Artificial intelligence for monitoring continuous flow engines and continuous operation devices

Abstract

Applications for monitoring complex devices are used in long term applications, and a large amount of data resulting from a plurality of data streams is generated, so that processing such data becomes a difficult and highly demanding task. The method enables significant support of persons responsible for interacting, monitoring and controlling with the complex device, providing not only improved efficiency, but also improved security. Furthermore, the invention relates to a corresponding monitoring device, an upgrade kit, a computer program product and a storage device comprising such a computer program product.

Description

[0001] The present method relates to a method for monitoring a continuous data generating device such as a continuous flow engine. Furthermore, the present invention relates to a monitoring device suitable for implementing the method of the present invention. Furthermore, the present invention relates to an upgrade kit comprising the monitoring device of the present invention. Furthermore, the present invention relates to a computer program product comprising instructions that can be run to cause a computing entity to perform the method of the present invention. Furthermore, the present invention relates to a storage device for providing the computer program product of the present invention.

[0002] The industrial sector uses a wide variety of equipment. Many or even most of the equipment is used intermittently. However, in many industrial sectors, continuously operated equipment essentially represents the backbone on which the corresponding industrial sector relies. Such continuously used equipment, which provides uninterrupted operation for months or even longer, offers different possibilities and challenges compared to intermittently used equipment. For example, intermittently used equipment can be repeatedly checked and restarted. However, such continuously operated equipment does not allow for simple shutdown without good reasons. Continuous operation usually generates a large amount of data and experience acquired over years and decades. This allows technicians who use such equipment to detect subtle changes early, thus anticipating problems and correctly responding to specific needs at a specific point in time. However, upgrades and more specific adjustments based on very specific needs and applications lead to constant changes, making it difficult for even experienced technicians to keep up to date. The next generation of experts is left alone with the huge amount of collected data and the task of, for example, correctly judging whether certain historical data is relevant for a specific application. Not only the specific historical data must be taken into account, but also the maintenance and upgrades that have been carried out in the meantime.

[0003] For example, continuous flow engines are very important and highly complex devices in modern industry. They are used in a variety of applications and perform their tasks for years and decades. They are subject to maintenance and upgrades, but remain essentially the same. Although the control of such continuous flow engines has naturally developed very well, they are still developing at a faster pace today. The new available possibilities (such as additive manufacturing) that can provide a small number of specially customized parts effectively provide the possibility and challenge of adapting such continuous flow engines more to very specific applications. It is possible to continuously improve the benefits obtained thereby. However, at the same time, people who use such continuous flow engines are faced with solving the above challenges. The increased risk of damage or danger or more rigorous use makes it very limited, for example, the possible benefits of upgrading, or in the worst case even reduces the output of such equipment to below the value before the upgrade.

[0004] The field of power generation and distribution is a known example of conservative equipment treatment, since reliability and safety are the primary goals that need to be ensured. For example, circuit breakers and power grids used in such fields are basically in continuous operation. Often, replacements are not available, or some local solutions for rerouting some power or using some removable solutions are available only for a short time, and providing such methods usually requires a lot of time and effort to prepare. Identifying problems in time to include them in some scheduled annual maintenance or similar work is a demanding task. And whether some deviations from the typical behavior can be ignored, whether they may be caused by a past upgrade of some equipment within the managed power grid, or whether some immediate action is needed to shut down the equipment safely before a serious danger to people occurs.

[0005] This and other problems are solved by the products and methods disclosed below and in the claims. It should be noted that the above problems are particularly prominent in industries related to energy production and the use of continuous flow engines, of course especially including the use of continuous flow engines in the field of energy production. Such applications can be unexpectedly well supported by the solutions specified below. Further beneficial embodiments are disclosed in the dependent claims and in the further description and drawings. These benefits can be used to adapt the corresponding solutions to specific needs or to solve additional problems.

