JP7490924B2 - Apparatus and method for user-configured predictive plant maintenance - Google Patents

Description

The present application relates to a user-setting-based plant predictive maintenance device and method. More specifically, the present application relates to a plant predictive maintenance device and method that can be applied to plant predictive maintenance by importing a pre-processing process (or a pre-processing module) and an artificial intelligence model based on user settings.

In plant equipment, predictive maintenance is essential to predict various risks due to the condition of the equipment and take measures to maintain productivity.

Typically, there are two methods for maintaining equipment: the Time Based Maintenance (TBM) method, which performs maintenance after a certain period of time has elapsed, and the Condition Based Maintenance (CBM) method, which uses inspection data, measurement data, and diagnostic data on the equipment status to determine whether the conditions for equipment maintenance are met and performs maintenance based on the results of the determination.

In this case, predictive maintenance uses a monitoring system to track raw data such as temperature, displacement, current, and acceleration generated by the equipment in real time to check the performance and condition of the equipment. This is a CBM method that detects equipment abnormalities through the values of sensors attached to the equipment and takes maintenance measures only when maintenance and repairs are required.

On the other hand, when applying artificial intelligence to a predictive maintenance system, the artificial intelligence model or pre-processing model used is subordinate to the predictive maintenance system, making it impossible to select an artificial intelligence model or pre-processing model based on the plant or user preferences, and only the artificial intelligence model designated for each plant can be used, which is inconvenient.

Korean Patent Publication No. 10-1282244

The present application is intended to solve the problems of the prior art described above, and aims to provide a user-configured plant predictive maintenance device and method that can train an artificial intelligence model using multiple equipment data collected at the plant and simultaneously perform predictive maintenance on the plant.

The present application aims to solve the problems of the prior art described above and provide a user-configured plant predictive maintenance device and method that can import external pre-processing files and external artificial intelligence models and selectively apply them according to plant or user preferences.

However, the technical problems that the embodiments of this application aim to achieve are not limited to those described above, and other technical problems may exist.

As a technical means for achieving the above technical object, a user-setting-based plant predictive maintenance device according to one embodiment of the present application includes a collection unit that collects multiple equipment data for a plant from multiple equipment sensors, a pre-processing unit that performs pre-processing on the multiple equipment data based on an external pre-processing file imported by user settings, a learning unit that trains an internal artificial intelligence model by inputting the multiple equipment data pre-processed by the pre-processing unit, a management unit that manages model data for the internal artificial intelligence model trained by the learning unit or an external artificial intelligence model imported by user settings, and at least one of the artificial intelligence models stored in the management unit and the collection unit. and an operation unit that calculates predictive data for the maintenance of the plant using at least one of the multiple equipment data collected by the predictive maintenance device. The preprocessing unit includes an embedded preprocessing unit that performs preprocessing on the multiple equipment data using an internal preprocessing function provided within the plant predictive maintenance device and embedded in the plant predictive maintenance device, and a user preprocessing unit that performs preprocessing on the multiple equipment data according to a rule set by the user using the external preprocessing file, and the multiple equipment data is preprocessed by selectively applying one of the embedded preprocessing unit using the internal preprocessing function or the user preprocessing unit using the external preprocessing file.

According to one embodiment of the present application, the system further includes a storage unit that stores the plurality of equipment data collected by the collection unit in real time or periodically according to a preset cycle, and the preprocessing unit can read at least a portion of the plurality of equipment data stored in the storage unit and perform preprocessing.

According to one embodiment of the present application, the management unit can import an external artificial intelligence model existing outside the device in file format, format it into a usable format, and upload it.

According to one embodiment of the present application, the management unit can store and provide the model data for at least one of the model name, algorithm used, characteristics, verification method, hyperparameter values, applied preprocessing method, and whether or not it is an internal/external model for the internal AI model or the external AI model.

According to one embodiment of the present application, the predictive data may include at least one of anomaly detection for the plant, prediction of equipment failure life, prediction of failure probability, and prediction of maintenance timing.

According to one embodiment of the present application, the operation unit can input at least one of the multiple equipment data stored in the storage unit into at least one of the artificial intelligence models stored in the management unit to calculate the prediction data.

According to one embodiment of the present application, the system may further include a visualization unit that visualizes the prediction data calculated by the operation unit so that it can be monitored.

