Method for operating a sensor arrangement and sensor arrangement and apparatus for data processing and device

Abstract

The present invention relates to a method for operating a sensor arrangement. The present invention also relates to an apparatus for data processing, which is suitable for carrying out such a method. In addition, the invention relates to a sensor arrangement which is suitable for being used in such a method and / or interacting with such an apparatus, as well as to a device comprising such a sensor arrangement and / or such an apparatus.

Description

[0001] Method for operating a sensor arrangement and sensor arrangement and device for data processing and apparatus

[0002] field of technology

[0003] The present invention relates to a method for operating a sensor arrangement. The present invention also relates to a data processing device adapted to carry out such a method. Furthermore, the invention relates to a sensor arrangement adapted to be used in such a method and / or to cooperate with such a device, as well as to a device comprising such a sensor arrangement and / or such a device.

[0004] State of the art

[0005] Processes, especially in industrial environments, are increasingly dependent on the availability of sensor data. Such sensor data can be recorded, for example, by the sensors of a sensor array and used to control and / or regulate devices such as machines.

[0006] If a defect in one or more sensors in such a sensor array results in invalid sensor data being processed, or perhaps no sensor data at all, this can lead to significant disruptions in the process flow. In the worst case, this could result in the shutdown of an entire factory or parts of it, which can have far-reaching economic consequences.

[0007] Reliable operation of sensor arrays is therefore of great importance. Ensuring reliable operation, however, is becoming increasingly complex due to the wide variety of different types of sensor arrays and the growing number of sensors per sensor array.

[0008] Summary of the invention

[0009] It is therefore an object of the present invention to overcome the described disadvantages of the prior art and in particular to provide means by which sensor arrangements can be operated in a reliable and simple but nevertheless cost-effective manner.

[0010] The object is achieved by the invention according to a first aspect by a method for operating a sensor arrangement with a plurality of sensors, wherein sensor data is received from each sensor of the plurality of sensors as received sensor data of the respective sensor and is provided as source sensor data of the respective sensor for further data processing. If an error condition is detected in connection with at least one specific sensor of the plurality of sensors, replacement sensor data is then determined for the specific sensor and provided as source sensor data of the specific sensor instead of received sensor data of the specific sensor.

[0011] The substitute sensor data is determined using at least one first trained data model with the support of machine learning, wherein received sensor data and / or source sensor data of at least one sensor selected as an auxiliary sensor from the plurality of sensors, which is not the specific sensor, and / or data based thereon are used as input data for the first trained data model of the machine learning. The invention is therefore based on the surprising finding that a process that relies on the sensor data of the sensors of a sensor arrangement can be reliably continued even during a failure, temporary failure, or temporary disruption of a sensor of the arrangement and / or a transmission channel used to transmit the sensor data if the actual sensor data of the affected sensor is at least temporarily replaced by synthetic sensor data.

[0012] This allows for a reliable bridge to the time until the sensor data from a specific sensor is properly available again. For example, this can be a time period until the sensor is repaired or replaced with a new one and / or until the transmission channel for the sensor data is functioning properly again.

[0013] In this way, the ongoing process can continue without or with only minimal interruption if the sensor data from the affected sensor is faulty or even completely missing. This makes it possible to operate the sensor array with increased reliability and thus reliably avoid or at least reduce downtimes of the process that relies on the sensor data.

[0014] It was surprisingly found that the synthetic data can be determined particularly reliably if they are determined at least on the basis of the sensor data of at least some of the other sensors in the arrangement in conjunction with a trained machine learning data model.

[0015] In other words, the specific (defective) sensor is temporarily replaced by a virtual sensor in the form of a trained machine learning data model, and the synthetic data of this virtual sensor is made available for processing and used as a substitute for the actual sensor data of the specific sensor.

[0016] In other words, it was primarily recognized that the real defective sensor (the specific sensor) can be replaced, so to speak, by a virtual sensor, so that the virtual sensor can continue to function in place of the real sensor until the specific sensor is replaced or repaired. The virtual replacement sensor is then based on a machine learning data model.

[0017] It has proven particularly advantageous that the proposed method can be applied, preferably at least in principle, independently of the number of sensors in the arrangement and also independently of the type of sensors in the arrangement. This makes the proposed method particularly flexible for use with different sensor arrangements. Furthermore, the proposed method is also particularly easy to use with existing sensor arrangements, since the physical sensor arrangement does not need to be adapted, or only needs to be adapted to a manageable extent. This means that the method can also be used economically in conjunction with existing sensor arrangements. In this way, the advantages of improved operation can be exploited in a variety of ways for both existing and newly installed sensor arrangements.

[0018] The proposed method can be used to advantageously counteract technical sensor failures in devices, particularly machines. This is because processes (or functionalities or other entities in general) that depend on the sensor data can continue to run as usual. This reliably avoids or at least reduces the risk of operational interruptions. For example, if one of the sensors (load cells) fails in a device such as a platform scale with a large number of sensors, for example four load cells, the device (platform scale) can continue to be used when operated according to the proposed method, since sensor data for the failed sensor continues to be available as a substitute due to the recourse to the virtual sensor.In the case of the platform scale, the total weight therefore advantageously does not deviate from the true total weight by around 25%, as is the case in conventional situations.