[0006] According to one aspect, the present invention relates to a method for monitoring an industrial plant comprising at least one continuous flow engine providing a current state, a power plant comprising at least one continuously operating device providing a current state, or a distribution facility comprising at least one continuously operating device providing a current state, wherein the method uses a monitoring device and a first user interface, wherein the monitoring device comprises artificial intelligence, wherein the artificial intelligence uses historical data, simulation data and interactive data, wherein the historical data represents data collected during the use of at least one continuous flow engine or a similar continuous flow engine or at least one continuously operating device or a similar continuously operating device, and the historical data is stored in a historical database, wherein the simulation data represents data obtained by running a simulation of the physical properties and / or behavior of the continuous flow engine on a processing unit. The method comprises the following steps: providing evaluation data related to at least one continuous flow engine or at least one continuously running device to a first user interface, wherein the evaluation data comprises at least one operation suggestion and reference data, wherein at least one operation suggestion comprises data about an actionable action that the artificial intelligence suggests to be implemented, wherein the reference data comprises data about the reason why the artificial intelligence provides at least one operation suggestion, and the at least one operation suggestion comprises at least one reference to historical data, simulation data and / or interaction data.

[0007] Surprisingly, it has been noted that the inclusion of such data about the reasons enables a significant simplification of the evaluation of such suggested actions by a person responsible for interacting with such a device or a technician who is not familiar with a specific device. In the practical application of artificial intelligence for such a device, there is the problem of deciding whether to follow the respective suggestion. There are a variety of options to choose from to provide additional information, such as simulated data of the possible consequences of such an action or data about the benefits that may be obtained thereby. However, putting such a solution into practice has shown that such apparently simple data forwarded along the respective suggestions actually show a very beneficial improvement of such a solution for a large number of typical application cases.

[0008] As used herein, the term "power plant" refers to conventional power generation facilities, such as power plants using steam turbines and gas turbines, as well as wind-based power generation facilities and solar-based facilities.

[0009] As used herein, the term "continuously operating equipment" refers to equipment that operates continuously for a long period of time, which generates a continuous stream of data related to its operation. Such a long period of time is preferably at least one month, more preferably at least three months, even more preferably at least one year. Examples of such continuously operating equipment are circuit breakers for power distribution, hydrolyzers for converting renewable energy into hydrogen, or turbines for wind-based power plants.

[0010] The term "continuous flow engine" as used herein refers to a device that uses a continuous flow of a fluid such as a gas or liquid. Herein, such a continuous flow engine generally provides a rotor that is located in the fluid and interacts with the fluid. Herein, such a fluid can be used to provide a rotational motion of the rotor, which can be converted into, for example, electrical energy. Examples of such continuous flow engines are gas turbines and steam turbines. Alternatively, the rotor can be rotated flexibly so that, for example, a fluid can be compressed. An example of such an application is a compressor, such as a compressor used in an oil refinery.

[0011] The phrase "similar continuous flow engine" refers to a continuous flow engine that provides similar technical performance compared to the continuous flow engine in question. For example, it refers to other continuous flow engines of the same model series. Such continuous flow engines of the same model series are established based on the wide application of continuous flow engines in the field. For example, Siemens' SGT-800 series represents a model series of gas turbines distributed in many countries and has been established to provide a reliable backbone for power generation.

[0012] The phrase "similar continuously operating equipment" refers to continuously operating equipment that offers similar technical characteristics compared to the continuously operating equipment in question. For example, it refers to other continuously operating equipment of the same model series.

[0013] The term "current state" as used herein refers to the current state in which the corresponding device (such as a continuous flow engine) is running. The term current state in this context includes the operating state of such a device as well as operating data and state transitions. Generally, the current state preferably particularly includes state transitions. In typical application cases, the current state includes at least two of the above information, more preferably all three, and the operating data in this context refers to conditions such as oil pressure, power output, cooling air temperature, etc.

[0014] In this article, the database can be located on a single server. However, it is usually preferred that the database is located in different locations. For example, it is usually preferred to locate the operator database and the simulation database in different locations. Although this requires additional work, it enables full use of the usually separate expertise and processing power required because the simulation database is usually preferably provided by a third party, and the third party needs to invest a lot of effort and processing power to provide such a database for a large number of users. And practical experience shows that users of such continuous flow engines are usually reluctant to share their insights into specific operations and hope to keep such expertise in-house. Resulting in such an operating method and a monitoring device that provides corresponding adjustments have unexpected benefits.