As a technical means for achieving the above technical object, a user-setting-based plant predictive maintenance method according to one embodiment of the present application includes a step of collecting a plurality of equipment data for a plant from a plurality of equipment sensors, a step of preprocessing the plurality of equipment data based on an external preprocessing file imported according to user settings, a step of training an internal AI model by inputting the plurality of equipment data preprocessed in the preprocessing step, a step of managing model data for the internal AI model trained in the training step or an external AI model imported according to user settings, and a step of managing the AI stored and managed in the model data management step. and calculating predictive data for the maintenance of the plant using at least one of the functional models and at least one of the equipment data collected in the step of collecting the equipment data. The pre-processing step includes: (a) performing pre-processing on the equipment data using an internal pre-processing function provided within the plant predictive maintenance device and incorporated in the plant predictive maintenance device; and (b) performing pre-processing on the equipment data according to a rule set by the user using the external pre-processing file. The pre-processing of the equipment data may be performed by selectively applying one of the steps (a) and (b).

According to one embodiment of the present application, the method further includes a step of storing the plurality of equipment data collected in the step of collecting the plurality of equipment data in real time or periodically according to a preset period, and the pre-processing step may read and pre-process at least a portion of the plurality of equipment data stored in the step of storing.

According to one embodiment of the present application, the step of managing the model data may involve importing an external artificial intelligence model existing outside the device in the form of a file, formatting it into a usable form, and uploading it.

According to one embodiment of the present application, the step of managing the model data may store and provide the model data for at least one of the model name, algorithm used, characteristics, verification method, hyperparameter values, applied preprocessing method, and whether the model is internal/external for the internal AI model or the external AI model.

According to one embodiment of the present application, the predictive data may include at least one of anomaly detection for the plant, prediction of equipment failure life, prediction of failure probability, and prediction of maintenance timing.

According to one embodiment of the present application, the step of calculating predictive data for the plant maintenance may include inputting at least one of the equipment data stored in the storing step into at least one of the artificial intelligence models stored in the model data managing step to calculate the predictive data.

According to one embodiment of the present application, the method may further include a step of visualizing the calculated predictive data for the plant maintenance in a manner that allows monitoring.

The above-described problem solving means are merely exemplary and should not be construed as limiting the present application. In addition to the exemplary embodiments described above, there may be additional embodiments in the drawings and detailed description of the invention.

According to the above-mentioned problem-solving means of the present application, by learning an artificial intelligence model using multiple equipment data collected at the plant and at the same time performing predictive maintenance on the plant, it is possible to reduce downtime and improve the efficiency of plant operations.

According to the above-mentioned solution to the problem of the present application, it is possible to import external pre-processing files and external artificial intelligence models and selectively apply them according to the plant or user preferences, thereby expanding the predictive maintenance function and improving performance compared to when only the existing internal pre-processing function and internal artificial intelligence model are used.

However, the effects obtained by this application are not limited to those described above, and other effects may also exist.

1 is a schematic diagram of a user-configured plant predictive maintenance system according to an embodiment of the present disclosure; 1 is a schematic diagram of a user-configured plant predictive maintenance system according to an embodiment of the present disclosure; 1 is a diagram illustrating an AutoEncoder model, which is a deep learning-based unsupervised learning model that can be provided by a learning unit according to an embodiment of the present application. 1 is a diagram illustrating model data managed by a management unit according to an embodiment of the present application; 11 is a diagram illustrating a process of importing an external artificial intelligence model in a management unit according to an embodiment of the present application. 1 is a diagram illustrating an example of predicted data provided by a visualization unit according to an embodiment of the present disclosure. 4 is an operational flowchart of a user-configured plant predictive maintenance method according to an embodiment of the present disclosure.

Hereinafter, the embodiments of the present application will be described in detail with reference to the accompanying drawings so that a person having ordinary skill in the art to which the present application pertains can easily implement the present application. However, the present application may be embodied in various different forms and is not limited to the embodiments described herein. In order to clearly explain the present application in the drawings, parts that are not relevant to the description have been omitted, and similar parts have been given similar reference numerals throughout the specification.

Throughout this specification, when one part is said to be "connected" to another part, this includes not only when it is "directly connected" but also when it is "electrically connected" or "indirectly connected" via another element in between.

