Intelligent power plant management method and system based on AI

Abstract

The invention provides an AI-based intelligent power plant management method and system, and the method comprises the steps: constructing a monitoring vector for each monitoring point based on monitoring data collected historically in an initialization process, inputting the monitoring vector into an AI edge model of the monitoring point, and outputting an abnormal value of the monitoring point; on the basis of abnormal values output by the AI edge models at all the time points, an abnormality statistical graph corresponding to each AI edge model is constructed; determining whether any two monitoring points have a synchronization relationship or not based on the position of each mark point in the anomaly statistical graph of the AI edge model and a connecting line between the adjacent mark points; in the real-time monitoring process, whether the monitoring point is in an abnormal state or not is judged based on abnormal values output by the AI edge models in real time, and if the monitoring point is in the abnormal state, a synchronous inspection task is constructed for the monitoring point and the monitoring points having the synchronous relation with the monitoring point based on the synchronous relation of the monitoring point.

Description

Technical Field

[0001] The present invention relates to the technical field of smart power plants, and in particular to an AI-based smart power plant management method and system. Background Art

[0002] The use of AI technology to regulate power plants has brought revolutionary changes to the power industry, bringing many benefits, covering multiple dimensions such as efficiency improvement, safety assurance, cost control, environmental protection, and management optimization.

[0003] First, the application of AI technology in power plant supervision has significantly improved operational efficiency. By real-time monitoring and analyzing the operating data of various equipment in the power plant, AI can quickly identify abnormal conditions and predict equipment failures, so as to take corresponding measures in advance to avoid production interruptions caused by equipment failures. This real-time and accurate data analysis capability enables power plants to fully control the status of equipment, ensure that the equipment is always in the best operating state, and effectively improve the overall operational efficiency. Secondly, AI technology provides strong support for the safety of power plants. Traditional manual supervision methods often find it difficult to achieve all-weather, no-dead-angle safety monitoring, while AI technology can achieve real-time monitoring of every corner of the power plant, and promptly discover and warn of potential safety hazards. In addition, AI can also mine and analyze the historical data of power plants through deep learning algorithms, predict possible safety risks, provide a scientific basis for the safety management of power plants, and greatly reduce the incidence of safety accidents.

[0004] However, the existing technology often sets an AI model for a monitoring end and implements supervision of a single monitoring end through the AI ​​model. However, the various devices in a power plant often have a relationship of mutual influence. The existing technology often simply implements the supervision of a monitoring end through a model, ignoring the relationship between the monitoring ends, and the supervision effect is poor.

[0005] In view of this, the present invention is proposed. Summary of the invention

[0006] The purpose of the present invention is to provide an AI-based smart power plant management method and system. Based on the relationship between monitoring points, this solution realizes synchronous supervision of multiple monitoring points through the AI ​​edge model of one monitoring point, thereby improving the supervision effect.

[0007] The present invention provides an AI-based smart power plant management method. A central control terminal and multiple monitoring terminals are provided in the power plant, and each monitoring terminal corresponds to a monitoring point of a power plant. The steps of the method include:

[0008] During the initialization process, the monitoring point of each monitoring point constructs a monitoring vector based on the historically collected monitoring data, inputs the monitoring vector into the AI ​​edge model of the monitoring point, and the AI ​​edge model outputs the abnormal value of the monitoring point;

[0009] Based on the abnormal values ​​output by the AI ​​edge model at each time point, construct an abnormality statistical graph corresponding to each AI edge model, wherein the abnormality statistical graph is a line graph;

[0010] Based on the position of each mark point in the anomaly statistics graph of the AI ​​edge model and the connection between adjacent mark points, determine whether there is a synchronization relationship between any two monitoring points;

[0011] During the real-time monitoring process, the abnormal values ​​output in real time by each AI edge model are used to determine whether the monitoring point is in an abnormal state. If it is in an abnormal state, a synchronous inspection task is constructed for the monitoring point and the monitoring points that have a synchronous relationship with the monitoring point based on the synchronization relationship of the monitoring point.