[0015] The term "monitoring device" as used herein refers to a physical device that contains at least one processing unit and data storage. Although the components of such a device may be placed in different locations within an industrial plant, they all work together and are managed together.

[0016] According to another aspect, the invention relates to a monitoring device suitable for implementing the method according to the invention, wherein the monitoring device comprises at least one data memory and at least one processing unit, wherein at least one data memory comprises an artificial intelligence.

[0017] According to another aspect, the present invention relates to an upgrade kit comprising a monitoring device according to the invention, wherein the upgrade kit is suitable for replacing a conventional monitoring device without artificial intelligence. It should be noted that it is very beneficial to introduce a monitoring device according to the invention into existing industrial plants comprising continuous flow engines. In this context, it should be noted that it is very easy to use such a monitoring device according to the invention in different industrial plants. In particular, it should be noted that surface retraining, typically using, for example, simulated data and / or historical data, enables the artificial intelligence to be fully adjusted so that it can be used in even very different industrial plants.

[0018] According to another aspect, the invention relates to a computer program product tangibly embodied in a machine-readable storage medium, the computer program product comprising instructions executable to cause a computing entity to perform the method of the invention.

[0019] According to another aspect, the invention relates to a storage device for providing a computer program product of the invention, wherein the device stores the computer program product and / or provides the computer program product for further use.

[0020] To simplify the understanding of the present invention, reference is made to the following detailed description and the accompanying drawings and their descriptions.Herein, the drawings should be understood not to limit the scope of the present invention, but to disclose preferred embodiments to further explain the present invention.

[0021] Figure 1 A schematic diagram of a system for implementing the method of the present invention is shown.

[0022] Preferably, unless otherwise stated, the embodiments below include at least one processor and / or data storage unit to implement the methods of the present invention.

[0023] Unless otherwise stated, terms such as "compute", "process", "determine", "generate", "configure", "reconstruct" and similar terms refer to actions and / or processes and / or steps of modifying data and / or generating data and / or transforming data, wherein the data is presented as physical variables or is otherwise usable.

[0024] The term "data storage" or similar terms as used herein refers, for example, to a temporary data storage, like a RAM (Random Access Memory) or a long-term data storage, like a hard drive or a data storage unit, like a CD, DVD, USB stick, etc. Such a data storage may additionally comprise or be connected to a processing unit to enable processing of the data stored on the data storage.

[0025] In the following, the invention will refer exemplarily to continuous flow engines such as gas turbines. It should be noted that the application of the invention in such fields is particularly beneficial. For example, corresponding continuous flow engines are often used as basic power supply units in power generation and distribution networks, which additionally have to handle fluctuations caused by uneven power generation from renewable energy sources.

[0026] According to one aspect, the invention relates to a method as described above.

[0027] In addition, it should be noted that using corresponding data to train artificial intelligence is beneficial for typical applications. According to another embodiment, it is preferred that at least training history data, training simulation data and training interaction data are used to train the artificial intelligence system, wherein the training history data is data collected during the use of a continuous flow engine or a similar continuous flow engine, and the training history data is stored in a training history database, wherein the training simulation data is data obtained by running a model of the physical characteristics and / or behavior of the continuous flow engine on a processing unit, and the training simulation data is stored in a training simulation database, wherein the training interaction data is obtained by monitoring the interaction during the operator of the control of a real continuous flow engine and / or a simulated continuous flow engine, and the training interaction data is stored in a training interaction database. It should be noted that the use of correspondingly trained artificial intelligence has provided a significantly improved starting point that has been able to solve many problems for many applications, which can be trained in real time using locally available stored data, supporting general-purpose trained artificial intelligence to solve specific needs of specific applications.

[0028] According to a further embodiment, it is preferred that the method comprises a step of retraining the artificial intelligence, preferably, wherein the artificial intelligence is retrained during the use of at least one continuous flow engine or at least one continuously operating device. It is noted that, based on the additional data provided, persons interacting with such a continuous flow engine or continuously operating device are able to maintain consistent and reliable work, even in cases where the artificial intelligence is retrained resulting in, for example, different suggestions or a different ranking of the suggestions provided. Even in cases where the retraining performed during the use of such a device results in spontaneous changes, it is possible to ensure the safe operation of the device quite easily. The possibility of implementing such a retraining while requiring little effort enables, for example, a reduction in downtime and other interruptions arising from tasks that require awareness to ensure that all interacting persons are aware of such changes and do not misinterpret the respective different outputs of the respective monitoring devices.