Throughout this specification, when a member is said to be located "on," "on top," "at the top end," "below," "at the bottom end," or "at the bottom end" of another member, this includes not only when the member is in contact with the other member, but also when another member is present between the two members.

Throughout this specification, when a part "comprises" certain elements, this means that it can further include other elements, not excluding other elements, unless specifically stated to the contrary.

Throughout this specification, "user setting" may be interpreted as meaning a setting value based on a user input received through a user terminal 30 described below, or a user input directly input through a wired or wireless communication device to a user-setting-based plant predictive maintenance device 100 according to one embodiment of this specification. In other words, throughout this specification, "user setting" refers to a setting value input by a user, and the input method therefor is not specifically limited.

This application relates to a user-setting-based plant predictive maintenance device and method. More specifically, this application relates to a plant predictive maintenance device and method that can be applied to plant predictive maintenance by importing a pre-processing process (or a pre-processing module) and an artificial intelligence model based on user settings.

Figure 1 is a schematic diagram of a user-configured predictive plant maintenance system according to one embodiment of the present application.

Referring to FIG. 1, the user-configured plant predictive maintenance system 10 may include a user-configured plant predictive maintenance device 100, a plant 20, and a user terminal 30.

In the following, for ease of explanation, the "user-configured plant predictive maintenance system 10" according to one embodiment of the present application will be identified as the "predictive maintenance system 10," and the "user-configured plant predictive maintenance device 100" according to one embodiment of the present application will be identified as the "predictive maintenance device 100."

Referring to FIG. 1, the predictive maintenance device 100 can communicate with the plant 20 and the user terminal 30 through a network. Specifically, according to one embodiment of the present application, the predictive maintenance device 100 can receive a plurality of equipment data from a plurality of equipment sensors included in the plant 20 through a network, can receive a user input for importing an external pre-processed file or an external artificial intelligence model from the user terminal 30, and can transmit prediction data for predictive maintenance calculated using an internal artificial intelligence model or an external artificial intelligence model to the user terminal 30.

At this time, the network includes 3GPP (registered trademark) (3rd Generation Partnership Project) network, LTE (Long Term Evolution) network, 5G network, WIMAX (registered trademark) (World Interoperability for Microwave Access) network, wired and wireless Internet (Internet), LAN (Local Area Network), Wireless LAN (Wireless Local Area Network), WAN (Wide Area Network), PAN (Personal Area Network), Bluetooth (registered trademark) network, Wi-Fi network, NFC (Near These may include, but are not limited to, a Digital Field Communication (DFT) network, a satellite broadcasting network, an analog broadcasting network, and a Digital Multimedia Broadcasting (DMB) network.

In addition, when an issue occurs, including when the predictive maintenance device 100 detects an abnormality in the plant 20 or determines that a situation requiring immediate maintenance has occurred based on the calculation results of the prediction data, the predictive maintenance device 100 can transmit an alarm signal for the issue that has occurred to the user terminal 30. At this time, the alarm signal can include information on the type of abnormality (problem) that has occurred, the location where the abnormality occurred, the time when the abnormality occurred, the urgency of the maintenance, and the reason why it has been determined that maintenance is necessary, but is not limited thereto.

The user terminal 30 may be, for example, a smartphone, a smart pad, a tablet PC, a wearable device, or a personal digital assistant (PDA), such as a personal digital assistant (PDAS), a personal digital assistant (PDC), a personal digital assistant (PHS ... This may be any type of wireless communication device such as W-CDMA (W-Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), or Wibro (Wireless Broadband Internet) terminals, as well as fixed terminals such as desktop computers and smart TVs.

Figure 2 is a schematic diagram of a user-configured plant predictive maintenance device according to one embodiment of the present application.

Referring to FIG. 2, the predictive maintenance device 100 may include a collection unit 110, a storage unit 120, a pre-processing unit 130, a learning unit 140, a management unit 150, an operation unit 160, and a visualization unit 170.

Referring to FIG. 2, the collection unit 110 can receive and collect multiple equipment data for the plant 20 from multiple equipment sensors included in the plant 20.

According to one embodiment of the present application, the equipment data may be data on at least one of temperature, displacement, acceleration, pressure, vibration, and flow rate, and the equipment sensors may include contact sensors attached to the plant 20 or non-contact sensors capable of measuring and sensing equipment data on at least one of the constituent equipment of the plant 20. In this case, if the equipment data to be collected (received) by the collection unit 110 is vibration data, the collection unit 110 may collect (receive) the data as values converted into the frequency domain or values converted into the time domain by each of the equipment sensors.