[0012] Using the above scheme, during the daily operation of the smart power plant of this scheme, first, based on the monitoring data collected by each AI edge model in the historical process, the abnormal value corresponding to each monitoring point in each historical process at each time point is calculated, and an abnormality statistical graph of each monitoring point is constructed. Since there may be mutual influence between the components in the power plant, and if there is mutual influence, the abnormality will increase successively, and then there are often similarities in the abnormality changes of the two. This scheme determines the monitoring points with a synchronization relationship by comparing the abnormality statistical graphs. Then, in the real-time monitoring process, if one of the monitoring points with a synchronization relationship has an abnormality, a synchronous inspection task is issued to perform synchronous inspections on multiple monitoring points. Based on the relationship between the monitoring points, the AI ​​edge model of one monitoring point is used to realize synchronous supervision of multiple monitoring points, thereby improving the supervision effect.

[0013] In some embodiments of the present invention, in the step of constructing a monitoring vector at the monitoring point based on historically collected monitoring data, inputting the monitoring vector into an AI edge model at the monitoring point, and the AI ​​edge model outputting an abnormal value of the monitoring point:

[0014] The monitoring data of each time point collected historically is constructed into a monitoring vector corresponding to a time point;

[0015] The monitoring vector of a monitoring point at each time point is input into the pre-trained AI edge model corresponding to the monitoring point, and the AI ​​edge model outputs an abnormal value of the monitoring point at each time point.

[0016] The above scheme is adopted. During the initialization process, data statistics are performed on a longer historical period. Since the mutual influence between components in a power plant is often not absolutely synchronized, the scheme ensures the accuracy of the synchronization relationship through statistical calculation of data over a longer period of time.

[0017] In some embodiments of the present invention, in the step of constructing an abnormality statistical graph corresponding to each AI edge model based on the abnormal values ​​output by the AI ​​edge model at each time point:

[0018] Construct a plane rectangular coordinate system, and use the horizontal axis of the plane rectangular coordinate system as the axis of the time point, and the vertical axis as the axis of the outlier value;

[0019] Based on the abnormal values ​​of the AI ​​edge model at each time point, a plurality of marking points are constructed in the plane rectangular coordinate system, and the marking points are connected in chronological order to obtain the abnormality statistical graph.

[0020] In some embodiments of the present invention, in the step of determining whether there is a synchronization relationship between any two monitoring points based on the position of each marked point and the connection line between adjacent marked points in the abnormality statistical graph of the AI ​​edge model:

[0021] Obtain anomaly statistics of the current monitoring point and the monitoring points that have a synchronization relationship with the current monitoring point;

[0022] Based on the preset comparison time window, the corresponding broken line segments of the comparison time window length are respectively intercepted from the two abnormality statistical graphs;

[0023] Calculate the numerical difference based on the abnormality of the corresponding marked points in the two polyline segments;

[0024] Calculate the slope difference based on the line between the corresponding marked points in the two polyline segments;

[0025] A synchronization relationship value is calculated based on the numerical difference and the slope difference, and whether there is a synchronization relationship between two monitoring points is determined based on the synchronization relationship value.

[0026] Using the above scheme, if there is a mutual influence between the components in the power plant, the abnormality will increase successively, and then there are often similarities in the abnormality changes of the two. This scheme first intercepts from the two abnormality statistical graphs through a preset comparison time window, and calculates the numerical difference and slope difference of the intercepted partial images respectively. The numerical difference reflects the intuitive abnormality difference, and the slope difference reflects the difference in the abnormality changes of the two. The synchronization relationship value is calculated together by the two differences to ensure the accuracy of the determination of the synchronization relationship.

[0027] In some embodiments of the present invention, in the step of calculating the numerical difference based on the abnormality of the corresponding marking points in the two broken line segments, the numerical difference is calculated based on the following formula:

[0028]

[0029] Among them, α1 represents the value of the numerical difference, N represents the number of marked points in the polyline segment, and x i Indicates the abnormality value of the i-th marked point in a polyline segment. Indicates the value of the abnormality of the i-th marked point in another polyline segment;

[0030] In the step of calculating the slope difference based on the connecting line between the corresponding marked points in the two polyline segments, the slope difference is calculated based on the following formula:

[0031]

[0032] Among them, α2 represents the value of the slope difference, N-1 represents the number of lines in the broken line segment, and y i Represents the value of the slope of the i-th line in a polyline segment. Represents the value of the slope of the i-th line in another polyline segment.