[0029] According to a further embodiment, preferably, the method comprises the step of receiving feedback data from an operator regarding the evaluation data, wherein the feedback data is used to retrain the artificial intelligence, preferably wherein the feedback data is used by the monitoring device to retrain the artificial intelligence during use of the at least one continuous flow engine or the at least one continuously running device. Using such an improvement cycle enables very efficient and easy improvement of the monitoring device of the present invention.

[0030] According to another embodiment, preferably, the method comprises the step of receiving feedback data about the evaluation data from the operator, wherein the feedback data is used to positively or negatively enhance future evaluation data. Herein, such enhanced future evaluation data can be achieved by retraining the artificial intelligence and / or adjusting the subsequent processing process that adjusts the evaluation data. For example, such adjustment to the evaluation data can be a change in the probability of following the suggestion or screening suggestion.

[0031] According to another embodiment, preferably, the artificial intelligence is a recurrent neural network. It should be noted that the corresponding artificial intelligence is generally particularly useful for many applications. Thereby further improving the processing speed and reliability. It should be particularly noted that such artificial intelligence can very effectively provide corresponding suggestions and generate reference data to be forwarded together.

[0032] According to another embodiment, it is preferred that the monitoring device forwards the evaluation data to a second user interface. Generally, it is preferred that the evaluation data forwarded to the second user interface is different from the evaluation data sent to the first user interface. For example, the evaluation data can be automatically split into a first part related to monitoring the active operation of at least one continuous flow engine or at least one continuously running device and a second part related to long-term monitoring and maintenance scheduling related to strategic operators by artificial intelligence. It should be noted that the method of the present invention is capable of automatically splitting the data generated by artificial intelligence to meet different needs and directly forwarding the data accordingly. The application and efficiency of the method of the present invention are further increased.

[0033] According to a further embodiment, it is preferred that the method comprises the step of receiving feedback data from an operator regarding the evaluation data, wherein the feedback data is stored for eventual inclusion in future evaluation data provided by the monitoring device. Typically, it is preferred that the feedback data comprises explanation data, wherein the explanation data comprises information as to why the operational suggestion was considered correct or incorrect. Storing such data enables the construction of a new historical database derived from the improved insights of the method of the present invention. Such a new historical database is indicated to provide significantly improved quality compared to existing historical databases, enabling further improvements in future processes and increasing the likelihood of future evaluations.

[0034] According to a further embodiment, preferably, the method comprises the step of receiving feedback data from an operator regarding the evaluation data, wherein the monitoring device comprises a second artificial intelligence, wherein the second artificial intelligence is adapted to process the feedback data and assign the feedback to future evaluation data. Although already available artificial intelligence can be used, it is noted that it is generally preferred to include such a second artificial intelligence for this explicit task to further improve the processing of such feedback data.

[0035] According to a further embodiment, it is preferred that the current state provides a fault associated with at least one continuous flow engine or at least one continuously running device, wherein the evaluation data contain the cause of the fault. It should be noted that in typical application cases, the relevant faults can be included with high reliability and relatively little effort (such as the processing power required to implement such a solution). At the same time, it should be noted that a person who interacts with the corresponding complex device can usually make full use of such data to further improve subsequent operations. The surprisingly very limited effort required to implement such a solution makes it generally possible to integrate this solution without any problems and as a standard solution.

[0036] According to a further embodiment, it is preferred that the evaluation data contain historical data relevant to the current state. It should be noted that by including such data in the evaluation data, an operator viewing the evaluation data can be surprisingly well supported. The inventors have found that artificial intelligence can very easily identify the relevant historical data and directly include it. And feedback from tests has shown that including such data significantly simplifies the work of operators responsible for monitoring and controlling industrial plants.

[0037] According to another embodiment, preferably, the artificial intelligence is suitable for providing evaluation data related to predictive maintenance. It should be noted that for typical application cases, the method of the present invention can provide corresponding data very effectively.