In other words, the collection unit 110 can receive multiple equipment data in real time from each of multiple equipment sensors that monitor the status of multiple pieces of equipment included in the plant 20.

Referring to FIG. 2, the collection unit 110 transmits the collected equipment data to the storage unit 120, and the storage unit 120 may receive the equipment data from the collection unit 110. At this time, the storage unit 120 may receive the equipment data from the collection unit 110 in real time and store the received equipment data in real time or periodically according to a preset period. However, the present invention is not limited thereto, and when the collection unit 110 transmits the equipment data to the storage unit 120 periodically according to a preset period, the storage unit 120 may store the received equipment data every time the collection unit 110 receives the equipment data. In other words, the storage unit 120 may store the equipment data collected by the collection unit 110 in real time or periodically according to a preset period.

In this case, the preset period may refer to a time period set by the user, and may be set and modified by the user before or during operation of the predictive maintenance device 100. For example, the predictive maintenance device 100 may receive a time period for collecting or storing multiple equipment data through the user terminal 30, and collect and store multiple equipment data according to the received time period.

In addition, the storage unit 120 may store the plurality of equipment data by dividing them into detailed categories, taking into consideration at least one of the plant type, plant size, data format, and the amount of data collected per unit time. For example, the storage unit 120 may store the plurality of equipment data by dividing them into structured data, unstructured data, and semi-structured data, taking into consideration the data format.

At this time, the structured data corresponds to a relational database, and may include, for example, temperature, pressure, and flow rate collected in the plant 20. In addition, the non-structured data may include, for example, noise data generated in the plant 20, equipment images, CCTV images, equipment-related documents such as PDF files, and plant drawing files, and the semi-structured data is data that includes an explanation of the structure in the data content, and may include, for example, data in file formats such as HTML (Hyper Text Markup Language), XML (extensible Markup Language), and JSON (Java Script Object Notation), and may more specifically include CAD shape information files, web log data, and signal data generated by sensors. However, the non-structured data is not limited thereto.

Referring to FIG. 2, the preprocessing unit 130 can read at least a portion of the equipment data stored in the storage unit 120 and perform preprocessing. Specifically, the preprocessing unit 130 can selectively preprocess data required to learn an artificial intelligence model, which will be described later, from among the equipment data stored in a manner classified according to the plant type, plant size, data format, and size of data collection per unit time.

The pre-processing unit 130 may include an embedded pre-processing unit 131 and a user pre-processing unit 132. The embedded pre-processing unit 131 performs pre-processing of multiple equipment data using an internal pre-processing function provided, i.e., embedded, within the predictive maintenance device 100 according to one embodiment of the present application, and the user pre-processing unit 132 can perform pre-processing of multiple equipment data based on an external pre-processing file imported according to user settings.

Specifically, the internal preprocessing functions may include preprocessing functions such as merging the provided data, handling missing values or outliers, interpolation (such as linear interpolation and multi-polynomial interpolation), adding columns, generating and converting columns, etc., and the built-in preprocessing unit 131 may preprocess (refine) multiple equipment data using such internal preprocessing functions.

Preprocessing methods for handling missing values include, for example, the Multi Polynomial Interpolation method, which is a method of interpolating using m values close to the missing value, rather than methods such as Replace and Linear Interpolation, which are common interpolation methods. The Polynomial Interpolation method is a well-known and well-understood technique among ordinary engineers, so a detailed explanation will be omitted below.

As described above, the user pre-processing unit 132 can import external pre-processing files that exist outside the predictive maintenance device 100, standardize and upload the external pre-processing files, and pre-process multiple pieces of equipment data using the uploaded external pre-processing files so that pre-processing can be performed according to rules set by the user using functions other than the pre-processing functions built into the predictive maintenance device 100. In other words, the user pre-processing unit 132 can be provided so that the user can utilize the external pre-processing functions as needed.