[0033] In some embodiments of the present invention, in the step of intercepting corresponding broken line segments of the comparison time window length from two abnormality statistical graphs based on a preset comparison time window, a first broken line segment of the comparison time window length before the current time point is intercepted from the abnormality statistical graph of the current monitoring point through the comparison time window; from the abnormality statistical graph of the monitoring point that determines the synchronization relationship with the current monitoring point through the comparison time window, in a reverse stepping manner, starting from the monitoring point at the current time point, one time point is stepped each time, and a second broken line segment of the comparison time window length is intercepted each time, and the first broken line segment and each second broken line segment are constructed as a comparison group;

[0034] In the steps of calculating the synchronization relationship value based on the numerical difference and the slope difference, and determining whether there is a synchronization relationship between two monitoring points based on the synchronization relationship value, the synchronization relationship values ​​of the first broken line segment and the second broken line segment in each comparison group are calculated respectively, and whether there is a synchronization relationship between the two monitoring points is determined based on the synchronization relationship values ​​of multiple comparison groups.

[0035] Using the above scheme, if there is mutual influence between the components in the power plant, the abnormality will increase successively, which will be reflected in the abnormality statistical graphs of the two monitoring points. The changes of the two are prone to advance or lag. This scheme compares the first broken line segment of the comparison time window length before the current time point with multiple second broken line graphs of another monitoring point, which can improve the perception sensitivity of the synchronization relationship between the two monitoring points.

[0036] In some embodiments of the present invention, in the step of respectively calculating the synchronization relationship values ​​of the first broken line segment and the second broken line segment in each comparison group, and determining whether two monitoring points have a synchronization relationship based on the synchronization relationship values ​​of multiple comparison groups, the synchronization relationship value of each comparison group is compared with a preset synchronization relationship threshold value. If the synchronization relationship value of any comparison group is greater than the preset synchronization relationship threshold value, it is determined whether the two monitoring points have a synchronization relationship.

[0037] Using the above scheme, according to the principle of the abnormality statistical chart of monitoring points that affect each other, this scheme constructs multiple comparison groups. If one of the comparison groups meets the judgment of the synchronization relationship, it means that there is a synchronization relationship between the two monitoring points, ensuring the accuracy of the synchronization relationship confirmation.

[0038] In some embodiments of the present invention, during the initialization process, based on the synchronization relationship between the monitoring points, each monitoring point is taken as a node, and edges are constructed between the nodes corresponding to the monitoring points having the synchronization relationship to obtain a synchronization relationship graph;

[0039] In the step of determining whether the monitoring point is in an abnormal state based on the abnormal values ​​output in real time by each AI edge model, if it is in an abnormal state, then based on the synchronization relationship of the monitoring point, constructing a synchronous inspection task for the monitoring point and the monitoring points that have a synchronous relationship with the monitoring point, locating the node corresponding to the monitoring point in the synchronization relationship graph, obtaining the nodes within a preset distance range of the node from the synchronization relationship graph, and constructing a synchronous inspection task for the monitoring points corresponding to the nodes within the preset distance range of the node and the monitoring points corresponding to the node.

[0040] Adopting the above scheme, this scheme constructs a synchronization relationship diagram based on the synchronization relationship between each monitoring point to represent the overall relationship between each monitoring point in the power plant. In the process of determining the synchronous inspection task, the preset distance can be directly set to one hop or multiple hops to determine the scope of the synchronous inspection task. By constructing a synchronization relationship diagram including the relationship of all monitoring points in the power plant, the efficiency of the construction process of the synchronous inspection task is guaranteed.

[0041] Another aspect of the present invention also relates to an AI-based smart power plant management system, which includes a computer device, wherein the computer device includes a processor and a memory, wherein the memory stores computer instructions, and the processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps implemented by the method.