[0038] According to a further embodiment, preferably, the method comprises the step of receiving an implementation request from the first user interface, wherein the implementation request triggers the monitoring device to implement an actionable action in the industrial plant. A further development of the invention not only provides suggestions for one or more actionable actions, but also implements the corresponding actions as required. Thus, for example, a person responsible for operating a complex device can simply select from the suggestions provided and trigger the corresponding action without additional work.

[0039] According to a further embodiment, preferably, the method comprises the following steps: the monitoring device evaluates the operation suggestion with respect to its potential impact on the operational safety of the at least one continuous flow engine or at least one continuously operating device, wherein, in case the evaluation of the operation suggestion will not negatively impact the safety of the at least one continuous flow engine or at least one continuously operating device, the monitoring device automatically implements the operation suggestion. Including such an evaluation may even automate the operation of such a complex device, at least partially enabling a further reduction in the effort required of a person interacting with such a complex device, such as operating such a complex device.

[0040] According to a further embodiment, preferably, the method comprises a step of artificial intelligence identifying a fault in at least one continuous flow engine or at least one continuously operating device, wherein the artificial intelligence includes data related to the fault in the evaluation data, includes a solution for such a fault in the evaluation data and / or accordingly marks the event to be stored in a database.

[0041] According to a further embodiment, preferably, the method comprises the step of storing a copy of the artificial intelligence, wherein the method comprises the step of monitoring the device for a malfunction based on the artificial intelligence in an active state and / or replacing the artificial intelligence in an active state with the stored artificial intelligence when a replacement signal is received from the first user interface. It is surprisingly easy to implement such an option. At the same time, the possibility to easily replace the artificial intelligence with a safe copy of an earlier version very easily enables to react immediately in cases where the output of the artificial intelligence seems to provide less beneficial results. Moreover, after an upgrade or a major maintenance action, it may be preferred to fall back to such a safe version in order to avoid that an unsuitable optimized artificial intelligence suffers from being over-optimized for a specific application.

[0042] According to a further embodiment, preferably, the method comprises a step of storing a copy of the artificial intelligence, wherein the method comprises a step of retraining the stored artificial intelligence while the artificial intelligence in active state is actively monitoring the industrial plant. It is noted that such parallel retraining of the artificial intelligence is a very beneficial feature, making it possible to optimize the output of the monitoring device even without stopping the at least one continuous flow engine or the at least one continuously running device. The retraining is preferably based on simulation data. For example, such simulation data can be retrieved from a remote database. The possibility of deriving optimized simulation data from a third party (such as an expert in continuous flow engines or continuously running devices) makes it possible to fully use such external expertise. This is particularly useful for the method of the present invention, which uses artificial intelligence to deal with such highly sensitive problems (such as monitoring and even ultimately controlling continuous flow engines such as gas turbines). Even small problems can easily accumulate into serious problems and dangers, and special attention needs to be paid to such applications. At the same time, the method of the present invention makes it possible to customize the monitoring device for very specific uses and to retrain the artificial intelligence thereafter according to new insights or even changes in the industrial plant that affect the operation of the at least one continuous flow engine or the at least one continuously running device.

[0043] According to another embodiment, it is preferred that the artificial intelligence of the monitoring device screens the operating data of the industrial plant to be provided to the first user interface, wherein the first user interface is suitable for monitoring by an operator monitoring at least one continuous flow engine or at least one continuously running device, wherein the monitoring device actively screens at least 50%, more preferably at least 70%, and even more preferably at least 90% of the data to be displayed on the first user interface. It should be noted that the method of the present invention even allows such strict restrictions to be set without compromising the operation or safety of complex equipment. At the same time, the strict reduction of the data provided enables the operator to reliably focus his attention and time to monitor additional complex equipment or take over additional tasks. In contrast, flexible systems can easily lead to different equipment that need to be paid attention to at the same time being monitored. There is the potential for serious losses.

[0044] The phrase "filtering operational data" refers to evaluating whether the corresponding data should be shown on the first user interface. For example, the filtered data may not be displayed on the first user interface, or may be marked by graying it out or removing the highlight indicating relevant data. Although it is usually preferred to limit a certain amount of data that is not filtered and therefore represents very important core data, the method of the present invention using the monitoring device enables unexpectedly significant reduction of this amount, allowing the operator to focus on the information that is important at the time.