For example, the user preprocessing unit 132 can upload an external preprocessing file through Python (registered trademark) modularization and extension format, and can perform the external preprocessing file. In another example, when unit conversion is required for data processing and a specific conversion formula is required rather than a simple operation such as a constant multiplication function, the user preprocessing unit 132 can import and upload an external preprocessing file for this and use it. Specifically, when it is necessary to convert the temperature unit from Celsius scale to Fahrenheit scale, the user preprocessing unit 132 can import and upload a relational formula (°F = 1.8°C + 32) showing the relationship between Celsius and Fahrenheit according to the user's settings to convert the temperature unit.

Although not shown in the drawing, the equipment data preprocessed by the preprocessing unit 130 can also be stored in the storage unit 120 and can be used to operate the artificial intelligence model in the operation unit 160 described below. In addition, the equipment data preprocessed by the preprocessing unit 130 can be transmitted to the learning unit 140 and used to learn the internal artificial intelligence model.

The preprocessing unit 130 can import and upload external preprocessing files according to user settings and use the external preprocessing files to perform preprocessing of multiple equipment data, thereby helping the user to perform preprocessing taking into account various cases (situations), such as the type and format of the artificial intelligence model to be learned by the user, and the input values or forms of the input values required for the artificial intelligence model to be used for predictive maintenance.

Referring to FIG. 2, the learning unit 140 can learn an internal artificial intelligence model by inputting a plurality of pieces of equipment data preprocessed by the preprocessing unit 130. The learning unit 140 can create an artificial intelligence model using a plurality of pieces of preprocessed equipment data. In this case, the artificial intelligence model generated or learned by the learning unit 140 can be disclosed as an "internal artificial intelligence model" in the following description and should be interpreted as a concept distinct from an "external artificial intelligence model."

Artificial intelligence models can include supervised learning-based models, unsupervised learning-based models, machine learning-based models (such as random forests and SVMs (Support Vector Machines)), and deep learning-based models.

The learning unit 140 may provide, for example, an autoencoder model, which is an unsupervised learning model among deep learning-based models. FIG. 3 is a diagram illustrating an autoencoder model, which is an unsupervised learning model based on deep learning, that can be provided by a learning unit according to one embodiment of the present application.

In addition, the learning unit 140 can generate or learn an internal artificial intelligence model using multiple equipment data stored in the storage unit 120 or data preprocessed through the preprocessing unit 130, as well as verifying the internal artificial intelligence model that has already been generated and learned.

Specifically, the learning unit 140 classifies multiple equipment data (including preprocessed data) into learning data necessary for generating and learning an AI model and verification data necessary for verifying an AI model that has already been generated and learned, generates or learns an AI model using the learning data, and evaluates the performance of an AI model that has already been generated and learned using the verification data. The learning unit 140 can store the generated, learned, and verified AI model by transmitting it to the management unit 150.

Referring again to FIG. 2, the management unit 150 can receive and store the artificial intelligence model generated, learned, and verified from the learning unit 140, i.e., the internal artificial intelligence model, and store and manage internal model data for the stored internal artificial intelligence model. The management unit 150 can also import and store an external artificial intelligence model according to user settings, and store and manage external model data for the stored external artificial intelligence model. In other words, the management unit 150 can manage model data for the internal artificial intelligence model learned by the learning unit 140 or the external artificial intelligence model imported according to user settings.

Specifically, the model data managed by the management unit 150 may include at least one of the model name, model type, algorithm used, characteristics, scale, verification method, hyper parameter value, applied preprocessing method, and whether or not it is an internal/external model (model registrant) for the AI model.

Figure 4 is a diagram illustrating an example of model data managed by a management unit according to one embodiment of the present application. Referring to Figure 4, it can be seen that two or more artificial intelligence models are stored and managed together with a plurality of model data. However, this is not limited thereto, and it is preferable that Figure 4 be interpreted as illustrating one embodiment of the management unit 150.

According to one embodiment of the present application, the management unit 150 may include an internal model management unit 151 and an external model management unit 152. Specifically, the internal model management unit 151 may store the internal AI model generated, learned, and verified from the learning unit 140 as described above, and manage the internal model data for the stored internal AI model.

In addition, the external model management unit 152 can import, in file format, an external artificial intelligence model that exists outside the predictive maintenance device 100, i.e., that has not been generated or learned by the learning unit 140, format it into a usable form, upload it, store it, and manage the external model data for the stored external artificial intelligence model.

Figure 5 is a diagram illustrating a process of importing an external AI model in a management unit according to one embodiment of the present application. Referring to Figure 5, the external model management unit 152 can import various external AI models that already exist, standardize them into a form that can be used for predictive maintenance, and upload and store them.

