Transformer operation state logical reasoning chain construction method and system

Abstract

The invention discloses a transformer operation state logical reasoning chain construction method and system, and relates to the technical field of transformer state evaluation. The method comprises the following steps: constructing a running state quantity set, an abnormal state type set and a mutual mapping relationship between elements in the running state quantity set and elements in the abnormal state type set; main characteristic state quantities are randomly selected from the operation state quantity set, and an abnormal state type matrix corresponding to the main characteristic state quantities and an operation state quantity matrix corresponding to the abnormal state type matrix are obtained according to the mutual mapping relation; according to the main characteristic state quantity and the abnormal state type matrix, obtaining a backbone network of a transformer operation state logical reasoning chain; and obtaining a branch network of the transformer operation state logical reasoning chain according to the operation state quantity matrix, and connecting the branch network to the backbone network to obtain the transformer operation state logical reasoning chain. The method can overcome the problems of excessive dependence on feature quantity integrity, excessive attention to partial state quantity and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of transformer state assessment, and in particular to a method and system for constructing a transformer operating state logic reasoning chain. Background Art

[0002] Transformers are numerous in power systems, and most are critical to power transmission. Failure can cause significant economic losses. To monitor the real-time operating status of transformers and reduce losses from transformer failures, power operations and maintenance organizations often employ online monitoring, routine maintenance, and offline testing, combined with methods like three ratios and artificial intelligence to provide transformer fault warnings and diagnostics.

[0003] Among them, the three-ratio method based on dissolved gas in oil is relatively advanced, but its application requires high on-site experience of operation and maintenance personnel, which can easily lead to misjudgment or omission of faults; the artificial intelligence method has high requirements for the integrity of basic data, and the data integrity of on-site transformers varies greatly, making it difficult to give full play to the advantages of artificial intelligence methods. In response to the problem of incomplete basic data in transformer fault diagnosis, the existing methods can be roughly divided into the following types: (1) discarding samples with incomplete data and only considering samples with complete data as valid, which increases the difficulty of sample collection to a certain extent; (2) based on statistical methods, using zero value, mean or median to replace missing sample data, or expressing missing data in the form of probability, but due to the large dispersion of transformer data, it causes a certain degree of deviation in the diagnosis results; (3) not doing any data processing and considering samples with incomplete data as valid samples, which can also cause errors in the diagnosis results to a certain extent. Summary of the Invention

[0004] The purpose of the present invention is to overcome the deficiencies in the prior art and provide a method and system for constructing a logical reasoning chain for transformer operating status, which provides reliable results for equipment status assessment through state quantity drive and logical reasoning, and effectively overcomes problems such as over-reliance on the integrity of characteristic quantities, excessive focus on partial state quantities, and incomplete assessment and diagnosis results.

[0005] To achieve the above object, the present invention is implemented by adopting the following technical solutions: In one aspect, the present invention provides a method for constructing a transformer operating state logic reasoning chain, comprising: According to the acquired historical operating state quantities and historical abnormal state types of the transformer, an operating state quantity set and an abnormal state type set are respectively constructed, as well as a mutual mapping relationship between elements in the operating state quantity set and elements in the abnormal state type set; Randomly select a main characteristic state quantity in the set of operating state quantities, and obtain an abnormal state type matrix corresponding to the main characteristic state quantity and an operating state quantity matrix corresponding to the abnormal state type matrix according to the mutual mapping relationship; Obtain a main network of the transformer operating state logical inference chain according to the main characteristic state quantity and the abnormal state type matrix. Reduce the order of the operating state quantity matrix, and obtain a branch network of the transformer operating state logical inference chain according to the elements of the operating state quantity matrix deleted by the order reduction and the abnormal state type matrix, and connect the branch network to the main network to obtain the transformer operating state logical inference chain.

[0006] Optionally, obtaining the abnormal state type matrix corresponding to the main characteristic state quantity and the operating state quantity matrix corresponding to the abnormal state type matrix according to the mutual mapping relationship comprises: Finding an abnormal state type S that has the mutual mapping relationship with the main characteristic state quantity from the set of abnormal state types, and constructing an abnormal state type matrix according to the abnormal state type S. Traversing the abnormal state types S in the abnormal state type matrix, finding operating state quantities M that have the mutual mapping relationship with the abnormal state types S from the set of operating state quantities, and constructing an operating state quantity matrix according to the operating state quantities M corresponding to all the abnormal state types S.