[0045] According to another aspect, the invention relates to a monitoring device as described above.

[0046] According to another aspect, the invention relates to an upgrade kit as described above.

[0047] According to another aspect, the invention relates to a storage device as described above.

[0048] The following detailed description of the drawings uses the drawings to discuss illustrative embodiments, which should not be construed as limiting, as well as features and further advantages thereof.

[0049] Figure 1 A schematic diagram of a system for implementing the method of the present invention is shown. The method is used to monitor three continuous flow engines. The continuous flow engines 2, 2', 2" in this case are gas turbines in a power plant for generating electricity. The three continuous flow engines 2, 2', 2" communicate with a monitoring device 2, which includes an artificial intelligence 4, which in this case is a recurrent neural network. Herein, current state data 3, 3', 3" about the current state received from the continuous flow engines 2, 2', 2" are continuously processed by the monitoring device 1. In order to perform the processing, the artificial intelligence retrieves historical data from a historical database 5, retrieves simulation data from a simulation database 6, and retrieves interactive data from an interactive database 7.

[0050] The historical database 5 contains data collected over many years of operation of the continuous flow engine 2, 2', 2" shown and other engines used in different power plants. The data was collected during the use of the continuous flow engine 2, 2', 2" or similar continuous flow engines. It contains normal operation as well as extreme conditions and fault data that occurred in the past time.

[0051] The simulation database 6 contains simulation data generated by running a model of the physical characteristics and / or behavior of the continuous flow engine on a processing unit. It also allows to understand very extreme situations that will never occur in reality. For example, serious operating errors due to wrong command options, etc. In addition, it can also include situations such as failure of important components that may cause serious damage or even danger to people in the vicinity of the continuous flow engine 2, 2', 2".

[0052] The interaction database 7 contains interaction data about the interaction between the continuous flow engine 2, 2', 2" and the people working therewith. For example, by monitoring the interaction of operators monitoring a real continuous flow engine and / or during a simulation of a continuous flow engine. Using the simulation data enables further understanding of the possible reactions of experienced operators. And whether the expected reaction is good or bad. It may even be helpful to solve the problems that arise, which require the use of different coping solutions.

[0053] The output of the artificial intelligence comprises evaluation data 9 relating to the three continuous flow engines 2, 2', 2", which are forwarded to a first user interface 8. The evaluation data comprises at least one operation suggestion comprising data on the actionable action that the artificial intelligence proposes to implement. Furthermore, it comprises reference data comprising data on the reasons why the artificial intelligence provides at least one operation suggestion comprising at least one reference to historical data, simulation data and / or interaction data.

[0054] In order to enable further reduction of downtimes and to ensure the greatest possible adaptation options, the artificial intelligence 4 is suitable for retraining during the use of the three continuous flow engines 2, 2', 2". This enables the correction of unanticipated developments of the artificial intelligence, as well as the adaptation of the artificial intelligence to new requirements resulting from changes originating from maintenance and upgrades. For example, for such actions, a copy of the artificial intelligence 4 can be stored. For example, in the event that the operation of the monitoring device 1 as described herein becomes unstable, the operator can trigger the copy to be used to replace the artificial intelligence with the previous version. In addition, such a copy of the artificial intelligence 4 can be run in parallel for a short period of time to obtain similar evaluation data, making it easy to compare changes in the operation of the artificial intelligence 4 before and after the change. It is also unexpectedly beneficial that the artificial intelligence 4 is evaluated not after such a retraining, but simply based on the process of continuous learning and development of the artificial intelligence 4 over time, resulting in reasonable lagging deviations from the expected performance.

[0055] In addition, the monitoring device is suitable for using feedback data 12. An operator using the first user interface 8 is able to input feedback data 12. The feedback data 12 is forwarded to the second artificial intelligence 13, which processes the feedback data to make full use of it. The second artificial intelligence continuously builds a feedback database 14 containing the high-quality data generated thereby. Compared with the feedback data stored in an unprocessed manner or the historical data of the historical database 5, it is noted that the data contained in the feedback database 14 provides significantly higher quality and reliability. Doing such additional work is very meaningful for further increasing future possibilities and decisions. In addition, the feedback database 14 is also suitable for communicating directly with the artificial intelligence 4, so that the artificial intelligence 4 can continuously further increase its evaluation data 9 and increase the feedback obtained over time. In this article, specific parts of the evaluation data are positively or negatively enhanced.