[0042] In summary, the present invention has the following beneficial effects:

[0043] 1. In the daily operation of the smart power plant of this solution, based on the monitoring data collected by each AI edge model in the historical process, the abnormal value corresponding to each monitoring point at each time point in each historical process is calculated, and the abnormality statistical graph of each monitoring point is constructed. Since there may be mutual influence between the components in the power plant, and if there is mutual influence, the abnormality will increase successively, and then there is often similarity in the abnormality changes of the two. This solution determines the monitoring points with synchronization relationship by comparing the abnormality statistical graph. Then, in the real-time monitoring process, if one of the monitoring points with synchronization relationship has an abnormality, a synchronous inspection task is issued to perform synchronous inspections on multiple monitoring points. Based on the relationship between the monitoring points, the AI ​​edge model of one monitoring point is used to realize synchronous supervision of multiple monitoring points, thereby improving the supervision effect;

[0044] 2. This scheme performs data statistics for a longer historical period during the initialization process. Since the mutual influence between components in a power plant is often not absolutely synchronized, this scheme ensures the accuracy of the synchronization relationship through data statistics calculation over a longer period of time;

[0045] 3. If there is mutual influence between the components in the power plant, the abnormality will increase successively, and the abnormality changes of the two often have similarities. This scheme first intercepts from the two abnormality statistical graphs through a preset comparison time window, and calculates the numerical difference and slope difference of the intercepted partial images respectively. The numerical difference reflects the intuitive abnormality difference, and the slope difference reflects the difference in the abnormality changes of the two. The synchronization relationship value is calculated by the two differences to ensure the accuracy of the determination of the synchronization relationship;

[0046] 4. This solution constructs a synchronization relationship diagram based on the synchronization relationship between each monitoring point to represent the overall relationship between each monitoring point in the power plant. In the process of determining the synchronization inspection task, the preset distance can be directly set to one hop or multiple hops to determine the scope of the synchronization inspection task. By constructing a synchronization relationship diagram including the relationship between all monitoring points in the power plant, the efficiency of the synchronization inspection task construction process is guaranteed. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0048] Figure 1 A schematic diagram of an implementation of an AI-based smart power plant management method of the present invention;

[0049] Figure 2 A schematic diagram of another implementation of the AI-based smart power plant management method of the present invention;

[0050] Figure 3 A schematic diagram of an implementation method of building a synchronous inspection task in the AI-based smart power plant management method of the present invention. DETAILED DESCRIPTION

[0051] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of systems and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0052] The terms used in the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The singular forms "a", "the" and "the" used in the present invention and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0053] like Figure 1 As shown, the present invention provides an AI-based smart power plant management method, in which a central control terminal and multiple monitoring terminals are provided in the power plant, each monitoring terminal corresponds to a monitoring point of a power plant, and the steps of the method include:

[0054] Step S100, during the initialization process, the monitoring point of each monitoring point constructs a monitoring vector based on historically collected monitoring data, inputs the monitoring vector into the AI ​​edge model of the monitoring point, and the AI ​​edge model outputs an abnormal value of the monitoring point;

[0055] In some embodiments of the present invention, the AI ​​edge models of the monitoring points are all pre-trained machine learning models, and the machine learning models can be convolutional neural network models, deep learning models, or recurrent neural network models, etc.

[0056] During the pre-training of the machine learning model, the mean square error, cross entropy, or SmoothL1 loss function can be used to calculate the loss to complete the pre-training of each AI edge model.

[0057] During the specific implementation process, each AI edge model can adopt a different model structure and can also be trained with different training data sets.

[0058] Step S200, based on the abnormal values ​​output by the AI ​​edge model at each time point, construct an abnormality statistical graph corresponding to each AI edge model, wherein the abnormality statistical graph is a line graph;

[0059] Step S300, based on the position of each marked point in the abnormality statistical graph of the AI ​​edge model and the connection line between adjacent marked points, determine whether there is a synchronization relationship between any two monitoring points;

[0060] In a specific implementation process, each marked point in the abnormality statistical graph corresponds to a time point, and the total vertical coordinate of the marked point corresponds to the value of the abnormality.

[0061] Step S400, during the real-time monitoring process, it is determined whether the monitoring point is in an abnormal state based on the abnormal values ​​output in real time by each AI edge model. If it is in an abnormal state, based on the synchronization relationship of the monitoring point, a synchronous inspection task is constructed for the monitoring point and the monitoring points that have a synchronous relationship with the monitoring point.

[0062] During the specific implementation process, in the step of determining whether the monitoring point is in an abnormal state based on the abnormal value output in real time by each AI edge model, the abnormal value output in real time by the AI ​​edge model is compared with a preset abnormality threshold. If it is greater than, the monitoring point is determined to be in an abnormal state.

[0063] During the specific implementation process, after the synchronous inspection task is constructed, the location of each monitoring point in the synchronous inspection task is fed back to the staff so that the staff can carry out inspection work on each monitoring point.