[0007] Optionally, constructing the operating state quantity matrix according to the operating state quantities M corresponding to all the abnormal state types S comprises: Taking the number of elements in the abnormal state type matrix as a replication number, replicating the operating state quantity M, taking each copy of the operating state quantity M as a row of the operating state quantity matrix, and obtaining the operating state quantity matrix.

[0008] Optionally, obtaining the main network of the transformer operating state logical inference chain according to the main characteristic state quantity and the abnormal state type matrix comprises: Taking the main characteristic state quantity as a starting node of the main network, taking the abnormal state types S in the abnormal state type matrix as terminal nodes of the main network, obtaining the main network of the transformer operating state logical inference chain, and if there are a plurality of abnormal state types S, arranging the plurality of abnormal state types S side by side as the terminal nodes of the main network.

[0009] Optionally, reducing the order of the operating state quantity matrix, obtaining the branch network of the transformer operating state logical inference chain according to the elements of the operating state quantity matrix deleted by the order reduction and the abnormal state type matrix, and connecting the branch network to the main network to obtain the transformer operating state logical inference chain comprises: Expressing the operating state quantity matrix as a sparse matrix, reducing the order of the sparse matrix according to the sparsity of the matrix elements, and obtaining the row elements of the sparse matrix deleted by the order reduction. According to the row elements of the sparsification matrix deleted by the order reduction, the elements of the reduced operating state quantity matrix deleted by the order reduction are obtained; The abnormal state type S which has a mutual mapping relationship with the elements of the reduced operating state quantity matrix deleted by the order reduction is searched from the abnormal state type matrix; The abnormal state type S is taken as the starting node of the branch network, and the elements of the reduced operating state quantity matrix deleted by the order reduction are taken as the terminal node of the branch network, so that the branch network of the transformer operating state logical reasoning chain is obtained; The starting node of the branch network is connected to the terminal node of the trunk network, so that the transformer operating state logical reasoning chain is obtained; If there are multiple abnormal state types S, the multiple abnormal state types S are taken as the starting node of the trunk network in parallel; if there are multiple elements of the reduced operating state quantity matrix deleted by the order reduction, any element of the reduced operating state quantity matrix deleted by the order reduction is taken as the parent node of the terminal node, and other elements of the reduced operating state quantity matrix deleted by the order reduction are taken as the child nodes of the parent node.

[0010] Optionally, the operating state quantity matrix is expressed as a sparsification matrix, including: If the element of any row in the operating state quantity matrix has a mapping relationship with any element in the abnormal state type matrix, the element in the operating state quantity matrix is valued as 1, otherwise, the element is valued as 0, so that a sparsification matrix containing 0 and 1 is obtained; wherein the element of any row in the operating state quantity matrix or the element in the abnormal state type matrix is not repeatedly mapped.

[0011] Optionally, the order reduction of the sparsification matrix includes: The columns of the sparsification matrix are traversed, the column with the least non-zero elements is deleted, and the row corresponding to any non-zero element in the column is deleted, and the step is repeated until the sparsification matrix does not have zero elements.

[0012] Optionally, it further includes: During the order reduction, if there are multiple columns with the least non-zero elements, any column in the multiple columns is deleted; if there are zero elements after the order reduction to 1 row or 1 column, the zero elements are directly deleted.

[0013] In a second aspect, the present application provides a system for constructing a transformer operating state logical reasoning chain, which is used to execute the construction method of the first aspect, and the system includes: A relationship construction module is configured to: according to the obtained operating state quantity and abnormal state type of the transformer, respectively construct an operating state quantity set and an abnormal state type set, and a mutual mapping relationship between the elements in the operating state quantity set and the elements in the abnormal state type set; The matrix obtaining module is configured to: randomly select a main characteristic state variable from the set of operating state variables, and obtain an abnormal state type matrix corresponding to the main characteristic state variable and an operating state variable matrix corresponding to the abnormal state type matrix according to the mutual mapping relationship; The backbone network constructing module is configured to: obtain a backbone network of the logical inference chain of the transformer operating state according to the main characteristic state variable and the abnormal state type matrix. The inference chain constructing module is configured to: reduce the order of the operating state variable matrix, obtain a branch network of the logical inference chain of the transformer operating state according to the operating state variable matrix elements deleted by the order reduction and the abnormal state type matrix, connect the branch network to the backbone network, and obtain the logical inference chain of the transformer operating state.

[0014] In a third aspect, the present application provides a computer readable storage medium, which stores a computer program, and the program is executed by a processor to implement the steps in the method of the first aspect.