[0056] Furthermore, the operator using the user interface 8 can also select a suggested option from the evaluation data 9 simply by clicking on a button. This very quick possibility is made possible because a longer decision process is possible due to the already very high quality of the suggested actions and the increased insight the operator has based on the reference data provided. While normally such options should only be available for small operations that do not pose a risk to the operation as a whole, the present method makes it possible to suggest even complex operating tasks summarizing multiple steps that can be easily evaluated as a whole for use. Based on the selection, an implementation request 15 is sent to the monitoring device 1, which further processes the implementation request 15. In this context, the monitoring device can be adapted to directly implement the corresponding operation. In the present case, the monitoring device 1 sends the corresponding data to the control units of the three continuous flow engines 2, 2', 2". The control unit is not shown in the accompanying drawings.

[0057] However, the monitoring device 2 not only sends the evaluation data to the first user interface 8. Additional evaluation data 11 is also sent to the second user interface 10. In this context, the additional evaluation data includes at least one operation evaluation and at least one operation suggestion and reference data, wherein the additional evaluation data is different from the evaluation data. In this context, at least one operation suggestion includes data about the feasible actions that the artificial intelligence recommends to implement, and at least one operation evaluation includes data about the evaluation of the feasible operations and their results. The reference data includes data about the reasons why the artificial intelligence provides at least one operation suggestion, and the operation suggestion includes at least one reference to historical data, simulation data and / or interaction data. While the evaluation data sent to the first user interface 8 is directed to the operator operating the continuous flow engine 2, 2', 2", in such a case the additional evaluation data is directed to the decision-making members among the personnel responsible for using the continuous flow engine 2, 2', 2". For example, it involves possible changes in the maintenance plan or optimizing operations to increase the overall benefits by changing the operating mode accordingly.

[0058] Although most operating commands are triggered by the operator, the artificial intelligence 4 is already adapted to further reduce the burden on the operator by taking over safety decisions. In such a case, the monitoring device 1 evaluates the potential impact of the operating suggestion on the operational safety of the three continuous flow engines 2, 2', 2". In the case where the operating suggestion is deemed not to negatively affect the safety of the three continuous flow engines 2, 2', 2", the artificial intelligence 4 of the monitoring device 1 automatically implements the operating suggestion. This enables a significant reduction in the burden on the operator. However, the operator is still able to review the decision and ultimately include feedback data 12 to be processed accordingly, resulting in the feedback database 14 including that in similar future situations the artificial intelligence 4 should not take action on its own.

[0059] The present invention is further described in detail for the purpose of explanation only. However, the present invention should not be understood to be limited to these embodiments, because they represent embodiments that provide benefits of solving specific problems or satisfying specific needs. It should be understood that the scope of protection is limited only by the appended claims.

Claims

1. A method for monitoring an industrial plant comprising at least one continuous flow engine (2, 2', 2") providing a current state, a power plant comprising at least one continuously running device providing a current state, or a power distribution facility comprising at least one continuously running device providing a current state, in, The method uses a monitoring device (1) and a first user interface (8), The monitoring device (1) comprises artificial intelligence (4), Wherein, the artificial intelligence (4) uses historical data, simulation data and interactive data, wherein the historical data represent data collected during use of the at least one continuous flow engine (2, 2', 2") or similar continuous flow engines (2, 2', 2") or the at least one continuously operating device or similar continuously operating device, and the historical data are stored in a historical database, wherein the simulation data represent data obtained by running a model of the physical properties and / or behavior of the continuous flow engine (2, 2', 2") on a processing unit and the simulation data are stored in a simulation database, wherein the interaction data represent data obtained by monitoring the interaction of an operator during monitoring of a real continuous flow engine (2, 2', 2") and / or a simulated continuous flow engine (2, 2', 2"), and the interaction data are stored in an interaction database, The method comprises the step of providing the first user interface (8) with evaluation data (9) related to the at least one continuous flow engine (2, 2', 2") or the at least one continuously operating device, The evaluation data (9) includes at least one operation suggestion and reference data. wherein the at least one action suggestion comprises data on an actionable action that the artificial intelligence (4) suggests performing, The reference data includes data on the reasons why the artificial intelligence (4) provides the at least one operation suggestion, and the at least one operation suggestion includes at least one reference to the historical data, the simulation data and / or the interaction data.

