configured to generate the data descriptive of the operation of the at least one entity of the elevator system with a testing procedure of the elevator system.

[0026] Alternatively, the computing apparatus may be configured to generate the data descriptive of the operation of the at least one entity of the elevator system during normal elevator operation.

[0027] Still further, the computing apparatus may be configured to apply as the set of elevator emergency operations comprise at least one of: a safety gear operation; an emergency stop operation; an operation to prevent unintended move of the elevator car; an operation to detect an elevator car buffer run.

[0028] The computing apparatus may further be configured to, in response to the identification of the elevator emergency operation, generate at least one operational parameter based on an analysis of the data descriptive of the operation of at least one entity of the elevator system.

[0029] According to a third aspect, a computer program is provided, the computer program comprising instructions to cause the computing apparatus according to the second aspect as defined above to execute the steps of the method according to the first aspect as defined above. According to a fourth aspect, a computer-implemented method for training a machine-learning model for identifying an elevator emergency operation among a set of elevator emergency operations is provided, the method, performed by a computing apparatus, comprises: receiving simulation data from a simulation model of at least one entity of an elevator system, the simulation data being descriptive of the set of elevator emergency operations; training the machine-learning model with at least part of the simulation data descriptive of an operation of at least one entity of an elevator system.

[0030] The at least part of the simulation data may be descriptive of at least one of the following: a safety gear operation; an emergency stop operation; an operation to prevent unintended move of the elevator car; an operation to detect an elevator car buffer run.

[0031] According to a fifth aspect, an elevator system is provided, the elevator system comprising a computing apparatus according to the second aspect as defined above.

[0032] The expression "a number of” refers herein to any positive integer starting from one, e.g. to one, two, or three.

[0033] The expression "a plurality of” refers herein to any positive integer starting from two, e.g. to two, three, or four.

[0034] Various exemplifying and non-limiting embodiments of the invention both as to constructions and to methods of operation, together with additional objects and advantages thereof, will be best understood from the following description of specific exemplifying and non-limiting embodiments when read in connection with the accompanying drawings.

[0035] The verbs “to comprise” and “to include” are used in this document as open limitations that neither exclude nor require the existence of unrecited features. The features recited in dependent claims are mutually freely combinable unless otherwise explicitly stated. Furthermore, it is to be understood that the use of “a” or “an”, i.e. a singular form, throughout this document does not exclude a plurality.

[0036] BRIEF DESCRIPTION OF FIGURES

[0037] The embodiments of the invention are illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings.

[0038] Figure 1 illustrates schematically a computing apparatus according to an example.

[0039] Figure 2 illustrates schematically a training mechanism of a machine-learning model according to an example.

[0040] Figures 3A and 3B illustrate schematically aspects relating to safety gear operation as an emergency operation according to an example.

[0041] Figures 4A and 4B illustrate schematically aspects relating to emergency stop operation as an emergency operation according to an example.

[0042] Figures 5A and 5B illustrate schematically aspects relating to unintended movement operation as an emergency operation according to an example.

[0043] Figures 6A and 6B illustrate schematically aspects relating to elevator car buffer run as an emergency operation according to an example.

[0044] Figure 7 illustrates schematically a method according to an example.

[0045] Figure 8 illustrates schematically an elevator system according to an example.

[0046] DESCRIPTION OF THE EXEMPLIFYING EMBODIMENTS

[0047] The specific examples provided in the description given below should not be construed as limiting the scope and / or the applicability of the appended claims. Lists and groups of examples provided in the description given below are not exhaustive unless otherwise explicitly stated.

[0048] The present invention is for performing safety operation in an elevator. The invention may be implemented with a computing apparatus configured to perform a predefined method to identify one or more predefined situations in an operation of the elevator. The predefined situations in the operation of the elevator are called as elevator emergency operations herein.