[0015] Compared with the prior art, the present application has the following beneficial effects: 1. The present application provides analysis results for equipment state evaluation in a state variable driving and logical inference manner through multi-modal parameter information, can comprehensively establish the causal logical relationship between the transformer state variable and the abnormal state, and can be applied to fault diagnosis, trace analysis and state evaluation in a visual manner, can provide more scientific and detailed technical references for on-site personnel, and can effectively overcome the problems of excessive dependence on the integrity of characteristic variables, excessive attention to part of the state variables, and incomplete evaluation and diagnosis results. 2. The backbone network of the present application focuses on the core fault mode and ensures the basic diagnosis capability; through order reduction of a sparse matrix in mathematics, the essence is to preferentially process strong correlation characteristics, so that the added branches are the most necessary, and in general cases, a complete inference chain is directly constructed, and all mapping relationships between the state variables and the abnormalities need to be processed at one time, which is a relatively large amount of tasks and is easy to miss some fault types. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 Fig. 1 shows a flowchart of the transformer operating state evaluation method in an embodiment of the present application; Figure 2 Fig. 2 shows a logical inference backbone network diagram constructed by taking the oil level abnormal alarm signal as the main characteristic state variable in an embodiment of the present application; Figure 3 Fig. 3 shows a reduction process diagram of the operating state variable matrix when the oil level abnormal alarm signal is taken as the main characteristic state variable in an embodiment of the present application; Figure 4The figure shows a flow chart of adding inference branches corresponding to the operation state quantity matrix order reduction when the oil level abnormality alarm signal is used as the main characteristic state quantity in one embodiment of the present invention; Figure 5 The figure shows a schematic diagram of a transformer operating status evaluation process constructed with an abnormal oil level alarm signal as a main characteristic state quantity in one embodiment of the present invention. DETAILED DESCRIPTION

[0017] The technical solution of the present invention is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0018] The term "and / or" simply describes a relationship between related objects, indicating that three possible relationships exist. For example, "A and / or B" can mean: A exists alone, A and B exist simultaneously, or B exists alone. Additionally, the character " / " generally indicates an "or" relationship between the related objects.

[0019] Example 1

[0020] like Figure 1 As shown, this embodiment introduces a method for constructing a transformer operating state logic reasoning chain, which specifically includes the following steps: Step 1: Organize the transformer's operating status and abnormal status types.

[0021] Step 2: Construct the transformer's operating state quantity set A= [ x1, x2, …, xm ] and exception status type collection B = [ y1, y2, …, yn ], and the mutual mapping relationship between the set of running state quantities and the set of abnormal state types, where m Indicates the number of running state quantities; n Indicates the number of abnormal state types. Among them, the running state quantity set A and a collection of exception status types B There is a one-to-many mapping relationship between them.

[0022] Step 3: Set of running state quantities A Randomly select the i Main characteristic state quantity xi , find the abnormal state type S that has a mutual mapping relationship with the main characteristic state quantity from the abnormal state type set, and construct the abnormal state type matrix according to the abnormal state type S. The abnormal state type matrix is Yi= [yi1, yi2, …, yip traversing the abnormal state type S in the abnormal state type matrix, finding the running state quantity M which has a mutual mapping relationship with the abnormal state type S in the running state quantity set, and the running state quantity combination does not have a mapping relationship with each element in the abnormal state type matrix; copying the running state quantity M according to the number of elements in the abnormal state type matrix as the copy number, taking each running state quantity M as a row of the running state quantity matrix to obtain the running state quantity matrix Xi= [ x11, x12, …, x1q; x21, x22, …, x2q; ……; xp1, xp2, …, xpq ],wherein, p represents the number of elements in the abnormal state type matrix; q represents the number of elements of the running state quantity M obtained according to the mapping relationship of the elements in the abnormal state type matrix.

[0023] In the embodiment, the historical running state quantity, the running state quantity set and the running state quantity matrix are obtained according to the priority category; the priority category is numerical category, image category, text category and artificial judgment category in turn.

[0024] Step four: obtaining the main network of the transformer running state logical inference chain according to the i th main feature state quantity xi and the corresponding abnormal state type matrix Yi , specifically: taking the main feature state quantity as the starting node of the main network and taking the abnormal state type S in the abnormal state type matrix as the termination node of the main network to obtain the main network of the transformer running state logical inference chain; if there are multiple abnormal state types S, the multiple abnormal state types S are taken as the termination node of the main network in parallel, that is: xi → Yi ; wherein, → represents mapping.