[0064] Using the above scheme, during the daily operation of the smart power plant of this scheme, first, based on the monitoring data collected by each AI edge model in the historical process, the abnormal value corresponding to each monitoring point in each historical process at each time point is calculated, and an abnormality statistical graph of each monitoring point is constructed. Since there may be mutual influence between the components in the power plant, and if there is mutual influence, the abnormality will increase successively, and then there are often similarities in the abnormality changes of the two. This scheme determines the monitoring points with a synchronization relationship by comparing the abnormality statistical graphs. Then, in the real-time monitoring process, if one of the monitoring points with a synchronization relationship has an abnormality, a synchronous inspection task is issued to perform synchronous inspections on multiple monitoring points. Based on the relationship between the monitoring points, the AI ​​edge model of one monitoring point is used to realize synchronous supervision of multiple monitoring points, thereby improving the supervision effect.

[0065] like Figure 2 As shown, in some embodiments of the present invention, in the step of constructing a monitoring vector at the monitoring point based on historically collected monitoring data, inputting the monitoring vector into the AI ​​edge model of the monitoring point, and the AI ​​edge model outputting an abnormal value of the monitoring point:

[0066] Step S110, constructing the monitoring data of each time point collected historically into a monitoring vector corresponding to one time point;

[0067] During the specific implementation process, the monitoring data collected by the monitoring point at each time point may be image data collected by an image acquisition device, temperature data collected by a temperature sensor, and data on humidity, sound intensity, sound frequency, and harmful gas concentration. Specifically, the value of each data is normalized, and the normalized value is used as the value of each dimension in the monitoring vector.

[0068] Step S120, inputting the monitoring vector of a monitoring point at each time point into the pre-trained AI edge model corresponding to the monitoring point, and the AI ​​edge model outputs an abnormal value of the monitoring point at each time point.

[0069] The above scheme is adopted. During the initialization process, data statistics are performed on a longer historical period. Since the mutual influence between components in a power plant is often not absolutely synchronized, the scheme ensures the accuracy of the synchronization relationship through statistical calculation of data over a longer period of time.

[0070] In some embodiments of the present invention, in the step of constructing an abnormality statistical graph corresponding to each AI edge model based on the abnormal values ​​output by the AI ​​edge model at each time point:

[0071] Step S210, constructing a plane rectangular coordinate system, using the horizontal axis of the plane rectangular coordinate system as the axis of the time point, and the vertical axis as the axis of the outlier;

[0072] Step S220, based on the abnormal values ​​of the AI ​​edge model at each time point, construct a plurality of marking points in the plane rectangular coordinate system, connect the marking points in chronological order, and obtain the abnormality statistical graph.

[0073] like Figure 2 and 3 As shown, in some embodiments of the present invention, in the step of determining whether there is a synchronization relationship between any two monitoring points based on the position of each marked point and the connection between adjacent marked points in the abnormality statistical graph of the AI ​​edge model:

[0074] Step 310, obtaining an abnormality statistical graph of the current monitoring point and the monitoring points that have a synchronization relationship with the current monitoring point;

[0075] Step 320, based on the preset comparison time window, respectively intercepting a corresponding line segment of the comparison time window length from the two abnormality statistical graphs;

[0076] Step 330, calculating the numerical difference based on the abnormality of the corresponding marking points in the two broken line segments;

[0077] Step 340, calculating the slope difference based on the line connecting the corresponding marked points in the two broken line segments;

[0078] In a specific implementation process, the marking points and connecting lines in the two broken line segments are numbered sequentially based on the time sequence, and the marking points or connecting lines with the same number are regarded as corresponding marking points or connecting lines.

[0079] Step 350: Calculate a synchronization relationship value based on the numerical difference and the slope difference, and determine whether there is a synchronization relationship between two monitoring points based on the synchronization relationship value.

[0080] In a specific implementation process, in the process of calculating the synchronization relationship value, the sum of the numerical difference and the slope difference is calculated, and the reciprocal of the sum is calculated as the synchronization relationship value.

[0081] In some embodiments of the present invention, in the step of determining whether there is a synchronization relationship between two monitoring points based on the synchronization relationship value, the calculated synchronization relationship value is compared with a preset synchronization relationship threshold. If the synchronization relationship value is greater than the preset synchronization relationship threshold, it is determined that there is a synchronization relationship between the two monitoring points.