[0019] It should be noted that during error-free operation of the sensor arrangement, the received sensor data of each sensor is preferably identical to the provided original sensor data of the respective sensor. However, if, as a result of a detected error condition, the replacement sensor data for the specific sensor is determined and provided as the original sensor data of the specific sensor, the received sensor data of the specific sensor (if any data is received at all) and the provided original sensor data of the specific sensor may differ, and advantageously will usually do so.

[0020] The replacement sensor data can advantageously be provided as original sensor data for the specific sensor instead of the received sensor data of the specific sensor until proper sensor data is received again from the specific sensor, for example after a replacement or repair of the specific sensor.

[0021] Invalid sensor data can be received from a sensor, for example, if the sensor has become detached from its measuring location, for example if it has become completely or partially detached from a device and / or has fallen off.

[0022] It is particularly advantageous if several or all of the sensors from the multitude of sensors that are not the specific sensor are selected as auxiliary sensors. This can make the database for determining the substitute sensor data more meaningful, as multiple sensors contribute to it.

[0023] Advantageously, the source sensor data of the auxiliary sensors, which are used here as input data for the first trained data model of the machine learning, comprise or represent the received sensor data currently received by the auxiliary sensors. This allows the current circumstances of the sensor arrangement to be taken into account particularly reliably.

[0024] In one embodiment, the first trained machine learning data model is provided after the fault condition is detected. It can then be used to determine the replacement sensor data.

[0025] In one embodiment, the replacement sensor data, in particular at least for the specific sensor, is continuously determined using the first trained machine learning data model. In this case, when the error condition is detected, this data can then advantageously be provided as the original sensor data instead of the sensor data of the specific sensor. In other words, the system switches to the replacement sensor data (only) when the error condition is detected. This allows a sequence of original sensor data for the specific sensor to be provided uninterruptedly or almost uninterruptedly.

[0026] The error condition associated with the specific sensor is preferably detected at a first point in time. The proposed method is particularly advantageous for operating a sensor arrangement of a conveying, measuring, weighing, grinding, mixing, filtering, screening, drying, and / or dosing device.

[0027] Preferably, the method is computer-implemented and / or executed using a data processing device configured to execute the method. This data processing device will be discussed in more detail below, with the statements there applying here accordingly, unless the context indicates otherwise.

[0028] In the context of the present application, a sensor arrangement is preferably understood to mean a collection of at least two sensors, advantageously of the same type, wherein preferably their sensor data are evaluated and / or processed within the framework of a joint data processing.

[0029] Advantageously, all sensors of the arrangement are mounted on a single device, for example to record operating parameters and / or physical quantities of the device and / or its parts.

[0030] In the present application, the terms "x-th trained machine learning data model", "trained x-th machine learning data model" and "x-th machine learning data model in trained form" (with "x-th" depending on the situation, for example, "first", "second" or "third") are preferably used synonymously, unless the respective context indicates otherwise.

[0031] Alternatively or additionally, it may also be provided that the training of the first data model of the machine learning is carried out after the error condition has been detected.

[0032] The use of such a freshly trained data model can lead to particularly meaningful replacement sensor data, so that a particularly reliable operation of the sensor arrangement is possible despite a failed or defective sensor.

[0033] In this case, training of the first machine learning data model is advantageously started immediately or almost immediately (i.e., less than 1 minute, less than 30 seconds, less than 10 seconds, less than 5 seconds, or less than 1 second) after the error condition is detected. This allows the data model to be made available in trained form particularly quickly and used to determine the replacement sensor data. At the same time, the most recent sensor data possible can be used for training, which can advantageously contribute to a reliable data model.

[0034] However, in embodiments, it may also be preferable to start training with a delay after the error condition has been detected and / or to add it to a task list with tasks to be processed. This allows for a particularly advantageous, variable consideration of the current workload of the computer system responsible for training. For example, depending on the system workload and / or the availability of resources, training can be started, in particular within a defined or definable time window.

[0035] In one embodiment, the training is performed using the data processing device. A storage device may also be provided there, on which the trained data models are stored and / or from which they can be retrieved.

[0036] Alternatively or additionally, it can also be provided that the training of the first data model of the machine learning is or was carried out at least by means of historical sensor data, in particular historical received sensor data, of the specific sensor and / or historical sensor data, in particular historical received sensor data, of at least one sensor, preferably all sensors, of the at least one auxiliary sensor.

[0037] It has been shown that, based on corresponding historical sensor data, reliable and particularly advantageous replacement sensor data can be determined, especially temporarily. This works particularly well if the environmental conditions existing at the time the historical sensor data was recorded have not changed or have changed only to a limited extent.

[0038] In advantageous embodiments, for example, the received sensor data of the specific sensor and the auxiliary sensors of the array before the failure are used to train the model. This advantageously allows the virtual sensor to estimate its raw data (and thus the raw data of the defective sensor) based on the raw data of the functioning sensors, in particular the selected auxiliary sensors.