[0025] Step five: taking the running state quantity matrix Xi as a sparse matrix; if the element of any row in the running state quantity matrix Xi has a mapping relationship with any element in the abnormal state type matrix Yi , the element in the running state quantity matrix Xi is assigned a value of 1, otherwise, a value of 0, to obtain a sparse matrix containing 0 and 1; wherein, the element of any row in the running state quantity matrix or the element in the abnormal state type matrix is not repeated mapping.

[0026] reducing the order of the sparse matrix, that is: ​Traverse the columns of the sparse matrix, delete the column with the least non-zero elements, and the rows corresponding to any non-zero element in the column, repeat the step until there are no zero elements in the sparse matrix, and obtain the row elements of the reduced and deleted sparse matrix; During the reduction process, if there are multiple columns with the least non-zero elements, any column in the multiple is deleted; if there are zero elements after reduction to 1 row or 1 column, the zero elements are directly deleted.

[0027] According to the row elements of the reduced and deleted sparse matrix, the elements of the reduced and deleted operating state quantity matrix are obtained.

[0028] At the same time, the abnormal state type S which has a mutual mapping relationship with the elements of the reduced and deleted operating state quantity matrix is searched from the abnormal state type matrix; The abnormal state type S is taken as the starting node of the branch network, and the elements of the reduced and deleted operating state quantity matrix are taken as the terminal node of the branch network, to obtain the branch network of the transformer operating state logical reasoning chain. The starting node of the branch network is connected to the terminal node of the trunk network, to obtain the transformer operating state logical reasoning chain, which is represented as: xi → Yi → xij ; wherein, xij represents the elements of the reduced and deleted operating state quantity matrix, i.e. the first i element in the operating state quantity matrix corresponding to the abnormal state type matrix corresponding to the first j main feature state quantity; When reducing each time, if there are multiple abnormal state types S, the multiple abnormal state types S are arranged side by side as the starting node of the trunk network; If there are multiple elements of the reduced and deleted operating state quantity matrix, any element of the reduced and deleted operating state quantity matrix is taken as the parent node of the terminal node, and other elements of the reduced and deleted operating state quantity matrix are taken as the child nodes of the parent node, i.e. the branch network is added according to the binary tree type, one branch network is added at a time, and the addition is performed multiple times.

[0029] Repeat the execution of the step-by-step downward expansion of the logical reasoning chain until all elements in the sparse matrix are non-zero, to obtain the transformer operating state logical reasoning chain.

[0030] Repeat the above steps until the transformer operating state logical reasoning chain covering all elements in the operating state quantity set A and the abnormal state type set B is established.

[0031] Embodiment 2

[0032] Based on embodiment 1, this embodiment introduces a method for constructing a transformer operating state logical reasoning chain, which comprises: In the transformer equipment evaluation, the transformer operation state logical inference chain constructed based on embodiment 1 is used for state inference evaluation, that is: According to the obtained current operation state quantity of the transformer to be evaluated, the transformer operation state evaluation result is obtained based on the pre-constructed transformer operation state logical inference chain, specifically: When the first k operation state quantity xk needs to be evaluated, the first k operation state quantity xk corresponding inference branch is selected in the transformer operation state logical inference chain, and the device state is inferred step by step downward combined with the conditions of other state quantities. If the node operation state quantity is missing, it can be defaulted that the subsequent branches of the node are all valid, and then it is determined whether to delete other branches according to the element judgment in the subsequent nodes. The device state inference result is the most serious one in the final valid branch, and the transformer operation state evaluation result is obtained.

[0033] In this embodiment, the multi-modal parameter information is used to provide analysis results for device state evaluation in a state quantity driven and logical inference manner, which can comprehensively establish the causal logical relationship between the transformer state quantity and the abnormal state.

[0034] Embodiment 3

[0035] Based on embodiment 2, this embodiment introduces a test example of a construction method of a transformer operation state logical inference chain, including the following steps: Taking the oil level abnormal alarm signal as the operation state quantity, assuming that the oil level abnormal alarm signal is A in x1 , the abnormal state types corresponding to it mainly include: the oil level gauge indication is close to zero scale, the oil level gauge indication is close to full scale, and the oil level gauge indication is intermediate scale, that is, three parameters in the abnormal state type set B , assuming that B in y1 , y2 and y3 , the mapping relationship of x1→ [ y1, y2, y3 ] is constructed. It should be noted that y1 may have more operation state quantities corresponding to A , such as oil level gauge damage, breather blockage, etc. These are some state quantities in A .