According to one embodiment of the present application, the predictive maintenance device 100 can selectively apply external pre-processing files and external artificial intelligence models according to the plant 20 or user preferences, as described above, and can import and upload them according to the plant 20 or user preferences. This allows the predictive maintenance device 100 to expand predictive maintenance functions and improve performance compared to when only the existing internal pre-processing functions and internal artificial intelligence models are used.

Referring also to FIG. 2, the operation unit 160 can receive at least one of the artificial intelligence models stored in the management unit 150 and at least one of the equipment data collected by the collection unit 110 and stored in the storage unit 120, and can use the received artificial intelligence model and equipment data to calculate predictive data for the maintenance of the plant 20.

Specifically, the operation unit 160 receives at least one artificial intelligence model to be used for predictive maintenance from the management unit 150 among the multiple internal artificial intelligence models or external artificial intelligence models stored in the management unit 150, and receives multiple equipment data stored in the storage unit 120 or multiple equipment data preprocessed through the preprocessing unit 130 (not shown in FIG. 2) for use as data to be applied (input) to the transmitted artificial intelligence model.

The operation unit 160 also inputs and applies the transmitted equipment data to the transmitted artificial intelligence model, making it possible to calculate the prediction data. In other words, the operation unit 160 inputs at least one of the multiple equipment data stored in the storage unit 120 to at least one of the artificial intelligence models stored in the management unit 150, making it possible to calculate the prediction data.

For example, when performing predictive maintenance for problems that may occur due to temperature changes in the plant 20, the operation unit 160 can receive an artificial intelligence model capable of predicting temperature changes and problems that may occur due to temperature changes, as well as temperature data for each component of the plant 20, and input the transmitted temperature data into the transmitted artificial intelligence model, thereby enabling prediction data to be calculated from the artificial intelligence model.

At this time, the prediction data may include at least one of abnormality detection for the plant 20, prediction of equipment failure life, prediction of failure probability, and prediction of maintenance timing.

According to one embodiment of the present application, the operation unit 160 receives and uses at least one of the artificial intelligence models already stored and managed by the management unit 150, rather than the internal artificial intelligence model generated and learned by the learning unit 140. Therefore, while the internal artificial intelligence model is being learned in the learning unit 140, predictive data can be calculated, i.e., predictive maintenance can be performed, using at least one of the multiple internal artificial intelligence models and multiple external artificial intelligence models stored and managed by the management unit 150.

In other words, the predictive maintenance device 100 can learn an artificial intelligence model using multiple equipment data collected in the plant 20 and at the same time perform predictive maintenance on the plant 20, thereby reducing downtime and improving the efficiency of plant operations.

Referring to FIG. 2, the visualization unit 170 can receive the predicted data calculated from the operation unit 160. The visualization unit 170 can visualize the predicted data calculated by the operation unit 160 so that it can be monitored.

FIG. 6 is a diagram illustrating an example of predicted data provided by a visualization unit according to an embodiment of the present application. As shown in FIG. 6, the visualization unit 170 can visualize the predicted data in various forms, such as a line graph or a bar chart, depending on an artificial intelligence model using the predicted data or a type of equipment data. In addition, the visualization unit 170 can transmit the visualized predicted data to a user terminal 30 or a PC including a display that can be directly monitored by a user. In addition, the visualized predicted data can be provided as an image file that can be downloaded or edited according to user settings.

Below, we will briefly explain the operational flow of this application based on the content explained in detail above.

Figure 7 is an operational flowchart for a user-configured plant predictive maintenance method according to one embodiment of the present application.

The user-configured plant predictive maintenance method shown in FIG. 7 can be performed by the user-configured plant predictive maintenance device 100 described above. Therefore, even if the content is omitted below, the content described about the user-configured plant predictive maintenance device 100 can be similarly applied to the description of the user-configured plant predictive maintenance method.

Referring to FIG. 7, in step S210, the collection unit 110 can collect multiple equipment data for the plant 20 from multiple equipment sensors.

Next, in step S220, the storage unit 120 may store the equipment data collected by the collection unit 110 in real time or periodically according to a preset period.