[0039] Training the first machine learning data model can, for example, involve using at least historical sensor data, in particular historical received sensor data, of the at least one auxiliary sensor (or the plurality of auxiliary sensors) as input data for the data model and / or using historical sensor data, in particular historical received sensor data, of the specific sensor as truth data. The input data and the truth data are advantageously associated with one another.

[0040] Alternatively or additionally, it can also be provided that the training of the first data model of the machine learning involves assuming a linear relationship between the historical sensor data of the specific sensor on the one hand and the historical sensor data of the auxiliary sensors on the other hand.

[0041] A linear relationship enables the use of a particularly well and efficiently implementable data model.

[0042] For example, in a sensor arrangement with four sensors, a linear relationship between the sensor data of a specific sensor Si and the sensor data of three auxiliary sensors S2 to S4 at a specific time t is advantageously described by the relationship

[0043] S (t) = a ■ S2(t) + b ■ S3(t) + c ■ S4(t) + d.

[0044] In the context of training a machine learning data model implementing this relationship, the coefficients a, b, c, and d could then be determined in the present example. After training, these coefficients could be used to provide the trained machine learning data model. Using the trained data model, the current received sensor data (S2 to S4) of the three auxiliary sensors could then be inserted into the equation with the coefficients determined during training, and the current substitute sensor data could be determined as the result (Si).

[0045] In one embodiment, the data model is of the linear regression type, in particular of the stepwise linear regression type. Advantageously, or alternatively, those sensors of the plurality of sensors (and which are not the specific sensor) that have the greatest predictive power are selected as auxiliary sensors.

[0046] Alternatively or additionally, it can also be provided that the historical sensor data of the specific sensor and the historical sensor data of the auxiliary sensors were recorded within the same time window. Preferably, the individual historical sensor data of the specific sensor and all auxiliary sensors are recorded at the same times, and preferably all times lie within the time window. For example, each sensor records one measured value per unit of time (e.g., per second). This means that one measured value is available from each sensor per unit of time. If the time window is 100 time units long, the sensor data of each sensor in this case comprises 100 measured values.

[0047] Alternatively or additionally, it can also be provided that the historical sensor data of the specific sensor and / or the auxiliary sensors are the received sensor data received during a defined or definable period before the detection of the error state or data of the respective sensors provided as original sensor data, wherein preferably (i) within the period no error state was detected for either the specific sensor or for one of the auxiliary sensors and / or (ii) sensors of the plurality of sensors for which an error state was detected within the period are not selected as auxiliary sensors.

[0048] For example, the period may end immediately upon detection or earlier, in particular with a defined or definable time offset.

[0049] The longer the period selected, the more information about the individual sensors and their behavior and properties is available as a database, on which the replacement sensor data can be determined and / or on which the data model can be trained. With a shorter period, the replacement sensor data can be advantageously determined based on more recent sensor behavior. This way, longer-term changes in the sensor properties, for example, due to changing environmental conditions or progressive sensor aging, can be ignored.

[0050] An advantageous period is 30 days or shorter, preferably 14 days or shorter, preferably 7 days or shorter, preferably 3 days or shorter, preferably 1 day or shorter, preferably 12 hours or shorter, preferably 6 hours or shorter, preferably 3 hours or shorter, preferably 1 hour or shorter, and / or 1 minute or longer, preferably 1 hour or longer, preferably 3 hours or longer, preferably 6 hours or longer, preferably 9 hours or longer, preferably 12 hours or longer, preferably 1 day or longer, preferably 3 days or longer, preferably 7 days or longer, preferably 14 days or longer, preferably 30 days or longer.

[0051] By selecting only those sensors for which no error condition has been detected as auxiliary sensors, a particularly solid database can be provided for determining the replacement sensor data, especially for training the data model.

[0052] Alternatively or additionally, it can also be provided that the sensor data, in particular the historical sensor data, of the specific sensor correlate at least slightly with the sensor data, in particular the historical sensor data, of each auxiliary sensor.

[0053] This makes it particularly advantageous to determine meaningful replacement sensor data based on the sensor data from the auxiliary sensors.

[0054] For the purposes of the present application, the sensor data are preferably at least slightly correlated if there is at least a slight relationship between the data. The existing relationship can be determined, for example, using a correlation coefficient of the sensor data.

[0055] For example, a correlation coefficient can take values ​​from -1 to 1. At -1, there is advantageously a perfectly negative correlation between the sensor data, at 0 there is no correlation (at least not a linear one and thus preferably not a correlation in the sense of the present definition) between the sensor data, and at 1 there is a perfectly positive correlation between the sensor data.

[0056] In one embodiment, there is at least a weak correlation between the sensor data if there is a correlation coefficient of, in particular in terms of amount, at least 0.1, preferably at least 0.2, preferably at least 0.3, preferably at least 0.4, preferably at least 0.5, preferably at least 0.6, preferably at least 0.7, preferably at least 0.8, preferably at least 0.9, preferably at least 0.95, between the respective sensor data.