[0049] A computing apparatus configured to perform the predefined method is schematically illustrated in Figure 1. For sake of clarity, it is worthwhile to mention that the block diagram of Figure 1 depicts some components of an entity that may be employed to implement a functionality of the computing apparatus 110. The computing apparatus 100 of Figure 1 comprises a processing unit 110 comprising one or more processors. The computing apparatus 110 also comprises a memory 120 storing various types of data, such as one or more computer programs executable with the processing unit. The computer programs are referred with references 122 and 124 in Figure 1. The computer program indicated with the reference 124 is a machine-learning model that is executable by a processor of the processing unit 110. The machine-learning model 124 refers to a computer program that is trained to execute a task as is described in the forthcoming description. Moreover, the computing apparatus 100 may comprise a communication interface 130, such as a wireless communication interface or a communication interface for wired communication, or both to communicate with other entities as described. The communication interface 130 may thus comprise one or more modems, antennas, and any other hardware and software for enabling an execution of the communication e.g. under control of the processing unit 110. Furthermore, I / O (input / output) components may be arranged, together with the processing unit 110 and a portion of the computer program code to provide a user interface for receiving input from a user, such as from a technician, and / or providing output to the user of the apparatus when necessary. In particular, the I / O components may include user input means, such as one or more keys or buttons, a keyboard, a touchscreen, or a touchpad, etc. The I / O components may include output means, such as a loudspeaker, a display, or a touchscreen. The components of the computing apparatus 100 may be communicatively connected to each other via data bus that enables transfer of data and control information between the components.

[0050] Moreover, the computing apparatus 100 may be equipped with a number of sensors 140 selected in accordance with a number of physical quantities desired to be measured. Thus, the computing apparatus 100 may house the number of sensors 140 or the sensors 140 may be separate to a housing of the computing apparatus 100 and communicatively connected, e.g. through the communication interface 130, with the other entities of the computing apparatus 100 so as to enable the computing apparatus 100 receiving measurement data obtained with the number of sensors 140. In accordance with the invention the sensors are preferably such that they may be used to measure a movement of an object they are arranged to measure, such as an elevator car into which they are associated to. For example, the type of the sensors 140 may be accelerometers and magnetometers for measuring elevator car vibration and elevator position in an elevator shaft, but also other types of sensors 140 may be included in the setup.

[0051] The memory 120 and at least a portion of the computer program 122, 124 stored therein may further be arranged, with the processing unit 110, to cause the apparatus to perform at least a portion of any of the methods as is described herein. The processing unit 110 may be configured to read from and write to the memory 120. Although the processing unit 1 10 is depicted as a respective single component, it may be implemented as respective one or more separate processing components. Similarly, although the memory 120 is depicted as a respective single component, it may be implemented as respective one or more separate components, some, or all of which may be integrated I removable and I or may provide permanent / semi-permanent / dynamic / cached storage. The computer program 122, 124 may comprise computer-executable instructions that implement functions that correspond to steps implemented in the method when loaded into the processing unit 110 of the respective computing apparatus 100. As an example, the computer program 122, 124 may include a computer program consisting of one or more sequences of one or more instructions. The processing unit 110 is able to load and execute the computer program by reading the one or more sequences of one or more instructions included therein from the memory 120. The one or more sequences of one or more instructions may be configured to, when executed by the processing unit 110, cause the computing apparatus 100, such as a computer, to perform at least some of the method steps as described in the forthcoming description. Hence, the computing apparatus 100 may comprise at least one processing unit 110 and at least one memory 120 including the computer program 122, 124 for one or more programs, the at least one memory 120 and the computer program 122, 124 implemented with a computer program code configured to, with the at least one processing unit 110, cause the computing apparatus 100 to perform at least some of the methods.

[0052] The computer program 122, 124, or at least some portion of it, may be provided e.g. a computer program product comprising at least one computer-readable non-transitory medium having the computer program 122, 124 stored thereon, which computer program 122, 124, when executed by the processing unit 110 causes the apparatus to perform the method. The computer-readable non- transitory medium may comprise a memory device or a record medium, such as a CD-ROM, a DVD, a Blu-ray disc, or another article of manufacture that tangibly embodies the computer program. As another example, the computer program may be provided as a signal configured to reliably transfer the computer program.

[0053] Still further, the computer program 122, 124 may comprise a proprietary application, such as computer program code for causing an execution of the method in the manner as described in the description herein. Any of the programmed functions mentioned may also be performed in firmware or hardware adapted to or programmed to perform the necessary tasks.