[0036] Taking the main feature state quantity in the state quantity set A at this time, that is, the oil level abnormal alarm signal x1 as the starting node, according to the mutual mapping relationship, the corresponding abnormal state type matrix Y1 [ y11, y12, y13]= [ y1, y2, y3 ],three abnormal state types y1, y2, y3 Total of 8 state quantities are mapped, i.e. the corresponding running state quantities at this time are obtained, including: unreasonable design of oil tank xi1 , oil leakage xi2 , low historical oil level xi3 , high historical oil level xi4 , damage of oil level gauge xi5 , blockage of respirator xi6 , existence of a large amount of gas in the oil tank xi7 , moisture in the secondary wiring box of the oil level meter xi8 , long-term no change of oil level scale xi9 , the number of elements in the abnormal state type matrix is taken as the number of copies, and the running state quantity is copied, i.e. the abnormal state type matrix Y1 =[ y11, y12, y13 ] and the running state quantity matrix X1 =[ x11, x12, …, x19; x21, x22, …, x29; x31, x32, …, x39 ] The mapping relationship between the abnormal state type and the running state quantity is shown in Table 1.

[0037] Table 1 Mapping relationship between abnormal state type and running state quantity Based on x1→Y1 , the main network of the logical reasoning chain is constructed, as shown in Figure 2 , which indicates three abnormal state types triggered by the oil level gauge alarm signal, which are oil level meter indicating near zero scale, oil level meter indicating near full scale, and oil level meter indicating middle scale.

[0038] The running state quantity matrix X1 can be represented as: .

[0039] The running state quantity matrix X1 is sequentially reduced in terms of the sparsity of row elements, i.e. according to the number of non-zero elements in the column elements from few to many, the column corresponding to the running state quantity matrix X1 and the row corresponding to the non-zero elements in the column are deleted, if the number of non-zero elements in each column of the sparse matrix is the same, then any column with the same number of non-zero elements and the row corresponding to the non-zero elements in the column are deleted, and it can be found that the running state quantity matrix X1 can be reduced, in this embodiment, the first column and the first row of the original matrix are deleted, and the following running state quantity matrix X1 can be updated: X1 : ; You can use any element of the deleted row to build a subsequent logical reasoning chain, that is, add x1 →Y1→x1j The reasoning branch, and according to x1j The impact of transformer operation status on equipment status is evaluated. Taking the unreasonable design of oil tank as an example, the following is obtained: Figure 3 According to the expanded logical reasoning chain shown, if the equipment status is that the oil conservator volume is too small, the equipment is in an abnormal state, and the oil conservator can be replaced or the oil volume can be adjusted according to the "oil temperature-oil level" curve.

[0040] The running state matrix X1 When the order is reduced again, the third column and the first row of the matrix are deleted to obtain the following running state matrix: X1 : ; At this time, the running state matrix X1 If there is only one row, stop reducing the order. If there is only one row left, it means there is only one abnormal state type. You can directly use the operating state quantity corresponding to the non-zero elements of this abnormal state type to make internal distinctions, that is, delete the zero elements until all the elements in the matrix are non-zero, and get the final operating state quantity matrix. X1 : .

[0041] Repeatedly expand the logical reasoning chain step by step until all the abnormal state matrices have non-zero elements, and the classification of the remaining abnormal states is achieved, and we get y1 The corresponding logical reasoning chain branches are as follows Figure 4 As shown, if the equipment status is that the oil conservator volume is too small, the equipment is in an abnormal state, and the oil conservator can be replaced or the oil volume can be adjusted according to the "oil temperature-oil level" curve; if the equipment status is not that the oil conservator volume is too small, further check whether the equipment status is oil leakage; if the equipment status is oil leakage, the oil tank, radiator or oil conservator is hot, and the equipment is in a serious state and can be shut down, and the valves on both sides of the leakage part can be closed or blocked; if the equipment status is not oil leakage, further check whether the equipment status is historically low oil level; if the equipment status is historically low oil level If the oil level is low, the oil level in the main oil storage tank is low. At this time, the equipment is in an abnormal state. You can add oil to the main oil storage tank according to the "oil temperature-oil level" curve of the main body; if the equipment status is not that the historical oil level is low, further check whether the equipment status is that the oil level scale remains unchanged for a long time; if the equipment status is that the oil level scale remains unchanged for a long time, the capsule is damaged. At this time, the equipment is in a serious state and you can replace the capsule; if the equipment status is not that the oil level scale remains unchanged for a long time, the oil level gauge is damaged. At this time, the equipment is in an abnormal state. You can replace the oil level gauge and adjust the oil level according to the "oil temperature-oil level" curve.