[0082] Using the above scheme, if there is a mutual influence between the components in the power plant, the abnormality will increase successively, and then there are often similarities in the abnormality changes of the two. This scheme first intercepts from the two abnormality statistical graphs through a preset comparison time window, and calculates the numerical difference and slope difference of the intercepted partial images respectively. The numerical difference reflects the intuitive abnormality difference, and the slope difference reflects the difference in the abnormality changes of the two. The synchronization relationship value is calculated together by the two differences to ensure the accuracy of the determination of the synchronization relationship.

[0083] In some embodiments of the present invention, in the step of calculating the numerical difference based on the abnormality of the corresponding marking points in the two broken line segments, the numerical difference is calculated based on the following formula:

[0084]

[0085] Among them, α1 represents the value of the numerical difference, N represents the number of marked points in the polyline segment, and x i Indicates the abnormality value of the i-th marked point in a polyline segment. Indicates the value of the abnormality of the i-th marked point in another polyline segment;

[0086] In the step of calculating the slope difference based on the connecting line between the corresponding marked points in the two polyline segments, the slope difference is calculated based on the following formula:

[0087]

[0088] Among them, α2 represents the value of the slope difference, N-1 represents the number of lines in the broken line segment, and y i Represents the value of the slope of the i-th line in a polyline segment. Represents the value of the slope of the i-th line in another polyline segment.

[0089] As Figure 3 As shown, in some embodiments of the present invention, in the step of intercepting corresponding broken line segments of the comparison time window length from two abnormality statistical graphs based on a preset comparison time window, the step includes step S321, intercepting a first broken line segment of the comparison time window length before the current time point from the abnormality statistical graph of the current monitoring point through the comparison time window; from the abnormality statistical graph of the monitoring point that determines the synchronization relationship with the current monitoring point through the comparison time window, in a reverse stepping manner, starting from the monitoring point at the current time point, stepping one time point each time, intercepting a second broken line segment of the comparison time window length each time, and constructing the first broken line segment and each second broken line segment into a comparison group;

[0090] In the steps of calculating the synchronization relationship value based on the numerical difference and the slope difference, and determining whether there is a synchronization relationship between two monitoring points based on the synchronization relationship value, step S351 is included, respectively calculating the synchronization relationship value of the first broken line segment and the second broken line segment in each comparison group, and determining whether there is a synchronization relationship between two monitoring points based on the synchronization relationship values ​​of multiple comparison groups.

[0091] Using the above scheme, if there is mutual influence between the components in the power plant, the abnormality will increase successively, which will be reflected in the abnormality statistical graphs of the two monitoring points. The changes of the two are prone to advance or lag. This scheme compares the first broken line segment of the comparison time window length before the current time point with multiple second broken line graphs of another monitoring point, which can improve the perception sensitivity of the synchronization relationship between the two monitoring points.

[0092] In some embodiments of the present invention, in the step of respectively calculating the synchronization relationship values ​​of the first broken line segment and the second broken line segment in each comparison group, and determining whether two monitoring points have a synchronization relationship based on the synchronization relationship values ​​of multiple comparison groups, the synchronization relationship value of each comparison group is compared with a preset synchronization relationship threshold value. If the synchronization relationship value of any comparison group is greater than the preset synchronization relationship threshold value, it is determined whether the two monitoring points have a synchronization relationship.

[0093] Using the above scheme, according to the principle of the abnormality statistical chart of monitoring points that affect each other, this scheme constructs multiple comparison groups. If one of the comparison groups meets the judgment of the synchronization relationship, it means that there is a synchronization relationship between the two monitoring points, ensuring the accuracy of the synchronization relationship confirmation.

[0094] like Figure 3 As shown, in some embodiments of the present invention, the initialization process further includes step S360, based on the synchronization relationship between the monitoring points, each monitoring point is taken as a node, and edges are constructed between the nodes corresponding to the monitoring points having the synchronization relationship to obtain a synchronization relationship graph;

[0095] In the step of determining whether the monitoring point is in an abnormal state based on the abnormal values ​​output in real time by each AI edge model, if it is in an abnormal state, based on the synchronization relationship of the monitoring point, constructing a synchronous inspection task for the monitoring point and the monitoring points that have a synchronous relationship with the monitoring point, the step includes step S410, locating the node corresponding to the monitoring point in the synchronization relationship graph, obtaining the node within a preset distance range of the node from the synchronization relationship graph, and constructing a synchronous inspection task for the monitoring point corresponding to the node within the preset distance range of the node and the monitoring point corresponding to the node.