[0057] Alternatively or additionally, it may also be provided that an error condition is detected in the specific sensor,

[0058] (i) if no sensor data is received from the specific sensor, at least temporarily,

[0059] (ii) if the received sensor data or a statistical value thereof received by the specific sensor is above or below a defined or definable threshold,

[0060] (iii) if the received sensor data received by the specific sensor does not meet a defined or definable quality measure,

[0061] (iv) if a result of a test of an electrical resistance of the specific sensor, in particular in the form of a load cell, indicates a defect in the sensor, and / or

[0062] (v) if a value, in particular a maximum value, of a correlation between the received received sensor data of the specific sensor and the received received sensor data of at least one other sensor of the plurality of sensors, in particular the auxiliary sensors, is above or below a defined or definable threshold value.

[0063] For example, receiving the received sensor data from the specific sensor may be prevented by a malfunction or failure of the transmission channel (e.g., a cable or radio channel). This is particularly possible in the case of a sensor from which the sensor data is received via a data cable or wirelessly via the air interface.

[0064] The statistical value can, for example, be an average of the sensor data, in particular over a defined or definable period of time.

[0065] A corresponding quality measure can, for example, be the noise behavior of the sensor and / or the sensor data recorded with it. Therefore, it may be advantageous to check the sensor and / or the sensor data for noise behavior and, preferably, to use the result of the check as a quality measure.

[0066] The test of the electrical resistance can advantageously be carried out by means of a test device which is in operative connection with the sensor or is brought into connection with it and / or is provided by it, in particular continuously or intermittently.

[0067] For example, a correlation can also be carried out between the received sensor data of the specific sensor and the received sensor data of each of two or more than two, preferably all, of the other sensors of the plurality of sensors, in particular the auxiliary sensors, and an error condition can be determined if a defined or definable number of correlations (in each case between the sensor data of the specific sensor and the sensor data of another sensor) result in a value, in particular a maximum value, which is above or below a defined or definable threshold value.

[0068] Alternatively or additionally, it can also be provided that after the error state in connection with the specific sensor has been detected, a cessation of the error state in connection with the specific sensor is detected and then the received sensor data of the specific sensor are received again and / or provided as original sensor data of the specific sensor and in particular the replacement sensor data are no longer provided as original sensor data of the specific sensor.

[0069] This allows for a particularly reliable and automatic return to normal operating mode when the sensors, especially the specific sensor, in the sensor array are once again providing correct sensor data. For example, after a replacement or repair of the specific sensor.

[0070] The elimination of the fault condition is preferably determined after the fault condition has been detected, in particular at a second point in time which is after the first point in time.

[0071] In one embodiment, the determined substitute sensor data are compared continuously and / or repeatedly at intervals with the received sensor data received by the specific sensor, and based on the result of the comparison, the elimination of the error condition associated with the specific sensor is determined. This allows the substitute sensor data provided by the respective data model to also be used for the further purpose of detecting the renewed proper operation of the specific sensor. Normal operation can then be returned to, and the received sensor data received (by the specific sensor) can be provided again as the original sensor data (of the specific sensor).

[0072] Alternatively or additionally, it can also be provided that the received sensor data is received continuously by each sensor of the plurality of sensors, that the received sensor data is received in parallel by all sensors of the plurality of sensors and / or that the result data of the first trained data model of the machine learning is used as replacement sensor data.

[0073] Advantageously, when a sensor receives sensor data continuously, sensor data is constantly received from the sensor. For example, the sensor data can be digital values ​​received at a specific clock frequency.

[0074] Advantageously, when sensor data from several sensors is received in parallel, the sensor data from the several sensors are received in parallel via one or more transmission channels.

[0075] Alternatively or additionally, it can also be provided that a second machine learning data model is kept ready in trained form for at least the specific sensor, and wherein the determination of the replacement sensor data comprises determining the replacement sensor data, in particular at least temporarily, preferably at least until the training of the first machine learning data model is completed, using the second trained machine learning data model kept ready for the specific sensor, wherein the second trained machine learning data model is preferably identical to the first trained machine learning data model. By using the second trained data model, after the error state has been detected, without a large time delay (especially without having to wait in preferred embodiments until, if necessary,the training of the first data model is completed) the replacement sensor data is determined using the second trained data model.

[0076] For example, the second trained data model can be used until the training of the first data model is complete. After the training of the first data model is complete, the replacement sensor data is determined using the first trained data model (and, in particular, is no longer determined using the second trained machine learning data model). Alternatively, the second machine learning data model can be identical to the first machine learning data model. The (first / second) data model used to determine the replacement sensor data is therefore advantageously already kept in trained form and ready for use at the time the error condition is detected. The replacement sensor data is then determined using this data model. It is then no longer necessary to change the trained data model to determine the replacement sensor data.

[0077] Preferably, the received sensor data and / or the original sensor data of at least one of the auxiliary sensors, preferably all auxiliary sensors, and / or data based thereon are used at least partially as input data for the second trained machine learning data model and / or the result data of the second trained machine learning data model are used as replacement sensor data.

[0078] Preferably, the statements made regarding the training of the first data model apply accordingly with regard to the training of the second data model. The corresponding features can therefore also be provided individually and in any combination during the training of the second data model. It goes without saying, however, that the reference point of the historical data in this case is no longer the detection of the error condition, but is preferably determined by the start of the training of the second data model. This means that received sensor data that was received before the said training of the second data model or that was provided as original sensor data are advantageously to be understood in this context as historical sensor data of the respective sensors.

[0079] Preferably, the statements made regarding the determination of the substitute sensor data using the first data model apply accordingly with regard to the determination of the substitute sensor data using the second data model. The corresponding features can therefore also be provided individually and in any combination when determining the substitute sensor data using the second data model.

[0080] The second data model of machine learning in trained form is advantageously already available when the error condition is detected.

[0081] Preferably, such a second trained machine learning data model is kept available for each sensor of the arrangement.

[0082] Alternatively or additionally, it can also be provided that a third machine learning data model is kept ready in trained form for at least the specific sensor, preferably the third trained machine learning data model is identical to the second trained machine learning data model, and wherein, with the third trained machine learning data model, test sensor data for the specific sensor is determined at least temporarily, preferably continuously, and compared with the sensor data received from the specific sensor, and wherein, based on a result of the comparison, a fault condition in the specific sensor is determined. By using the third trained data model, a virtual sensor, so to speak, continuously supplies comparison sensor data that can be checked against the sensor data of the specific sensor. Thus, fault conditions can be detected particularly reliably.

[0083] For example, the third machine learning data model can be identical to the first and / or second machine learning data model. Using the (first / second / third) data model used to provide the test sensor data, the replacement sensor data can then be provided seamlessly from the moment the fault condition is detected.

[0084] Preferably, the received sensor data and / or the original sensor data of at least one of the auxiliary sensors, preferably all auxiliary sensors, and / or data based thereon are used at least partially as input data for the third trained machine learning data model and / or the result data of the third trained machine learning data model are used as test sensor data and / or replacement sensor data.

[0085] Preferably, the statements made regarding the training of the first data model apply accordingly with regard to the training of the third data model. The corresponding features can therefore also be provided individually and in any combination during the training of the third data model. It goes without saying, however, that the reference point of the historical data in this case is no longer the detection of the error condition, but is preferably determined by the start of the training of the third data model. This means that received sensor data that was received before the aforementioned training of the third data model or that was provided as original sensor data are advantageously to be understood in this context as historical sensor data of the respective sensors.

[0086] Preferably, the statements made regarding the determination of the replacement sensor data with the first data model apply accordingly with regard to the determination of the test sensor data and / or the replacement sensor data with the third data model. The corresponding features can therefore also be provided individually and in any combination when determining the test sensor data and / or replacement sensor data with the third data model.

[0087] Preferably, such a third trained machine learning data model is kept available for each sensor of the arrangement.

[0088] Alternatively or additionally, it can also be provided that the received sensor data and / or the source sensor data are assigned or assignable to the individual sensors and / or the source sensor data are provided to a process, a module, a device and / or in the form of a control signal.

[0089] For example, the received sensor data and / or source sensor data can be sorted by sensor and / or the origin of the source sensor data can be known for the individual sensors in another way.

[0090] Generally speaking, providing the source sensor data can involve providing the data to an entity (which can preferably be implemented in software, hardware, or a combination of both). This allows other entities, such as software and / or hardware modules, processes, functionalities, software functions, and / or devices such as systems and machines, to access and / or receive this sensor data. It is particularly flexible if the source sensor data is provided as a control signal. Such a control signal can, for example, be digital and / or analog in nature. The control signal can optionally comprise multiple sub-control signals, in particular as many as there are sensors in the arrangement. For example, each sub-control signal can then represent the source sensor data of a single sensor in the arrangement.

[0091] Alternatively or additionally, it can also be provided that the sensor arrangement has two or more than two, in particular three or more than three, in particular four or more than four, in particular five or more than five, in particular six or more than six, in particular seven or more than seven, in particular eight or more than eight, in particular nine or more than nine, in particular ten or more than ten, sensors, and / or wherein all sensors of the arrangement are of the same type.

[0092] For example, the sensors of the arrangement, in particular the specific sensor and the auxiliary sensors, are all or at least partially of the type current measuring sensor, voltage measuring sensor, force sensor, load cell, acceleration sensor, motion sensor, speed sensor, rotational speed sensor, temperature sensor, ultrasonic sensor and / or eddy current sensor.

[0093] The object is achieved by the invention according to a second aspect in that a device for data processing, in particular having one or more interfaces for receiving sensor data from a plurality of sensors, wherein the device is adapted to carry out a method according to the first aspect of the invention, is proposed.

[0094] The data processing device can be implemented, for example, in software, hardware, or a combination of both. Alternatively or additionally, the data processing device can comprise a memory (in particular for storing the first, second, and / or third data model of the machine learning), a processor, a receiving device, a transmitting device (for example, for transmitting the original sensor data, in particular the control signal, to an internal or external entity), or any combination thereof. The data processing device preferably comprises one or more interfaces for receiving sensor data from a plurality of sensors (in particular the plurality of sensors of the sensor arrangement used in the method according to the first aspect of the invention).