[0042] Repeat the above steps until a set of running state quantities is established. Aand a collection of exception status types B The transformer operating status logic reasoning chain of all elements in the transformer is as follows: Figure 5 As shown, specifically: When the oil level gauge indicates the middle scale, if the equipment status shows that the oil level gauge and secondary circuit operations have been carried out recently, the equipment is in an abnormal state and you can stop working immediately to verify whether it is an erroneous operation; if the equipment status shows that the oil level gauge and secondary circuit operations have not been carried out recently, further check whether the equipment status has been rainy or snowy recently, and whether the transformer fire water spraying, water flushing and other operations have been carried out; if the equipment status shows that the transformer fire water spraying, water flushing and other operations have been carried out recently, the secondary junction box of the oil level gauge has been dampened by water. At this time, the equipment is in a serious state and its insulation strength can be restored, and measures can be taken to prevent water ingress and moisture; if the equipment status shows that the transformer fire water spraying, water flushing and other operations have not been carried out recently, the oil level gauge is damaged. At this time, the equipment is in an abnormal state and you can replace the oil level gauge and adjust the oil level according to the "oil temperature-oil level" curve.

[0043] When the oil level gauge indicates close to the full scale, if the equipment status is that the historical oil level is higher, the oil level is higher and the oil quantity is higher. At this time, the equipment is in the attention state, and the oil can be withdrawn according to the "oil temperature-oil level" curve; if the equipment status is not that the historical oil level is higher, further check whether the equipment status is that the breather is blocked; if the equipment status is that the breather is blocked, the equipment is in an abnormal state, and the breather can be replaced, and the oil level can be adjusted according to the "oil temperature-oil level" curve; if the equipment status is not that the breather is blocked, it is a false oil level, and further check whether the equipment status is that the oil level scale has not changed for a long time; if the equipment status is that the oil level scale has not changed for a long time, the oil level gauge is damaged. At this time, the equipment is in an abnormal state, and the oil level gauge can be replaced, and the oil level can be adjusted according to the "oil temperature-oil level" curve; if the equipment status is not that the oil level scale has not changed for a long time, there is a large amount of gas in the diaphragm oil conservator. At this time, the equipment is in an abnormal state, and the oil conservator can be vented, and the oil level can be adjusted according to the "oil temperature-oil level" curve.

[0044] In the case that the oil level gauge indicates close to zero scale, if the device state is that the oil tank volume is too small, the device is in an abnormal state, and the oil tank can be replaced or the oil quantity is adjusted according to the oil temperature-oil level curve; if the device state is not that the oil tank volume is too small, the device state is further detected whether it is oil leakage; if the device state is oil leakage, the oil tank, radiator or oil tank is overheated, at this time the device is in a serious state, and the device can be shut down, the valves on both sides of the leakage part are closed or plugging is adopted; if the device state is not oil leakage, the device state is further detected whether it is a low historical oil level; if the device state is a low historical oil level, the oil level of the main oil tank is low, at this time the device is in an abnormal state, and the main oil tank can be refilled according to the main oil temperature-oil level curve; if the device state is not a low historical oil level, the device state is further detected whether it is that the oil level scale is long-term unchanged; if the device state is that the oil level scale is long-term unchanged, the capsule is damaged, at this time the device is in a serious state, and the capsule can be replaced; if the device state is not that the oil level scale is long-term unchanged, the oil level gauge is damaged, at this time the device is in an abnormal state, and the oil level gauge can be replaced, and the oil level is adjusted according to the oil temperature-oil level curve.

[0045] Based on the construction process of the above logical reasoning chain, for the transformer in operation, when the oil level gauge alarm occurs, the reasoning chain of Figure 5 is used for defect cause analysis and state evaluation, and then corresponding branches are selected according to the oil level gauge scale and the corresponding state quantity, and the reasoning is pushed down level by level.

[0046] At this time, the transformer may issue an oil level abnormal alarm and the oil level gauge indicates close to zero scale, and it is found that the main transformer oil level scale is long-term unchanged, so the transformer may have two potential defects of oil level gauge damage or capsule damage, and related states and maintenance suggestions are pushed.