[0054] As mentioned, the computer program code may comprise a portion that represents the machine-learning model as referred with the reference 124 in Figure 1 . The machine-learning model may be trained in a manner as shown in Figure 2. In accordance with the invention the training data for training the machine-learning model to identify an emergency operation of an elevator is performed so that a simulation model 210 corresponding to the elevator with a required accuracy is established. In other words, the simulation model 210 refers to a virtual digital model, i.e. computer-implemented model also callable as a digital twin, corresponding to the elevator whose operation is to be monitored with a help of the trained machine-learning model. In other words, the simulation model 210 may be an exact model of the elevator in question or a model that corresponds, e.g. in structure and / or in its operational sense, to the elevator at least in an extent that the machine-learning model 124 is achieved to operate with a required accuracy. Now, the simulation model 210 is executed with various input parameters and / or in any other manner so as to generate training data to train the machine-learning model executed. Thus, the operation of the elevator is simulated with the simulation model 210 in a various manner. For example, the training data may be generated so that the simulation model 210 is executed with parameters causing the simulation model 210 to follow a normal operation of the elevator and / or its entities therein. Alternatively or in addition, the simulation model 210 may be executed with parameters causing the simulation model 210 to operate as the elevator operates in a number of emergency situations referred with 220 in Figure 2 may be generated in order to generate training data corresponding to the respective emergency operations 220. In other words, the data, i.e. training data, generated with the simulation model 210 defines at a desired accuracy the emergency situations and / or also normal situations of the elevator corresponding to the simulation model 210. In case there is generated training data for a plurality of emergency operations 220 the data descriptive of a certain emergency operation is advantageously provided with an identifier of the respective emergency operation in order to create a classification model in the machine-learning model, and, thus, to enable a differentiation of various situation with the machine-learning model 124. In other words, the machine-learning model 124 is arranged to be setup, through training, so that it is capable to output information identifying emergency situation(s) in response to an information input to the model 124.

[0055] For avoidance of any doubt it is worthwhile to mention that in some approaches the training of the machine-learning model may apply further data in addition to the simulation data generated with the simulation model 210. Namely, there may be available history data generated e.g. by one or more existing elevators corresponding with a predefined accuracy the one for which the machinelearning model 124 is trained for. The history data may refer to an operational data of the existing elevator which is received from the system data of the elevator or from a sensor arrangement arranged to measure at least some operational parameters of the existing elevator.

[0056] In other words, the training of the machine-learning model, i.e. the machinelearning algorithm is performed with the training data, aka. training model, generated as described. As mentioned, the training data consists of the sample output data and the corresponding sets of input data that have an influence on the output and the machine-learning algorithm is executed to run the input data to correlate the processed output against the sample output. The result from this correlation is used to modify the model. The process itself may be iterative to generate a machine-learning model that is accurate enough for the intended task. For example, the training procedure may be supervised, unsupervised or semi-supervised learning.

[0057] Without saying it is clear that the generation of the simulation data, i.e. the training data, and the training of the machine-learning model in the described manner as schematically illustrated in Figure 2 as a flow chart is a computer- implemented process. The computing entity applied for the simulation and training is selected so that it is provided with necessary computing resources to perform the task. The computing entity may e.g. be a standalone computer or a network of computers harnessed to execute distributed computing. In some embodiments, the computing entity may be the computing apparatus 100. However, if the computing entity applied to the training is separate to the computing apparatus 100, the trained machine-learning model 124 is transferred to the computing apparatus 100 with any known data transfer mechanisms, such as data transfer over an applied communication channel or by storing the trained machine-learning model 124 to a computer-readable non-transitory medium, such as a memory device or a record medium, to transport the model 124 from the computing entity applied for training, and possibly also for simulation, to the computing apparatus 100.

[0058] The simulation data descriptive of emergency operations 220 of the elevator may e.g. relate to the following situations:

[0059] • Safety gear operation

[0060] • Emergency stop operation

[0061] • Operation to detect an unintended move of the elevator car

[0062] • Operation to detect an elevator car buffer run.

[0063] The safety gear operation relates to a procedure that allows a braking of the elevator in case of intervention of so-called overspeed governor that is activated when there are anomalies during the elevator travel. In other words, in case abnormal acceleration and / or speed is detected, the overspeed governor intervenes by giving an instruction to the safety gear function to initiate the safety operation. Thus, the training data is generated with the simulation model so that it comprises a set of values for an acceleration of the elevator car within a predefined portion of the elevator shaft in terms of the position wherein the position is defined between the bottom-most landing and the top-most landing. For example, it may generate peak values for acceleration that are at least 20 times higher than a nominal value defined for the acceleration of the elevator type in question. For example, the nominal acceleration value may be less than 0,5 m / s2. Thus, the training data for training the machine-learning model 124 to detect the safety gear operation may correspond to data shown in Figures 3A and 3B. Thus, it may be generated simulation data for different loads of the elevator car, such as for an empty car and for a full car since there are some difference in the acceleration curves in the respective cases. Naturally, the training data may comprise a huge number of the curves as shown in Figures 3A and 3B generated with slightly different input parameters for the simulation model 210. As said, the simulation is performed within the defined portion of the elevator shaft i.e. the training data may comprise position data corresponding to the portion of the elevator shaft.