[0096] In a specific implementation process, the preset distance range may be a one-hop distance, a two-hop distance, a three-hop distance, or the like.

[0097] Adopting the above scheme, this scheme constructs a synchronization relationship diagram based on the synchronization relationship between each monitoring point to represent the overall relationship between each monitoring point in the power plant. In the process of determining the synchronous inspection task, the preset distance can be directly set to one hop or multiple hops to determine the scope of the synchronous inspection task. By constructing a synchronization relationship diagram including the relationship of all monitoring points in the power plant, the efficiency of the construction process of the synchronous inspection task is guaranteed.

[0098] In some implementations of the present invention, the method further comprises updating the synchronization relationship diagram at every preset time slice, and during the updating process, collecting monitoring data of a past time slice for initialization.

[0099] In a specific implementation process, the synchronization relationship diagram is updated at the control point.

[0100] Another aspect of the present invention also relates to an AI-based smart power plant management system, which includes a computer device, wherein the computer device includes a processor and a memory, wherein the memory stores computer instructions, and the processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps implemented by the method.

[0101] The embodiment of the present invention also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the aforementioned AI-based smart power plant management method is implemented. The computer-readable storage medium can be a tangible storage medium, such as a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a floppy disk, a hard disk, a removable storage disk, a CD-ROM, or any other form of storage medium known in the technical field.

[0102] It should be understood by those skilled in the art that the exemplary components, systems and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software or a combination of the two. Whether it is performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present invention are programs or code segments used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link via a data signal carried in a carrier.

[0103] It should be clear that the present invention is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present invention.

[0104] In the present invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with or replace features of other embodiments.

[0105] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the embodiments of the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A smart power plant management method based on AI, characterized in that: A central control terminal and multiple monitoring terminals are provided in the power plant, each monitoring terminal corresponds to a monitoring point of a power plant, and the steps of the method include: During the initialization process, the monitoring point of each monitoring point constructs a monitoring vector based on the historically collected monitoring data, inputs the monitoring vector into the AI ​​edge model of the monitoring point, and the AI ​​edge model outputs the abnormal value of the monitoring point; Based on the abnormal values ​​output by the AI ​​edge model at each time point, construct an abnormality statistical graph corresponding to each AI edge model, wherein the abnormality statistical graph is a line graph; Based on the position of each mark point in the anomaly statistics graph of the AI ​​edge model and the connection between adjacent mark points, determine whether there is a synchronization relationship between any two monitoring points; During the real-time monitoring process, the abnormal values ​​output in real time by each AI edge model are used to determine whether the monitoring point is in an abnormal state. If it is in an abnormal state, a synchronous inspection task is constructed for the monitoring point and the monitoring points that have a synchronous relationship with the monitoring point based on the synchronization relationship of the monitoring point.

2. The AI-based smart power plant management method according to claim 1 is characterized in that: In the step of constructing a monitoring vector at the monitoring point based on historically collected monitoring data, inputting the monitoring vector into an AI edge model of the monitoring point, and the AI ​​edge model outputting an abnormal value of the monitoring point: The monitoring data of each time point collected historically is constructed into a monitoring vector corresponding to a time point; The monitoring vector of a monitoring point at each time point is input into the pre-trained AI edge model corresponding to the monitoring point, and the AI ​​edge model outputs an abnormal value of the monitoring point at each time point.

3. The AI-based smart power plant management method according to claim 1 is characterized in that: In the step of constructing an abnormality statistical graph corresponding to each AI edge model based on the abnormal values ​​output by the AI ​​edge model at each time point: Construct a plane rectangular coordinate system, and use the horizontal axis of the plane rectangular coordinate system as the axis of the time point, and the vertical axis as the axis of the outlier value; Based on the abnormal values ​​of the AI ​​edge model at each time point, a plurality of marking points are constructed in the plane rectangular coordinate system, and the marking points are connected in chronological order to obtain the abnormality statistical graph.