[0047] Embodiment 4

[0048] Based on the embodiment 1, this embodiment introduces a construction system of a transformer operation state logical reasoning chain, which is used to execute the construction method described in embodiment 1, and the system comprises: a relationship construction module, configured to: according to the obtained transformer operation state quantity and abnormal state type, respectively construct a set of operation state quantities and a set of abnormal state types, and a mutual mapping relationship between elements in the set of operation state quantities and elements in the set of abnormal state types; a matrix acquisition module, configured to: randomly select a main feature state quantity in the set of operation state quantities, and according to the mutual mapping relationship, obtain an abnormal state type matrix corresponding to the main feature state quantity, and an operation state quantity matrix corresponding to the abnormal state type matrix; a backbone network construction module, configured to: according to the main feature state quantity and the abnormal state type matrix, obtain a backbone network of the transformer operation state logical reasoning chain. The reasoning chain construction module is configured to: reduce the order of the operating state quantity matrix, obtain a branch network of the transformer operating state logic reasoning chain according to the operating state quantity matrix elements deleted by the reduction of order and the abnormal state type matrix, and connect the branch network to the main network to obtain the transformer operating state logic reasoning chain.

[0049] The specific function implementation of each module is described in the method of Embodiment 1, and is not described herein.

[0050] Embodiment 5

[0051] Based on Embodiment 1, this embodiment introduces a computer readable storage medium, which stores a computer program, and the program is characterized in that the program is executed by a processor to implement the steps in the method of Embodiment 1.

[0052] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can be in the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program code.

[0053] The present application is described with reference to flowcharts and / or block diagrams of the method, device (system), and computer program product according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to produce a machine, so that the instructions executed by the computer or other programmable data processing devices produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks. Figure 1 The device that implements the functions specified in one or more blocks.

[0054] These computer program instructions can also be stored in a computer readable storage medium that can guide the computer or other programmable data processing devices to work in a specific way, so that the instructions stored in the computer readable storage medium produce a product including instruction devices that implement the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks. Figure 1 The device that implements the functions specified in one or more blocks.

[0055] These computer program instructions can also be loaded into a computer or other programmable data processing devices, so that a series of operational steps are generated to realize the computer-implemented processes, and the instructions executed on the computer or other programmable devices provide a process for implementing the functions specified in the flowchart Figure 1 one flow or multiple flows and / or the functions specified in the block Figure 1 one flow or multiple flows and / or the functions specified in the block

[0056] The embodiments of the present application are described above with reference to the accompanying drawings, but the present application is not limited to the specific embodiments described above, and the specific embodiments described above are merely illustrative, but not restrictive, and those of ordinary skill in the art can make many forms under the inspiration of the present application without departing from the purpose of the present application and the scope protected by the claims, which are all within the protection of the present application.

Claims

1. A method for constructing a transformer operating status logic reasoning chain, characterized in that: include: According to the acquired operating state quantity and abnormal state type of the transformer, an operating state quantity set and an abnormal state type set are respectively constructed, as well as a mutual mapping relationship between elements in the operating state quantity set and elements in the abnormal state type set; Randomly select a main characteristic state quantity from the operating state quantity set, and obtain an abnormal state type matrix corresponding to the main characteristic state quantity and an operating state quantity matrix corresponding to the abnormal state type matrix based on a mutual mapping relationship; According to the main characteristic state quantity and abnormal state type matrix, a backbone network of the transformer operation state logic reasoning chain is obtained; The operating state quantity matrix is ​​reduced in order, and the branch network of the transformer operating state logic reasoning chain is obtained according to the deleted operating state quantity matrix elements and abnormal state type matrix. The branch network is connected to the trunk network to obtain the transformer operating state logic reasoning chain.

2. The method for constructing a transformer operating status logic reasoning chain according to claim 1, characterized in that: According to the mutual mapping relationship, the abnormal state type matrix corresponding to the main characteristic state quantity and the operating state quantity matrix corresponding to the abnormal state type matrix are obtained, including: Searching for an abnormal state type S that has a mutual mapping relationship with the main characteristic state quantity from the abnormal state type set, and constructing an abnormal state type matrix based on the abnormal state type S; Traverse the abnormal state type S in the abnormal state type matrix, search the running state quantity M that has a mutual mapping relationship with the abnormal state type S from the running state quantity set, and construct the running state quantity matrix according to the running state quantity M corresponding to all abnormal state types S.

3. The method for constructing a transformer operating status logic reasoning chain according to claim 2, characterized in that: The running state quantity matrix is ​​constructed based on the running state quantity M corresponding to all abnormal state types S, including: The number of elements in the abnormal state type matrix is ​​used as the number of copies, and the running state quantity M is copied. Each copy of the running state quantity M is used as a row of the running state quantity matrix to obtain the running state quantity matrix.

4. The method for constructing a transformer operating status logic reasoning chain according to claim 1, characterized in that: According to the main characteristic state quantity and abnormal state type matrix, the backbone network of the transformer operation state logic reasoning chain is obtained, including: The main characteristic state quantity is used as the starting node of the backbone network, and the abnormal state type S in the abnormal state type matrix is ​​used as the ending node of the backbone network to obtain the backbone network of the transformer operation state logic reasoning chain; if there are multiple abnormal state types S, the multiple abnormal state types S are listed in parallel as the ending nodes of the backbone network.

5. The method for constructing a transformer operating status logic reasoning chain according to claim 1, characterized in that: The operating state matrix is ​​reduced in order. Based on the deleted operating state matrix elements and abnormal state type matrix, a branch network of the transformer operating state logic reasoning chain is obtained. The branch network is connected to the main network to obtain the transformer operating state logic reasoning chain, which includes: The running state matrix is ​​expressed as a sparse matrix, and the sparse matrix is ​​reduced in rank according to the sparsity of the matrix elements to obtain the reduced-order deleted sparse matrix row elements; According to the reduced-order deleted sparse matrix row elements, the reduced-order deleted running state quantity matrix elements are obtained; Searching the abnormal state type S from the abnormal state type matrix that has a mutual mapping relationship with the elements of the running state quantity matrix deleted by order reduction; The abnormal state type S is used as the starting node of the branch network, and the reduced-order deleted operation state matrix elements are used as the ending nodes of the branch network to obtain the branch network of the transformer operation state logic reasoning chain. Connect the starting node of the branch network to the ending node of the main network to obtain the transformer operation status logical reasoning chain; Among them, if there are multiple abnormal state types S, the multiple abnormal state types S are listed in parallel as the starting nodes of the backbone network; if there are multiple running state quantity matrix elements that are reduced in order and deleted, any running state quantity matrix element that is reduced in order and deleted is used as the mother node of the ending node, and the other running state quantity matrix elements that are reduced in order and deleted are used as the child nodes of the mother node.

6. The method for constructing a transformer operating status logic reasoning chain according to claim 5, characterized in that: Representing the running state matrix as a sparse matrix includes: If there is a mapping relationship between an element in any row of the running state quantity matrix and an element in the abnormal state type matrix, then the element in the running state quantity matrix is ​​assigned a value of 1; otherwise, it is assigned a value of 0, and a sparse matrix containing 0 and 1 is obtained; among which, an element in any row of the running state quantity matrix or any element in the abnormal state type matrix is ​​not mapped repeatedly.

7. The method for constructing a transformer operating status logic reasoning chain according to claim 5, characterized in that: Reducing the rank of a sparse matrix involves: Traverse the columns of the sparsified matrix, delete the column with the least non-zero elements and the row corresponding to any non-zero element in the column, and repeat this step until there are no zero elements in the sparsified matrix.

8. The method for constructing a transformer operating status logic reasoning chain according to claim 7, characterized in that: Also includes: During the order reduction process, if there are multiple columns with the least non-zero elements, any of the multiple columns will be deleted; If there are still zero elements after reducing to 1 row or 1 column, delete the zero elements directly.

9. A system for constructing a transformer operating status logic reasoning chain, characterized in that: For executing the construction method according to any one of claims 1 to 8, the system comprises: A relationship building module is used to: construct an operating state quantity set and an abnormal state type set according to the acquired operating state quantity and abnormal state type of the transformer, and a mutual mapping relationship between elements in the operating state quantity set and elements in the abnormal state type set; A matrix acquisition module is used to randomly select a main characteristic state quantity from the operating state quantity set, and obtain an abnormal state type matrix corresponding to the main characteristic state quantity and an operating state quantity matrix corresponding to the abnormal state type matrix based on a mutual mapping relationship; A backbone network construction module is used to obtain a backbone network of a transformer operation state logic reasoning chain according to the main characteristic state quantity and the abnormal state type matrix; The inference chain construction module is used to: reduce the order of the operating state quantity matrix, obtain the branch network of the transformer operating state logic reasoning chain based on the operating state quantity matrix elements and abnormal state type matrix deleted by the order reduction, connect the branch network to the trunk network, and obtain the transformer operating state logic reasoning chain.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images