[0064] The emergency stop operation, in turn, relates to a procedure that causes stopping of the elevator car withing a predefined period of time. In accordance with an embodiment the training data is generated with the simulation model so that it resembles with the emergency stop wherein the characteristics may e.g. be that the elevator car resides between a first landing floor (e.g. a ground floor or any other lowest floor) and the top-most landing floor. As regards to the acceleration, or deceleration, the training data may indicate that the acceleration changes in two steps. During a first period of time, such as during 200 ms, the deceleration is achieved with the elevator drive system and after that machinery brake system is activated causing a predefined deceleration being e.g. 5 times higher that a nominal acceleration, i.e. a deceleration, but also less than 20 times the nominal acceleration value. As already mentioned, the nominal acceleration, or deceleration, may refer to a value of 0,5 m / s2. Since the emergency stop causes an acceleration ramp, i.e. the acceleration oscillates, it is possible to detect from the direction of the first peak if the elevator car was traveling upwards (negative first peak) or downwards (positive first peak) at the instant of the emergency stop. These are shown in Figures 4A (travel direction upwards) and 4B (travel direction downwards) for exemplifying purposes. The training data generated with the simulation may also take into account different loads in the elevator car to at least some extent. The Figure 4A and 4B schematically illustrate the acceleration ramps with an empty car wherein the first peak is generated in response to an activation of hoisting machinery brakes. Prior to the first peak a very small ramp is detectable which results from the elevator drive control causing a braking movement to the elevator car with the hoisting motor (cf. marking “Emergency stop starts” in Figures 4A and 4B). Again, the training data may be generated with the simulation model 210 for various starting points and for various situations within the framework as provided in the foregoing description.

[0065] As it comes to an operation to detect an unintended movement of the elevator car the training data may be generated with the simulation model 210 so that an acceleration of the elevator car is limited to a peak value of being at least five times higher than the nominal value. The value results from an activation of so- called unintended car movement protection. Further, the simulation model 210 may be defined to operate so that the nominal acceleration value is again less than 0,5 m / s2and also that the peak value of the acceleration is detected in a predefined position, such as from -1 ,2 m to +1 m, from a landing zone. Furthermore, the acceleration peak is identified within a predefined instant of time after the elevator car moves from the landing zone, such as with 2 seconds from the initiation of the movement. With the described parameters the simulation model may be executed to generate training data for the machinelearning model 124 to identify the unintended movement of the elevator car. Figures 5A and 5B illustrate schematically curves descriptive of the acceleration and the position of the elevator car both when the elevator car is moving downwards (Figure 5A) and when the elevator car is moving upwards (Fig. 5B). The curves may be considered as the training data generated with the simulation model 210. Again, the simulation may be performed with various loads in the elevator car and so on so as to train the machine-learning model 124 to perform the identification of an emergency operation 220 corresponding to the unintended movement of the elevator car.

[0066] The operation to detect an elevator car buffer run relates to a rare situation wherein the elevator car travels against buffers installed at a bottom of an elevator shaft. Thus, in order to generate training data the simulation model 210 is executed so that the simulation occurs in a situation that the elevator car travels below the bottom-most landing, such as the ground floor, and the acceleration, i.e. the deceleration, of the elevator car is at least 20 times higher than a nominal value defined for the acceleration of the elevator type in question. For example, the nominal acceleration value may be less than 0,5 m / s2Hence, the simulation of the buffer run differs from the simulation of the safety gear operation at least in that the position of the elevator car is below the bottommost landing in the simulation of the buffer run whereas the position of the elevator car in the simulation of the safety gear operation is between the bottommost and the top-most landing even if the acceleration values correspond to each other. Figures 6A and 6B schematically illustrate acceleration curves according to an example for an empty car (Figure 6A) and for a full car (Figure 6B) which may be generated with the simulation model as to be applied as the training data for the machine-learning model 124 with the position data as defined. As a result, the trained machine-learning model 124 may detect the car buffer run from the operational data of the elevator if the described patterns exists in the measurement data comprising both acceleration and position data.