Next, in step S230, the pre-processing unit 130 can load at least a portion of the equipment data stored in the storage unit 120 and perform pre-processing. Specifically, the pre-processing unit 130 includes an embedded pre-processing unit 131 and a user pre-processing unit 132. In step S231, the embedded pre-processing unit 131 can perform pre-processing on the equipment data using an internal pre-processing function provided in the predictive maintenance device 100, and in step S232, the user pre-processing unit 132 can perform pre-processing on the equipment data based on an external pre-processing file imported according to user settings.

Next, in step S240, the learning unit 140 can train the internal artificial intelligence model by inputting multiple pieces of equipment data preprocessed by the preprocessing unit 130.

Next, in step S250, the management unit 150 can manage model data for the internal AI model learned by the learning unit 140 or the external AI model imported according to user settings. Specifically, in step S250, the management unit 150 can store and provide model data for at least one of the model name, algorithm used, characteristics, verification method, hyper parameter value, applied preprocessing method, and whether it is an internal/external model for the internal AI model or the external AI model. In addition, the management unit 150 includes an internal model management unit 151 and an external model management unit 152. In step S251, the internal model management unit 151 manages internal model data for the internal AI model generated and learned by the learning unit 140, and in step S252, the external model management unit 152 can import an external AI model existing outside the predictive maintenance device 100 in file form, standardize it into a usable form, upload it, and manage the external model data for the imported and uploaded external AI model.

Next, in step S260, the operation unit 160 can calculate prediction data for the maintenance of the plant 20 using at least one of the artificial intelligence models stored and managed in the management unit 150 and at least one of the plurality of equipment data collected by the collection unit 110. Specifically, in step S260, the operation unit 160 can input at least one of the plurality of equipment data stored in the storage unit 120 to at least one of the artificial intelligence models stored in the management unit 150 to calculate prediction data. At this time, the prediction data can include at least one of anomaly detection for the plant 20, prediction of equipment failure life, prediction of failure probability, and prediction of maintenance timing.

Next, in step S270, the visualization unit 170 can visualize the prediction data calculated by the operation unit 160 in a manner that allows monitoring.

In the above description, steps S210 to S270 may be further divided into additional steps or combined into fewer steps depending on the embodiment of the present application. Also, some steps may be omitted as necessary, and the sequence between steps may be changed.

The user-configured plant predictive maintenance method according to an embodiment of the present application may be embodied in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, either alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the present invention, or may be known and usable by those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, magnetic-optical media such as floptical disks, and hardware devices specially configured to store and execute program instructions, such as ROMs, RAMs, and flash memories. Examples of program instructions include machine code, such as produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

The above-mentioned user-configured plant predictive maintenance method can also be embodied in the form of a computer program or application executed by a computer stored on a recording medium.

The above description of the present application is for illustrative purposes only, and a person having ordinary skill in the art to which the present application pertains will understand that the present application may be easily modified into other specific forms without changing the technical ideas or essential features of the present application. Therefore, it should be understood that the above described embodiments are illustrative in all respects and are not limiting. For example, each component described as being single may be implemented in a distributed manner, and similarly, each component described as being distributed may be implemented in a combined form.

The scope of this application is defined by the claims set forth below rather than by the above detailed description, and all modifications and variations derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of this application.

10: User-configured plant predictive maintenance system 20: Plant 30: User terminal 100: User-configured plant predictive maintenance device 110: Collection unit 120: Storage unit 130: Pre-processing unit 131: Built-in pre-processing unit 132: User pre-processing unit 140: Learning unit 150: Management unit 151: Internal model management unit 152: External model management unit 160: Operation unit 170: Visualization unit

Claims (10)

