Patent Application
20220413477
The present invention relates to the field of industrial automation technology and, in particular, methods and techniques for data-reduced transmission between edge devices and cloud systems based on prediction models.
During the course of the Fourth Industrial Revolution and the all-pervasive digital transformation as well as, and in particular, in industrial automation technology, cloud solutions are increasingly being used to analyze, manage and possibly even control the automation environment and its components. Within a cloud system, the received data is usually stored, evaluated and made available to the users in prepared form for any type of terminal. It is also possible to send data (e.g., commands, firmware updates) from the cloud system to the components or devices of an automation environment. The connection between the devices and the cloud system(s) may be designed to be wirebound or wireless (e.g., via mobile radio technologies).
In this context, the devices within the automation environment are often also referred to as edge devices. An edge device is generally to be understood to be a device, which may be connected to a remote computer system, in particular a cloud system, but nevertheless carries out local and therefore decentralized data processing, for example, to prepare data prior to being transmitted to the cloud or to the remote computer system.
However, sending data from devices to cloud systems and the data link needed for this purpose are afflicted with negative characteristics. For example, transmitting data in the industrial context is often associated with use-dependent costs, i.e., the more data that are transmitted, the more expensive this may become. The costs result, for example, from the arising traffic and/or the number of incoming messages to the cloud system.
In addition, the data transmission rate is often physically limited, which means that an arbitrary volume of data may not be transmitted. The more devices that share a connection, the lower is the possible data transmission rate.
In the context of the Internet, in particular, but also in an intranet, a data link may also be impaired by outages.
One way to overcome these disadvantages is disclosed, for example, by the document, “Deep Learning: Edge Cloud Data Analytics for IoT” by A. Ghosh et al. (Department of Electrical and Computer Engineering, Western University, London, ON, Canada, May 2019). The document discloses a method, in which sensor data are received by an edge device, reduced within the edge device, and subsequently transmitted to the cloud. Here, the reduced data are converted back into the original data.
It is therefore an object of the present invention to provide techniques for transmitting data from an industrial automation environment to a computer system, which at least partially overcome the negative effects of the characteristics of the prior art described above.
In an exemplary embodiment, a method is provided for providing process data of a device in an industrial automation environment to a computer system. A process data model is executed on the device for the purpose of generating particular estimated process data. The process data model is preferably also executed on the computer system, so that estimated process data are also generated there. It is also determined (for example, on the device) whether the estimated process data deviate from the real process data by more than a threshold value. Only if the estimated process data deviate from the real process data by more than the threshold value is information representing the real process data transmitted from the device to the computer system.
As described at the outset in reference to the prior art, a computer system would usually use real process data, which it has received from the device in the automation environment. The invention deviates therefrom in that the computer system now initially uses estimated process data, which were generated by a process data model, instead of real process data. In other words, the computer system works to a certain extent on a simulation of the real process data.
To prevent the process data model from supplying inaccurate estimated process data, the invention further provides for also executing the process data model on the device, so that the estimated process data may be compared with the real process data. In the case of too great a deviation (it also being possible for the aforementioned threshold value to be zero, i.e., in the case of some other small deviation), information representing the real process data is transmitted from the device to the computer system. This information is preferably transmitted only if the deviation occurs, i.e., during normal operation (while the computer system is working on the basis of the process data model), no information is transmitted from the device to the computer system.
This is technically advantageous in several respects: the better the process data model, the less information needs to be transmitted from the device to the computer system or the less often does it have to be transmitted. This results in significant savings when it comes to exchanging data. Fewer data transmissions result in lower costs for traffic and incoming messages to the computer system; the transmission time for the data from the device to the computer system is also largely eliminated, which may result in shorter response times; due to the fact that the data no longer has to be sent continuously, the consequences of a data theft may possibly be less severe, since the person intercepting the data is able to seize only a portion of the data; and if the data transmission is interrupted, this is only relevant if the real process data deviate from the estimated process data. The comparison must then take place at a later point in time when the data transmission is functioning again.
In an aspect of the invention, the information representing the real process data may comprise the real process data. In other words, in the case of a deviation between the process data model and the real process data, the real process data themselves are sent to the computer system.
Alternatively, the information representing the real process data may also include reconstruction data, the reconstruction data allowing the computer system to reconstruct the real process data from the estimated process data. For example, the reconstruction data may comprise difference values.