[0095] The object is achieved by the invention according to a third aspect in that a sensor arrangement, in particular in the form of a plurality of sensors, which is adapted to be used in a method according to the first aspect of the invention and / or to cooperate with a device for data processing according to the second aspect of the invention is proposed.

[0096] The statements made with regard to the first aspect of the invention also apply to the third aspect of the invention, unless the context indicates otherwise. In particular, all advantages and features explained with regard to the sensor arrangement used in a method according to the first aspect of the invention also apply here accordingly. Reference can therefore be made to the previous statements in this regard.

[0097] Therefore, in preferred embodiments, all features explained with regard to the sensor arrangement used in a method according to the first aspect of the invention can also be provided, individually and in any combination, in the sensor arrangement according to the third aspect of the invention.

[0098] The object is achieved by the invention according to a fourth aspect in that a device, in particular a machine, with a sensor arrangement arranged thereon according to the third aspect of the invention and / or comprising such a sensor arrangement and / or comprising a device for data processing according to the second aspect of the invention is proposed.

[0099] The statements made with regard to the first and second aspects of the invention also apply with regard to the fourth aspect of the invention, unless the context indicates otherwise. In particular, all advantages and features explained with regard to the sensor arrangement used in a method according to the first aspect of the invention and with regard to a sensor arrangement according to the third aspect of the invention, as well as alternatively or additionally those explained with regard to a data processing device according to the second aspect of the invention, also apply here accordingly. Reference can therefore be made to the previous statements in this regard.

[0100] Therefore, in preferred embodiments, all features which have been explained with regard to the sensor arrangement used in a method according to the first aspect of the invention, which have been explained with regard to the sensor arrangement according to the third aspect of the invention and / or which have been explained with regard to the data processing device according to the second aspect of the invention, can also be provided individually and in any combination in the sensor arrangement and / or the data processing device of the device according to the fourth aspect of the invention.

[0101] Preferably, the device is or comprises a conveying, measuring, weighing, grinding, mixing, filtering, screening, drying and / or dosing device.

[0102] Alternatively or additionally, it can also be provided that (i) the device is or has a platform scale and / or the sensors of the sensor arrangement are load cells, (ii) the device is or has a sieve and / or the sensors of the sensor arrangement are motion and / or acceleration sensors, (iii) the device is or has a dosing device and / or the sensors of the sensor arrangement are speed sensors and / or load cells, (iv) the device is or has a belt scale and / or the sensors of the sensor arrangement are load cells and / or (v) the device is or has a crane scale and / or the sensors of the sensor arrangement are force sensors, in particular load cells.

[0103] Short description of the drawings

[0104] Further features and advantages of the invention will become apparent from the following description, in which preferred embodiments of the invention are explained with reference to schematic drawings.

[0105] Showing:

[0106] Fig. 1 is a schematic view of a sensor arrangement according to the third aspect of

[0107] Invention together with a data processing device according to the second aspect of the invention;

[0108] Fig. 2 is a schematic view of an apparatus according to the fourth aspect of the invention;

[0109] Fig. 3 is a flowchart of a method according to the first aspect of the invention; and

[0110] Fig. 4 Comparison of real sensor data and calculated substitute sensor data. Description of the embodiments

[0111] Fig. 1 shows a schematic view of a sensor arrangement 1 according to the third aspect of the invention, which is operatively connected to a data processing device 3 according to the second aspect of the invention.

[0112] The sensor arrangement 1 comprises four identical sensors 5a..d, each in the form of a load cell. Each sensor 5a..d is connected to an interface 9a..d of the device 3 via a transmission channel 7a..d, each in the form of a cable. The device 3 is configured to carry out a method according to the first aspect of the invention.

[0113] The sensor arrangement 1 can advantageously be used in a platform scale, as realized by the device 11 according to the fourth aspect of the invention, schematically illustrated in Fig. 2 and shown in Fig. 2 together with the data processing device 3. The platform scale has a support surface 13, on which an object to be weighed can be placed and which is mounted on the sensors 5a..d (which are hidden by the support surface 13 in Fig. 2 and are therefore only shown in dashed lines) of the sensor arrangement 1. When an object is placed on the support surface 13, a force acts on the sensors 5a..d due to its weight. Each sensor 5a..d generates sensor data in the form of digital measured values ​​corresponding to the respective force applied.

[0114] Fig. 3 shows a flowchart 100 of a method according to the first aspect of the invention.

[0115] In 101, during operation of the sensor arrangement 1, the device 3 receives sensor data from the individual sensors 5a..d as received sensor data via the transmission channels 7a..d and provides it as source sensor data (for example, to a process within a software). In the case of load cells, the received sensor data also varies accordingly over time depending on the time-dependent force acting on the individual sensors. The current sensor data is continuously received in parallel from all sensors. At least during error-free operation of the sensor arrangement 1, the received received sensor data of each sensor is advantageously identical to the provided source sensor data of the respective sensor.