In a user-configured plant predictive maintenance device,
A collection unit that collects a plurality of equipment data for the plant from a plurality of equipment sensors;
a storage unit that stores the plurality of equipment data collected by the collection unit in real time or periodically according to a preset cycle;
A pre-processing unit that performs pre-processing on the plurality of equipment data based on an external pre-processing file imported according to a user setting;
a learning unit that learns an internal artificial intelligence model by inputting a plurality of equipment data preprocessed by the preprocessing unit;
a management unit for managing model data for the internal AI model trained by the learning unit or an external AI model imported according to a user setting;
an operation unit that calculates prediction data for maintenance of the plant using at least one of the artificial intelligence models stored in the management unit and at least one of the multiple equipment data collected by the collection unit;
Including,
The pre-treatment unit includes:
an embedded pre-processing unit that is provided within the plant predictive maintenance device and that performs pre-processing on the plurality of equipment data using an internal pre-processing function embedded in the plant predictive maintenance device;
a user pre-processing unit that imports the external pre-processing file existing outside the plant predictive maintenance device, and standardizes and uploads the external pre-processing file to perform pre-processing on the plurality of equipment data according to a rule set by the user using the external pre-processing file;
Including,
the plant predictive maintenance device preprocesses the plurality of equipment data by selectively applying one of the built-in preprocessing unit using the internal preprocessing function or the user preprocessing unit using the external preprocessing file;
The management unit imports an external artificial intelligence model existing outside the plant predictive maintenance device in a file format, formats it into a usable format, and uploads it;
The operation unit inputs at least one of the plurality of equipment data stored in the storage unit into at least one artificial intelligence model among the plurality of internal artificial intelligence models and the plurality of external artificial intelligence models stored in the management unit according to a specific problem that may occur in the plant and a user setting, thereby calculating the prediction data.
Plant predictive maintenance equipment. The pre-treatment unit includes:
At least a part of the equipment data stored in the storage unit is read and pre-processed.
The plant predictive maintenance system according to claim 1 . The management unit
The model data for at least one of the model name, algorithm used, characteristics, verification method, hyper parameter value, applied pre-processing method, and whether the model is internal/external model for the internal AI model or the external AI model is stored and provided;
The plant predictive maintenance system according to claim 1 . The prediction data is
The method includes at least one of detecting an abnormality in the plant, predicting equipment failure life, predicting failure probability, and predicting maintenance timing.
The plant predictive maintenance system according to claim 1 . Further comprising a visualization unit for visualizing the prediction data calculated by the management unit so that the data can be monitored;
The plant predictive maintenance system according to claim 1 . 1. A user-configured predictive plant maintenance method comprising:
collecting a plurality of equipment data for a plant from a plurality of equipment sensors;
storing the plurality of equipment data collected in real time or periodically according to a preset period;
performing pre-processing on the plurality of equipment data based on an external pre-processing file imported according to a user setting;
training an internal artificial intelligence model by inputting the plurality of equipment data preprocessed in the preprocessing step;
managing model data for the internal AI model trained in the training step or an external AI model imported according to a user setting;
calculating prediction data for the maintenance of the plant using at least one of the artificial intelligence models stored and managed in the managing model data step and at least one of the plurality of equipment data collected in the collecting plurality of equipment data step;
Including,
The pretreatment step includes:
(a) providing within a plant predictive maintenance system and performing pre-processing on the plurality of equipment data using an internal pre-processing function incorporated in the plant predictive maintenance system;
(b) importing the external pre-processing file existing outside the plant predictive maintenance device, and standardizing and uploading the external pre-processing file to perform pre-processing on the plurality of equipment data according to a rule set by the user using the external pre-processing file ;
Including,
In the plant predictive maintenance method, the plurality of equipment data are pre- processed by selectively applying one of the steps (a) and (b);
The step of managing the model data includes importing an external artificial intelligence model existing outside the plant predictive maintenance device in a file format, formatting the model into a usable format, and uploading the formatted model;
The step of calculating predictive data for the maintenance of the plant comprises inputting at least one of the plurality of equipment data stored in the storing step into at least one of the plurality of internal artificial intelligence models and the plurality of external artificial intelligence models stored in the model data managing step according to specific problems that may occur in the plant and user settings, thereby calculating the predictive data . The pretreatment step includes:
The step of storing includes retrieving at least a portion of the equipment data stored in the step of storing and performing pre-processing.
The method of predictive maintenance of a plant according to claim 6 . The step of managing the model data includes:
Storing and providing the model data for at least one of the model name, algorithm used, characteristics, verification method, hyper parameter value, applied pre-processing method, and whether the model is internal/external model for the internal AI model or the external AI model;
The method of predictive maintenance of a plant according to claim 6 . The prediction data is
The method includes at least one of detecting an abnormality in the plant, predicting equipment failure life, predicting failure probability, and predicting maintenance timing.
The method of predictive maintenance of a plant according to claim 6 . A step of visualizing the predicted data calculated in the step of calculating the predicted data for the maintenance of the plant so as to be monitorable;
The method of claim 6 , further comprising: