Substation secondary equipment modeling method and device based on knowledge graph, and medium

Abstract

The application discloses a substation secondary equipment modeling method and device based on a knowledge graph and a medium. The method comprises the following steps: first classifying a basic database of substation secondary equipment according to the types of the secondary equipment, obtaining a first classification database, processing a data set of each type of secondary equipment in the first classification database according to characteristic information of the secondary equipment, obtaining a characteristic matrix and an equipment type matrix of all secondary equipment of the substation, calculating the correlation degree of all secondary equipment of the substation through a correlation degree calculation network trained according to the characteristic matrix and the equipment type matrix of all secondary equipment of the substation, constructing a knowledge graph according to obtained entities, attributes and relations, and completing substation secondary equipment modeling. The application classifies the secondary equipment first, extracts features from the classified data set, and calculates the correlation degree for each feature, thereby improving the working efficiency and the modeling accuracy.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of electrical secondary equipment management, and in particular to a substation secondary equipment modeling method based on a knowledge graph, equipment and a medium. BACKGROUND

[0002] The power secondary equipment is auxiliary equipment for monitoring, measuring, controlling, adjusting and protecting the primary equipment of the power system, and is not directly connected with the main circuit of the power. With the development of the knowledge graph technology, more and more people pay attention to the application of the knowledge graph to the power system. CN116187175A discloses a power transformation and distribution station neural network identification method based on an electrical industry knowledge graph, so that the graph elements and the data models are one-to-one corresponding, the electrical drawing standard is introduced, the knowledge graph for the complex user power transformation and distribution station design engineering field is constructed, on the basis of the knowledge graph, the electrical system diagram parameter intelligent identification method based on the recursive neural network learning algorithm is used, and the power distribution station equipment plane layout diagram generation method based on the graph reasoning is used, the electrical various design diagrams are automatically generated through the definition of the related rules by the graph reasoning, the automatic understanding of the electrical system diagram and the building plane diagram is realized, and the automatic generation of the power transformation and distribution station design scheme is realized on the basis of the automatic acquisition of the design parameters. However, in the actual scene, the cost of the perfect knowledge graph model is high, the data division is not standardized in the construction of the knowledge graph due to the various types of the secondary equipment, the working efficiency is low when the direct processing operation is performed, and the single feature attribute cannot be processed, so that the modeling accuracy is reduced. SUMMARY

[0003] Technical purpose: in view of the defects in the prior art, the application discloses a substation secondary equipment modeling method based on a knowledge graph, equipment and a medium, first, the secondary equipment is classified, the features of the classified data set are extracted, the correlation degree is calculated for each feature, the entity, the attribute and the relationship are obtained, and then the knowledge graph is constructed according to the entity, the attribute and the relationship, so that the working efficiency is improved and the modeling accuracy is improved.

[0004] Technical scheme: in order to realize the above technical purpose, the application adopts the following technical scheme.

[0005] A substation secondary equipment modeling method based on a knowledge graph, comprising the following steps:

[0006] Step S1, for the types of the secondary equipment, the first classification is performed on the basic database of the substation secondary equipment, and a first classification database is obtained;

[0007] Step S2, the data set of each type of secondary equipment in the first classification database is processed according to the feature information of the secondary equipment, and a feature matrix and a device type matrix of all secondary equipment of the substation are obtained;

[0008] Step S3, the correlation degree of all secondary devices in the substation is calculated by the trained correlation degree calculation network for the feature matrix and the device type matrix of all secondary devices in the substation;

[0009] Step S4, the knowledge graph is constructed according to the entities, attributes and relationships obtained in step S3, and the modeling of the secondary devices in the substation is completed.

[0010] Preferably, in step S1, the secondary devices are divided into T types according to the device type, and a first classification database DATA is constructed, wherein DATA = {data 1 ,data 2 ,...,data t ,...,data T}, 1≤t≤T, data t represents the data set of the tth type of secondary device, 1≤n≤N, N is the total number of the tth type of secondary device; is the nth secondary device set in the tth type of secondary device.