2. A method according to any one of the preceding claims, wherein: The method comprises the step of retraining the artificial intelligence (4), preferably wherein: The artificial intelligence (4) is retrained during use of the at least one continuous flow engine (2, 2', 2") or the at least one continuously operating device.

3. A method according to any one of the preceding claims, wherein: The method comprises the step of receiving feedback data (12) from an operator regarding the evaluation data (9), The feedback data (12) is used to retrain the artificial intelligence (4).

4. A method according to any one of the preceding claims, wherein: The method comprises the steps of receiving feedback data (12) from an operator regarding the evaluation data (9), The feedback data (12) is used to positively or negatively enhance future evaluation data (9).

5. A method according to any one of the preceding claims, wherein: The artificial intelligence (4) is a recurrent neural network.

6. A method according to any one of the preceding claims, wherein: The monitoring device (1) forwards the evaluation data (9) or the additional evaluation data (11) to a second user interface (10), The additional evaluation data (11) includes at least one operation evaluation and / or at least one operation suggestion and reference data, wherein the additional evaluation data (1) is different from the evaluation data (9), wherein the at least one action suggestion comprises data on an actionable action that the artificial intelligence (4) suggests performing, wherein the at least one operational assessment comprises data regarding an assessment of feasible operations and their outcomes, The reference data includes data on the reasons why the artificial intelligence (4) provides the at least one operation suggestion, and the at least one operation suggestion includes at least one reference to the historical data, the simulation data and / or the interaction data.

7. A method according to any one of the preceding claims, wherein: The method comprises the steps of receiving feedback data (12) from an operator regarding the evaluation data (9), Therein, the feedback data (12) is stored for eventual inclusion in future evaluation data (9) provided by the monitoring device (1).

8. A method according to any one of the preceding claims, wherein: The method comprises the steps of receiving feedback data (12) from an operator regarding the evaluation data (9), The monitoring device (1) comprises a second artificial intelligence (13), Therein, the second artificial intelligence (13) is adapted to process the feedback data (12) and to assign the feedback to future evaluation data (9).

9. A method according to any one of the preceding claims, wherein: The artificial intelligence (4) is suitable for providing evaluation data (9) relevant for predictive maintenance.

10. A method according to any one of the preceding claims, wherein: The method comprises the step of receiving an implementation request (15) from the first user interface (8), wherein the implementation request (15) triggers the monitoring device (1) to implement the actionable action in the industrial plant.

11. A method according to any one of the preceding claims, wherein: The method comprises the following steps: the monitoring device (1) evaluates the operation proposal with respect to its potential impact on the operational safety of the at least one continuous flow engine (2, 2', 2") or the at least one continuously operating device, Wherein, in the event that the evaluation of the operational recommendation is deemed not to negatively affect the safety of the at least one continuous flow engine (2, 2', 2"), the monitoring device (1) automatically implements the operational recommendation.

12. A method according to any one of the preceding claims, wherein: The method comprises the step of storing a copy of the artificial intelligence (4), The method comprises the step of retraining the stored artificial intelligence (4) while the artificial intelligence (4) in an active state is actively monitoring the industrial plant.

13. A monitoring device (1), adapted to implement the method according to any one of claims 1 to 12, in, The monitoring device (1) comprises at least one data storage device and at least one processing unit, Therein, the at least one data storage device contains the artificial intelligence (4).

14. An upgrade kit comprising a monitoring device (1) according to claim 13, wherein: The upgrade kit is suitable for replacing a conventional monitoring device (1) without artificial intelligence (4).

15. A computer program product tangibly embodied in a machine-readable storage medium, the computer program product comprising instructions executable to cause a computing entity to perform the method according to any one of claims 1 to 12.

Images