A further aspect of the present invention, which is advantageous in combination with the above aspects as well as on its own, relates to the generation of the process data model by a machine learning system based on real process data. The machine learning system is preferably arranged in the computer system or a second computer system. The process data model may then be generated on the device, and if the process data model was generated on the second computer system, it may be transmitted to the computer system.
If the estimated process data deviates from the real process data by more than the threshold value, it is also possible to retrain the process data model and to update the process data model accordingly on the device and on the computer system.
The device may comprise an edge device, a field device a control device and/or a programmable logic controller (e.g., WAGO PFC controller, WAGO touch panel). Sensors and/or actuators may usually be connected to a device of this type, and the latter may be used to implement the various tasks in the context of automation technology.
The computer system may comprise, in particular, a cloud system and/or a server; the same also applies to the second computer system. The computer system and/or the second computer system is/are preferably located at a distance from the device. Examples of cloud systems are WAGO Cloud, Microsoft Azure, Amazon AWS, SAP Cloud and/or IMB Cloud, this list not intending to be exhaustive.
The invention also provides a computer program including instructions for implementing any of the above methods and their different aspects, as well as a device and a computer system for use in one of the described methods. Finally, the invention also relates to a system, which comprises a device and a computer system as described, the system, being configured to carry out any of the above methods.
Further scope of applicability of the present invention will become apparent from the detailed description given hereinafter. However, it should be understood that the detailed description and specific examples, while indicating preferred embodiments of the invention, are given by way of illustration only, since various changes, combinations, and modifications within the spirit and scope of the invention will become apparent to those skilled in the art from this detailed description.
The present invention will become more fully understood from the detailed description given hereinbelow and the accompanying drawings which are given by way of illustration only, and thus, are not limitive of the present invention, and wherein:
The currently preferred exemplary embodiments of a method according to the invention for the reduced data transmission between a device and a computer system are explained in greater detail below with reference to
In a preparation phase, which is illustrated in
Machine learning system 300 may be based on different algorithms, each of which is used depending on the problem. In one preferred implementation, a structurally predefined statistical model is trained on the basis of input/output data pairs. A portion of the data points iteratively runs through the model. The learning algorithm may adapt the model step by step, based on the deviation between the calculated results and the expected results. Suitable techniques from machine learning and artificial intelligence may be used for this purpose. One objective of a monitored learning of this type is that the network is trained to be able to establish associations after multiple computing passes using different inputs and outputs.
Prediction model 400 is subsequently distributed to edge device 100 in step S10, as illustrated in
At runtime (cf.
In step S20, a (continuous) comparison between actual process data D100 of the real process and estimated process data D200, which prediction model 400 generates, additionally takes place on edge device 100.
If a deviation between real process data D100 and process data D200 of prediction model 400 is determined on edge device 100 (it being possible to set a certain deviation tolerance here using a threshold value), this means that predicted process data D200 are no longer correct within cloud system 200. To have correct process data present within cloud system 200 at any time, either real process data D100 must be transmitted to cloud system 200 in a case of this type (cf. step S25), or so-called reconstruction data D300 must be sent, which are used to reconstruct the process data from prediction model 400 (e.g., based on differences), so that these data subsequently correspond to real process data D100.
It is understood that the preparation phase illustrated in
Prediction model 400 may possibly also be retrained in step S30 and be updated on edge device 100 as well as in cloud system 200 in step S35 for the purpose of keeping its prediction accuracy as high as possible, which may decrease due to seasonalities or changed environmental conditions.
As illustrated in
The invention being thus described, it will be obvious that the same may be varied in many ways. Such variations are not to be regarded as a departure from the spirit and scope of the invention, and all such modifications as would be obvious to one skilled in the art are to be included within the scope of the following claims.
| Number | Date | Country | Kind |
|---|---|---|---|
| 10 2020 202 463.4 | Feb 2020 | DE | national |
This nonprovisional application is a continuation of International Application No. PCT/EP2021/050516, which was filed on Jan. 13, 2021, and which claims priority to German Patent Application No. 10 2020 202 463.4, which was filed in Germany on Feb. 26, 2020, and which are both herein incorporated by reference.
| Number | Date | Country | |
|---|---|---|---|
| Parent | PCT/EP2021/050516 | Jan 2021 | US |
| Child | 17895839 | US |