4. The AI-based smart power plant management method according to any one of claims 1 to 3, characterized in that: In the step of determining whether there is a synchronization relationship between any two monitoring points based on the position of each marked point and the connection between adjacent marked points in the abnormality statistics graph of the AI ​​edge model: Obtain anomaly statistics of the current monitoring point and the monitoring points that have a synchronization relationship with the current monitoring point; Based on the preset comparison time window, the corresponding broken line segments of the comparison time window length are respectively intercepted from the two abnormality statistical graphs; Calculate the numerical difference based on the abnormality of the corresponding marked points in the two polyline segments; Calculate the slope difference based on the line between the corresponding marked points in the two polyline segments; A synchronization relationship value is calculated based on the numerical difference and the slope difference, and whether there is a synchronization relationship between two monitoring points is determined based on the synchronization relationship value.

5. The AI-based smart power plant management method according to claim 4 is characterized in that: In the step of calculating the numerical difference based on the abnormality of the corresponding marked points in the two broken line segments, the numerical difference is calculated based on the following formula: Among them, α1 represents the value of the numerical difference, N represents the number of marked points in the polyline segment, and x i Indicates the abnormality value of the i-th marked point in a polyline segment. Indicates the value of the abnormality of the i-th marked point in another polyline segment; In the step of calculating the slope difference based on the connecting line between the corresponding marked points in the two polyline segments, the slope difference is calculated based on the following formula: Among them, α2 represents the value of the slope difference, N-1 represents the number of lines in the broken line segment, and y i Represents the value of the slope of the i-th line in a polyline segment. Represents the value of the slope of the i-th line in another polyline segment.

6. The AI-based smart power plant management method according to claim 4 is characterized in that: In the step of respectively extracting line segments of the length of the corresponding comparison time window from the two abnormality statistical graphs based on the preset comparison time window: Intercepting a first broken line segment of the length of the comparison time window before the current time point from the abnormality statistical graph of the current monitoring point through the comparison time window; Through the comparison time window, from the abnormality statistical chart of the monitoring point that determines the synchronization relationship with the current monitoring point, by reverse stepping, starting from the monitoring point of the current time point, stepping one time point each time, each time intercepting a second broken line segment of the length of the comparison time window, the first broken line segment and each second broken line segment are constructed as a comparison group.

7. The AI-based smart power plant management method according to claim 4 is characterized in that: In the steps of calculating the synchronization relationship value based on the numerical difference and the slope difference, and determining whether there is a synchronization relationship between two monitoring points based on the synchronization relationship value, the synchronization relationship values ​​of the first broken line segment and the second broken line segment in each comparison group are calculated respectively, and whether there is a synchronization relationship between the two monitoring points is determined based on the synchronization relationship values ​​of multiple comparison groups.

8. The AI-based smart power plant management method according to claim 7 is characterized in that: In the step of respectively calculating the synchronization relationship values ​​of the first broken line segment and the second broken line segment in each comparison group, and determining whether there is a synchronization relationship between two monitoring points based on the synchronization relationship values ​​of multiple comparison groups, the synchronization relationship value of each comparison group is compared with a preset synchronization relationship threshold. If the synchronization relationship value of any comparison group is greater than the preset synchronization relationship threshold, it is determined whether there is a synchronization relationship between the two monitoring points.

9. The AI-based smart power plant management method according to claim 1 or 8, characterized in that: During the initialization process, based on the synchronization relationship between monitoring points, each monitoring point is taken as a node, and edges are constructed between the nodes corresponding to the monitoring points with synchronization relationship to obtain a synchronization relationship graph; In the step of determining whether the monitoring point is in an abnormal state based on the abnormal values ​​output in real time by each AI edge model, if it is in an abnormal state, then based on the synchronization relationship of the monitoring point, constructing a synchronous inspection task for the monitoring point and the monitoring points that have a synchronous relationship with the monitoring point, locating the node corresponding to the monitoring point in the synchronization relationship graph, obtaining the nodes within a preset distance range of the node from the synchronization relationship graph, and constructing a synchronous inspection task for the monitoring points corresponding to the nodes within the preset distance range of the node and the monitoring points corresponding to the node.

10. An AI-based smart power plant management system, characterized by: The system includes a computer device, which includes a processor and a memory, wherein the memory stores computer instructions, and the processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps implemented by the method described in any one of claims 1 to 9.

Images