[0067] The above-given examples of the emergency operations provides insight on how the machine-learning model 124 may be trained by simulation and especially by generating training data that addresses acceleration with or without position of the elevator car.

[0068] Next, further aspects in relation to the computing apparatus 100 are described wherein the computing apparatus 100 is provided with the trained machinelearning model 124. The computing apparatus 100 is coupled with an elevator car, e.g. by mounting it on a roof of the elevator car as a device configured to analyse the operation of the elevator. The computing apparatus 100 is configured to perform a method as schematically illustrated in Figure 7. In step 710, the computing apparatus 100 receives 710 data that is descriptive of an operation of at least one entity of an elevator system. The receipt of data may refer to an operation in which the one or more sensors 140 are configured to generate measurement data descriptive of the operation of the at least one entity wherein the operation may refer to a movement of the elevator car in the elevator shaft. The measurement data may be one or more of the following: acceleration data, speed data, position data. The mentioned types of data may be obtained from applicable sources, such as from sensors 140 dedicated to measure the defined physical quantities either directly or indirectly. The position data of the elevator car in the elevator shaft may be received from other sources, such as from the encoder of the electric motor, over the communication interface 130 of the computing apparatus 100. Naturally, the encoder may be considered as one type of sensor.

[0069] In response to the receipt 710 of the data as described the computing apparatus 100 may be configured to manipulate the received data in a predefined manner. The manipulation may refer to a conversion of the data descriptive of the operation of the entity, such as the movement of the elevator car, applicable for the machine-learning model. For example, the computing apparatus 100 may generate position data from acceleration data in known mathematical methods in order to enable classification of the data descriptive of the movement of the elevator car, such as the acceleration, i.e. the first data descriptive of the movement of the elevator car, to a plurality of categories on a basis of the position data. Corresponding conversions between an acceleration, a speed and a position may be applied to the received data also taking into account time information in relation to the movement where necessary. This allows the analyse the data in the plurality of categories in view of the first, or original, physical quantity in order to identify the elevator emergency operation.

[0070] Next, the computing apparatus 100, i.e. the processing unit 110 of the computing apparatus 100, is configured to input 720 the data derivable from the data descriptive of the operation of at least one entity of the elevator system to the machine-learning model 124 as described. The data derivable from the data descriptive of the operation of the at least one entity shall be understood at least to cover e.g. the original data received from the sensor(s), any derivation of it or a combination of the original measurement data with any other data, such as with data derived from the original measurement data. As described in the foregoing description the machine-learning model 124 is trained to identify an elevator emergency operation among the set of elevator emergency operations through a classification operation wherein the identification is performed on the basis of the input data. For example, the classification operation within the machine-learning model 124 may be based on taking the position data into account when selecting the class for the emergency operation in question in view of the acceleration data of the entity, e.g. the elevator car. Thus, in accordance with the invention the computing apparatus 100 performs an analysis with the trained machine-learning model 124 to detect, or identify, the data descriptive of the operation of the elevator system corresponds to an emergency operation 220 the machine-learning model 124 is trained to identify.

[0071] Finally, in the step 730 the computing apparatus 100 is configured to generate 730 an indication descriptive of the elevator emergency operation 220 identified through the classification operation. The generation of the indication may e.g. comprise a generation of a signal, or a message, that carries data indicative of the identified emergency operation if any. The generation 730 of the signal may also comprise a transmit of the signal to a predefined destination, such as to a remote service center configured to manage the elevator in question. The recipient of the signal may initiate further actions with respect to the situation which may e.g. comprise requesting a maintenance personnel to visit the site and / or requesting medical help to the site over predefined communication channels and so on.