[0011] Preferably, in step S2, The feature information of the secondary device is included in the feature information of the secondary device, and the feature information includes ID feature attribute, static feature attribute, dynamic feature attribute, early warning feature attribute and control feature attribute; the ID feature attribute is a feature used by the secondary device to represent its own identity, the static feature attribute is a feature in the secondary device that does not change with time, the dynamic feature attribute is a feature in the secondary device that changes with time, the early warning feature attribute is whether the secondary device can feedback early warning information, and the control feature attribute is whether the secondary device can control the signal or loop;

[0012] The feature information of all secondary devices in the substation is encoded to obtain the feature matrix FM e of the secondary device = [fm e0 ,fm e1 ,...,fm er ,...,fm eR ], wherein 1≤r≤R, R is the number of feature information, fm e0 is the type of the e th secondary device, and fm er is the encoding of the rth feature information of the e th secondary device. The device type matrix TYPE is constructed separately for the type of the secondary device = [fm 10 ,fm 20 ,...,fm e0 ,...,fm E0 ], and the feature information matrix of the e th secondary device is FM' e = [fm e1 ,...,fm er..., fm eR ]; wherein, 1≤e≤E, E is the total number of all secondary devices of the substation.

[0013] Preferably, the correlation degree calculation network in step S3 includes a first network and a second network, the first network includes an ID feature attribute correlation degree network for outputting entities in the knowledge graph, and the second network includes a static feature attribute correlation degree network, a dynamic feature attribute correlation degree network, a pre-warning feature attribute correlation degree network, and a control feature attribute correlation degree network for outputting attributes and relationships in the knowledge graph.

[0014] Preferably, the ID feature attribute correlation degree network includes:

[0015] The input unit inputs and outputs a device type matrix TYPE, TYPE=[fm 10 ,fm 20 ..., fm e0 ..., fm E0 ]; the dimension is 1,

[0016] The hidden unit one connected with the input unit and the hidden unit three has an input of the device type matrix TYPE, ID feature attributes in the feature information, and a linear conversion output OUT3' of the hidden unit three; the weight between the hidden unit one and the input unit is 1≤i≤I, I is the dimension of the hidden unit one, and the activation function of the hidden unit one is f 1 (x); the output of the hidden unit one is

[0017] The hidden unit two connected with the hidden unit one and the input unit has an input of the device type matrix TYPE and the output OUT1 of the hidden unit one, and the weight between the hidden unit two and the hidden unit one and the input unit is The dimension of the hidden unit two is E, the activation function of the hidden unit two is f 2 (x), and the output of the hidden unit two is

[0018] The hidden unit three connected with the hidden unit two has an input of the output OUT2 of the hidden unit two, and the weight between the hidden unit three and the hidden unit two is The activation function of the hidden unit three is f 3 (x), the dimension of the hidden unit three is I, and the output of the hidden unit three is The output threshold interval [δ min ,δ max ] of the preset hidden unit three is set, δ min ,δ maxrespectively, and the output OUT3 of the hidden unit three is input to the output unit if the output OUT3 of the hidden unit three is within the output threshold interval, and the output OUT3 of the hidden unit three is linearly converted to obtain the linearly converted output OUT3' of the hidden unit three if the output OUT3 of the hidden unit three is not within the output threshold interval, the linearly converted output OUT3' of the hidden unit three is returned to the hidden unit one processing;

[0019] an output unit, an input of the output unit being the output OUT3 of the hidden unit three, and a weight between the output unit and the hidden unit three being a dimension of the output unit being E, and an output of the output unit being the correlation degree between the e-th secondary device and other secondary devices,

[0020] the correlation degree obtaining the correlation degree between the secondary devices based on the ID feature attributes as entities in the knowledge graph.

[0021] Preferably, the early warning feature attribute correlation degree network comprises:

[0022] a first input unit, an input and an output of the first input unit both being a feature information matrix of the e-th secondary device, i.e., FM' e = [fm e1 ,...,fm er ,...,fm eR ], 1≤r≤R, R being a feature information quantity, and a dimension of the first input unit being 1;

[0023] a first hidden unit connected with the first input unit, an input of the first hidden unit being the feature information matrix of the e-th secondary device, and a weight between the first hidden unit and the first input unit being a dimension of the first hidden unit being 1, and an activation function of the first hidden unit being g 1 (x); an output of the first hidden unit being λ1 being a first hidden unit parameter, and 0<λ1<1, by adjusting the value of the first hidden unit parameter λ1, the output of the first hidden unit is within the first early warning feature attribute threshold interval [0,1], and the output of the first hidden unit is input to the first output unit;

[0024] a first output unit connected with the first hidden unit, an input of the first output unit being the output of the first hidden unit, and an output of the first output unit being wherein θ is a preset second early warning feature attribute threshold; the output OUT 2 of the first output unit being a secondary device early warning feature attribute, and an output result being 0 or 1;

[0025] the control feature attribute correlation degree network comprises:

[0026] The second input unit inputs and outputs the feature information matrix of the e-th secondary device, that is, FM e e1 ,...,fm er ,...,fm eR , 1≤r≤R, R is the number of feature information, and the dimension of the second input unit is 1;