[0116] Due to a failure of one of the sensors, such as sensor 5a, which will be referred to as the specific sensor for ease of reference, device 3 will no longer receive any sensor data from the specific sensor 5a from a certain point in time. The sensor failure may be caused, for example, by a defect within the specific sensor 5a or a (perhaps physical) interruption of the transmission channel 7a (e.g., due to a severed cable).

[0117] Due to the fact that no more sensor data is received from the specific sensor 5a, an error condition associated with the specific sensor 5a is detected in 103.

[0118] In 105, a training of a first machine learning data model is then carried out.

[0119] For this purpose, the remaining three sensors 5b..d of arrangement 1 are selected as auxiliary sensors, and their received sensor data from the last two days before the detection of the error state are used as input data for training. At the same time, the received sensor data of the specific sensor 5a from the last two days before the detection of the error state are used as truth data associated with the input data for training. The received sensor data from the last two days are therefore historical sensor data. A linear relationship is assumed between the historical sensor data of the specific sensor 5a on the one hand and the historical sensor data of the auxiliary sensors 5b..d on the other.

[0120] Care is taken to ensure that no error condition was detected for any of the sensors 5a..d during the period from which the historical sensor data from sensors 5a..d originate, i.e., the last two days. If this had been the case, a shorter and / or later period could have been selected, for example, in which no error condition would have been detected, and the historical sensor data from sensors 5a..d from this period could have been used.

[0121] Once the training of the first machine learning data model is complete, the received sensor data from the auxiliary sensors 5b..d are used as input data for the first data model in 107. The result data obtained by calculating the trained data model as substitute sensor data are provided as source sensor data of the specific sensor 5a. This means that while the provided source sensor data from the auxiliary sensors 5b..d continue to be the received sensor data received by these sensors 5b..d, the provided source sensor data from the specific sensor 5a are the determined substitute sensor data.

[0122] After a certain period of time, the sensor 5a functions properly again, for example because the defective sensor 5a has been repaired or replaced, so that the device 3 again receives sensor data from the specific sensor 5a.

[0123] At 109, therefore, the elimination of the error state associated with the specific sensor 5a is determined. Subsequently, the received sensor data of the specific sensor 5a is received again and provided as the original sensor data of the specific sensor 5a. In particular, the substitute sensor data is no longer used. Therefore, for example, the use of the first trained data model of the machine learning can then also be terminated. The received received sensor data of each sensor is then again identical to the original sensor data provided for the respective sensor.

[0124] Fig. 4 shows the course of the sensor data received from a real sensor of a sensor arrangement according to the invention (curve A) during a specific period of time, together with the course of substitute sensor data calculated for this same real sensor for this same real sensor (curve B) using a trained machine learning data model. The sensor arrangement had four sensors. The machine learning data model used was trained using historical sensor data from the four sensors in a manner comparable to the first machine learning data model used in the method described with reference to the flowchart in Fig. 3. During the period of time shown in the diagram in Fig. 4 (time axis T), the trained machine learning data model was calculated using the current sensor data from the remaining three sensors to determine the shown substitute sensor data (curve B) for the sensor.

[0125] The almost identical course of the two curves A and B confirms the particularly advantageous and reliable operation of the proposed method according to the first aspect of the invention for the operation of a sensor arrangement.

[0126] The features disclosed in the foregoing description, in the drawings, and in the claims may be essential to the invention in its various embodiments, both individually and in any combination. List of reference symbols

[0127] 1 sensor arrangement

[0128] 3 Data processing facility

[0129] 5a Specific sensor

[0130] 5b, 5c, 5d Auxiliary sensor

[0131] 7a, 7b, 7c, 7d transmission channel

[0132] 9a, 9b, 9c, 9d interface

[0133] 11 Device

[0134] 13 Support surface

[0135] 100 Flowchart

[0136] 101 Receiving sensor data from a plurality of sensors and providing this as source sensor data of the individual sensors

[0137] 103 Determining an error condition associated with a specific sensor of the plurality of sensors

[0138] 105 Training a first machine learning data model using historical sensor data from a multitude of sensors

[0139] 107 Determining replacement sensor data using the first trained machine learning data model and providing the replacement sensor data instead of the sensor data of the specific sensor as the original sensor data of the specific sensor

[0140] 109 Determining the elimination of the error condition in connection with the specific sensor and receiving and providing sensor data of the specific sensor again as original sensor data of the specific sensor