[0072] In accordance with some example embodiments the computing apparatus 100 may further be configured to analyse a performance of one or more entities behind the emergency operation identified by the trained machine-learning model 124. This may refer to an approach in which, in response to the detection of the emergency operation, the computing apparatus 100 is configured to analyse the measurement data received from the elevator system in order to generate data descriptive of the performance of the one or more entities on which it is possible to generate information from the measurement data. As a non-limiting example, if the computing apparatus 100 detects that an emergency stop has occurred, the measurement data may be arranged to analyse so that an instant of time when the braking by the machinery brakes is identified as well as a travel distance from the activation of the braking until the stop of the movement of the elevator car. Based on analysis of the measurement data it may e.g. be determined whether the braking effect is achieved too slowly (such that reaction time of the machinery brakes may be too slow) and / or whether the travel distance during the braking is too long, and this may be informed to a responsible entity, such as to a technician, over messaging from the computing apparatus 100 e.g. to a data center and / or to a terminal device of the technician. Furthermore, with respect to a detection of the unintended movement of the elevator car by the machine-learning model 124 an analysis may be performed to a measurement data descriptive of a travel distance of the elevator car withing a door zone of the elevator prior to the stop of the elevator car when the safety operation for a prevention of the unintended movement is activated (i.e. the machinery brakes are instructed to brake). If the travel distance in the door zone is too big, a notification may be generated to indicate an unacceptable operation of the machinery brakes. When it comes to a detection of the car buffer run with the machine-learning model, an acceleration data as the measurement data may be analysed with an aim to detect whether a deceleration in the car buffer run is too large and if this is the case, necessary corrective actions may be initiated. The above-given approaches are non-limiting examples of a further operation according to at least some embodiments of the present invention in which, in response to the identification of the elevator emergency operation, a generation of a value of at least one operational parameter based on an analysis of the data, i.e. the measurement data, descriptive of the operation of at least one entity of the elevator system. The operational parameter may be utilized in evaluating a condition of the at least one entity and in maintaining the respective entity, and / or the elevator in general.

[0073] The computing apparatus 100 implementing the method as described may be applied in performing safety tests during an installation of the elevator or at least during a handover process of the elevator to the customer, but also in normal operation of the elevator. The “normal operation” here refers to transferring passengers and / or cargo between landing floors in accordance with service requests issued by elevator passengers. One advantage of the invention is that the computing apparatus 100 may be configured to operate as a stand-alone entity by arranging power to the apparatus 100 e.g. with a battery or by coupling the apparatus 100 to electrical network through the wiring of the elevator car, for example. Thus, it is not necessary to communicatively couple the computing apparatus 100 to a control portion of the elevator system in order to receive any data from the elevator system, but the computing apparatus 100 may be configured to perform alone in its task. This allows a utilization of the computing apparatus 100 with elevators into which access is limited in terms of obtaining operational data therefrom. In this kind of approach the computing apparatus 100 may be configured to communicate with external entities by implementing at least one predefined wireless communication technology over the communication interface 130 of the apparatus 100. Moreover, the measurement data, in response to the detection of the emergency operation, may be used for evaluating an operation of one or more entities of the elevator system and for initiating corrective actions if seen necessary.

[0074] For sake of completeness some aspects it is worthwhile to mention that some aspects of the present invention relate to an elevator system 800 as schematically illustrated in Figure 8 as a non-limiting example. The computing apparatus 100 is associated with an elevator car 810 of the elevator system 800. The elevator system 800 comprises further entities and functionalities than those shown in Figure 8 in order to make it operable. However, they are known from elevator systems according to prior art and not discussed in more detail herein. The specific examples provided in the description given above should not be construed as limiting the applicability and / or the interpretation of the appended claims. Lists and groups of examples provided in the description given above are not exhaustive unless otherwise explicitly stated.

Claims

WHAT IS CLAIMED IS:

1. A method for identifying an elevator emergency operation among a set of elevator emergency operations, the method, performed by a computing apparatus (100), comprises: receiving (710) data descriptive of an operation of at least one entity of an elevator system (800), inputting (720) data derivable from the data descriptive of the operation of at least one entity of the elevator system (800) to a machine-learning model (124), the machine-learning model (124) is trained to identify the elevator emergency operation among the set of elevator emergency operations through a classification operation, generating (730) an indication descriptive of the elevator emergency operation identified through the classification operation.

2. The method according to claim 1 , wherein the data descriptive of the operation of the at least one entity of the elevator system (800) is received from at least one sensor (140) associated to an elevator car (810), the at least one sensor (140) is configured to detect at least one physical quantity descriptive on a movement of the elevator car (810).

3. The method according to claim 2, the method further comprises: converting the data expressed with first physical quantity to a second data expressed with another physical quantity to perform the classification operation by: classifying the data expressed with the first physical quantity to a plurality of categories based on the second data, andidentifying the elevator emergency operation based on data ended up to each category.

4. The method according to claim 2 or 3, wherein the physical quantity descriptive of the movement of the elevator car (810) is at least one of the following: an acceleration, a speed.