[0141] A, B History of sensor data

[0142] X, T diagram axis

Claims

Patent claims Method for operating a sensor arrangement with a plurality of sensors, comprising that from each sensor of the plurality of sensors, sensor data is received as received sensor data of the respective sensor and is provided as original sensor data of the respective sensor for further data processing, wherein an error condition is detected in connection with at least one specific sensor of the plurality of sensors and then replacement sensor data is determined for the specific sensor and is provided as original sensor data of the specific sensor instead of received sensor data of the specific sensor,wherein the substitute sensor data is determined by means of at least one first trained machine learning data model, and at least partially the received sensor data and / or original sensor data of at least one sensor selected as an auxiliary sensor from the plurality of sensors, which is not the specific sensor, and / or data based thereon are used as input data for the first trained machine learning data model. The method according to claim 1, wherein the training of the first machine learning data model is carried out after the detection of the error state. The method according to one of the preceding claims, wherein the training of the first machine learning data model is carried out at least by means of historical sensor data, in particular historical received sensor data, of the specific sensor and / or historical sensor data, in particular historical received sensor data, of at least one sensor.preferably all sensors of the at least one auxiliary sensor is or has been carried out. Method according to one of the preceding claims, wherein the training of the first data model of the machine learning comprises assuming a linear relationship between the historical sensor data of the specific sensor, on the one hand, and the historical sensor data of the auxiliary sensors, on the other hand. Method according to one of claims 3 to 4, wherein the historical sensor data of the specific sensor and the historical sensor data of the auxiliary sensors were acquired within the same time window. Method according to one of claims 3 to 5, wherein the historical sensor data of the specific sensor and / or the auxiliary sensors are the received sensor data received during a defined or definable period before the detection of the error state or the data of the respective sensors provided as original sensor data.Preferably, (i) within the period, no fault condition was detected for either the specific sensor or any of the auxiliary sensors, and / or (ii) sensors of the plurality of sensors for which a fault condition was detected within the period are not selected as auxiliary sensors. Method according to one of the preceding claims, wherein the sensor data, in particular the historical sensor data, of the specific sensor correlate at least slightly with the sensor data, in particular the historical sensor data, of each auxiliary sensor. Method according to one of the preceding claims, wherein a fault condition is detected in the specific sensor, (i) if no sensor data is received from the specific sensor, at least temporarily, (ii) if the received sensor data or a statistical value thereof received by the specific sensor is above or below a defined or definable threshold, (iii) if the received sensor data received by the specific sensor does not meet a defined or definable quality measure, (iv) if a result of a test of an electrical resistance of the specific sensor, in particular in the form of a load cell, indicates a defect in the sensor, and / or (v) if a value, in particular a maximum value, of a correlation between the received received sensor data of the specific sensor and the received received sensor data of at least one other sensor of the plurality of sensors, in particular the auxiliary sensors, is above or below a defined or definable threshold value. Method according to one of the preceding claims, wherein, after the error state associated with the specific sensor has been detected, a cessation of the error state associated with the specific sensor is detected, and then the received sensor data of the specific sensor are received again and / or provided as original sensor data of the specific sensor, and in particular, the substitute sensor data are no longer provided as original sensor data of the specific sensor.Method according to one of the preceding claims, wherein the received sensor data is received continuously by each sensor of the plurality of sensors, wherein the received sensor data is received in parallel by all sensors of the plurality of sensors and / or wherein the result data of the first trained data model of the machine learning is used as replacement sensor data.Method according to one of the preceding claims, wherein a second machine learning data model is kept ready in trained form for at least the specific sensor, and wherein the determination of the replacement sensor data comprises determining the replacement sensor data, in particular at least temporarily, preferably at least until the training of the first machine learning data model is completed, using the second trained machine learning data model kept ready for the specific sensor, wherein the second trained machine learning data model is preferably identical to the first trained machine learning data model. Method according to one of the preceding claims, wherein a third machine learning data model is kept ready in trained form for at least the specific sensor. Preferably, the third trained machine learning data model is identical to the second trained machine learning data model, and wherein, with the third trained machine learning data model, test sensor data for the specific sensor is determined at least temporarily, preferably continuously, and compared with the sensor data received from the specific sensor, and wherein, based on a result of the comparison, a fault condition in the specific sensor is determined. Method according to one of the preceding claims, wherein the received sensor data and / or the original sensor data are respectively assigned or assignable to the individual sensors and / or the original sensor data are provided to a process, a module, a device and / or in the form of a control signal.Method according to one of the preceding claims, wherein the sensor arrangement has two or more than two, in particular three or more than three, in particular four or more than four, in particular five or more than five, in particular six or more than six, in particular seven or more than seven, in particular eight or more than eight, in particular nine or more than nine, in particular ten or more than ten, sensors, and / or wherein all sensors of the arrangement are of the same type. Device for data processing, in particular comprising one or more interfaces for receiving sensor data from a plurality of sensors, wherein the device is adapted to carry out a method according to one of the preceding claims 1 to 14.Sensor arrangement, in particular in the form of a plurality of sensors, which is adapted to be used in a method according to one of the preceding claims 1 to 14 and / or to cooperate with a data processing device according to claim 15. Device, in particular a machine, with a sensor arrangement arranged thereon according to claim 16 and / or comprising such a sensor arrangement and / or comprising a data processing device according to claim 15.Device according to claim 17, wherein (i) the device is or comprises a platform scale and / or the sensors of the sensor arrangement are load cells, (ii) the device is or comprises a sieve and / or the sensors of the sensor arrangement are motion and / or acceleration sensors, (iii) the device is or comprises a dosing device and / or the sensors of the sensor arrangement are speed sensors and / or load cells, (iv) the device is or comprises a belt scale and / or the sensors of the sensor arrangement are load cells and / or (v) the device is or comprises a crane scale and / or the sensors of the sensor arrangement are force sensors, in particular load cells.

Images