5. The method according to any of claims 2 to 4, wherein a position of the elevator car (810) is applied as the physical quantity in setting the data expressed with first physical quantity to the plurality of categories.

6. The method according to any of the preceding claims, wherein the data descriptive of the operation of the at least one entity of the elevator system (800) is generated with a testing procedure of the elevator system (800).

7. The method according to any of claims 1 -5, wherein the data descriptive of the operation of the at least one entity of the elevator system (800) is generated during normal elevator operation.

8. The method according to any of the preceding claims, wherein the set of elevator emergency operations comprises at least one of: a safety gear operation; an emergency stop operation; an operation to prevent unintended move of the elevator car (810); an operation to detect an elevator car (810) buffer run.

9. The method according to any of the preceding claims, the method further comprises, in response to the identification of the elevator emergency operation, a generation of a value of at least one operational parameter based on an analysis of the data descriptive of the operation of at least one entity of the elevator system (800).

10. A computing apparatus (100) for identifying an elevator emergency operation among a set of elevator emergency operations, the computing apparatus (100) is configured to:receive data (710) descriptive of an operation of at least one entity of an elevator system (800), input (720) data derivable from the data descriptive of the operation of at least one entity of the elevator system (800) to a machine-learning model (124), the machine-learning model (124) is trained to identify the elevator emergency operation among the set of elevator emergency operations through a classification operation, generate (730) an indication descriptive of the elevator emergency operation identified through the classification operation.

11. The computing apparatus (100) according to claim 10, wherein the computing apparatus (100) is configured to receive the data descriptive of the operation of the at least one entity of the elevator system (800) from at least one sensor (140) associated to an elevator car (810), the at least one sensor (140) is configured to detect at least one physical quantity descriptive on a movement of the elevator car (810).

12. The computing apparatus (100) according to claim 11 , the computing apparatus (100) further configured to: convert the data expressed with first physical quantity to a second data expressed with another physical quantity to perform the classification operation by: classifying the data expressed with the first physical quantity to a plurality of categories based on the second data, and identifying the elevator emergency operation based on data ended up to each category.

13. The computing apparatus (100) according to claim 11 or claim 12, wherein the computing apparatus (100) is configured to use as the physical quantitydescriptive of the movement of the elevator car (810) at least one of the following: an acceleration, a speed.

14. The computing apparatus (100) according to any of claims 11 to 13, wherein the computing apparatus (100) is configured to apply a position of the elevator car (810) as the physical quantity in setting the data expressed with first physical quantity to the plurality of categories.

15. The computing apparatus (100) according to any of the claims 10 to 14, wherein the computing apparatus (100) is configured to generate the data descriptive of the operation of the at least one entity of the elevator system (800) with a testing procedure of the elevator system (800).

16. The computing apparatus (100) according to any of claims 10 to 14, wherein the computing apparatus (100) is configured to generate the data descriptive of the operation of the at least one entity of the elevator system (800) during normal elevator operation.

17. The computing apparatus (100) according to any of the claims 10 to 16, wherein the computing apparatus (100) is configured to apply as the set of elevator emergency operations comprise at least one of: a safety gear operation; an emergency stop operation; an operation to prevent unintended move of the elevator car (810); an operation to detect an elevator car (810) buffer run.

18. The computing apparatus (100) according to any of the claims 10 to 17, the computing apparatus (100) is further configured to, in response to the identification of the elevator emergency operation, generate at least one operational parameter based on an analysis of the data descriptive of the operation of at least one entity of the elevator system (800).

19. A computer program comprising instructions to cause the computing apparatus (100) of claim 10 to execute the steps of the method of claim 1 .

20. A computer-implemented method for training a machine-learning model (124) for identifying an elevator emergency operation among a set of elevator emergency operations, the method, performed by a computing apparatus, comprises: receiving simulation data from a simulation model (210) of at least one entity of an elevator system (800), the simulation data being descriptive of the set of elevator emergency operations, training the machine-learning model (124) with at least part of the simulation data descriptive of an operation of at least one entity of an elevator system (800).

21. The method according to claim 20, wherein the at least part of the simulation data is descriptive of at least one of the following: a safety gear operation; an emergency stop operation; an operation to prevent unintended move of the elevator car (810); an operation to detect an elevator car (810) buffer run.

22. An elevator system (800) comprising a computing apparatus (100) according to any of claims 10 to 18.

Images