[0027] The second hidden unit connected with the second input unit inputs the feature information matrix of the e-th secondary device, and the weight between the second hidden unit and the second input unit is The dimension of the second hidden unit is 1, and the activation function of the second hidden unit is g 2 (x); the output of the second hidden unit is λ2 is the second hidden unit parameter, and 0<λ2<1; by adjusting the value of the second hidden unit parameter λ2, the output of the second hidden unit is in the first control feature attribute threshold interval [0, 1], and the output of the second hidden unit is input to the second output unit;

[0028] The second output unit connected with the second hidden unit inputs the output of the second hidden unit; the output of the second output unit is Wherein, η is a preset second control feature attribute threshold; the output OUT 4 of the second output unit is the control feature attribute of the secondary device, and the output result is 0 or 1;

[0029] The static feature attribute correlation degree network comprises:

[0030] The third input unit connected with the first output unit and the second output unit inputs and outputs the feature information matrix of the e-th secondary device and the output OUT 2 of the first output unit and the output OUT 4 of the second output unit, also recorded as IN 3 , and the dimension of the third input unit is R+2; wherein, IN 3 is represented as:

[0031]

[0032] Wherein, is the j-th value in the output of the third input unit;

[0033] The third hidden unit connected with the third input unit inputs the output of the third input unit and the output OUT4 of the knowledge graph model correlation degree calculation network; the weight between the third hidden unit and the third input unit is ​1≤j≤R+2, R is the number of characteristic information, 1≤e≤E, E is the total number of all secondary devices of the substation; the activation function of the third hidden unit is g 3 (x); the output of the third hidden unit is

[0034] The fourth hidden unit connected with the third hidden unit, the input of the fourth hidden unit is the output OUT 5 of the third hidden unit; the weight between the fourth hidden unit and the third hidden unit is 1≤h≤H, H is the dimension of the fourth hidden unit; the activation function of the fourth hidden unit is g 4 (x); the output of the fourth hidden unit is

[0035] The third output unit connected with the fourth hidden unit, the input of the third output unit is the output OUT 6 of the fourth hidden unit; the weight between the fourth hidden unit and the third output unit is The activation function of the third output unit is g 5 (x); the dimension S of the third output unit is also the number of static characteristic attributes; the output of the third output unit is also the static characteristic attribute of the secondary device;

[0036] The dynamic characteristic attribute correlation degree network comprises:

[0037] The fourth input unit connected with the first output unit, the second output unit and the third output unit, the input and the output are the feature information matrix of the e-th secondary device and the output OUT 2 of the first output unit, the output OUT 4 of the second output unit and the output OUT 7 of the third output unit, also recorded as IN 4 , the dimension of the fourth input unit is S+R+2; wherein, IN 4 is expressed as:

[0038]

[0039] wherein, is the v-th value in the output of the third input unit; 1≤v≤S+R+2, R is the number of characteristic information, S is the number of static characteristic attributes;

[0040] The fifth hidden unit connected with the fourth input unit, the input of the fifth hidden unit is the output of the fourth input unit and the output OUT4 of the knowledge graph model correlation degree calculation network; the dimension of the fifth hidden unit is E; the weight between the fifth hidden unit and the fourth input unit is 1≤e≤E, E is the total number of all secondary devices of the substation; the activation function of the fifth hidden unit is g 6 (x); the output of the fifth hidden unit is

[0041] The sixth hidden unit connected with the fifth hidden unit, the input of the sixth hidden unit is the output OUT 8 of the fifth hidden unit; the weight between the sixth hidden unit and the fifth hidden unit is 1≤p≤P, P is the dimension of the sixth hidden unit; the activation function of the sixth hidden unit is g 7 (x); the output of the sixth hidden unit is

[0042] The fourth output unit connected with the sixth hidden unit, the input of the fourth output unit is the output OUT 9 of the sixth hidden unit; the weight between the sixth hidden unit and the fourth output unit is The activation function of the fourth output unit is g 8 (x); the dimension D of the fourth output unit is also the number of dynamic characteristic attributes; the output of the fourth output unit is also the dynamic characteristic attribute of the secondary device;

[0043] The output of the second network is the correlation degree relate between the secondary devices of the substation, relate = [OUT 2 , OUT 4 , OUT 7 , OUT 10 ]; used as attributes and relations in the knowledge graph.

[0044] An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing the above-described method for modeling secondary devices of a substation based on a knowledge graph when executing the program.

[0045] A computer-readable storage medium storing computer-executable instructions for executing the above-described method for modeling secondary devices of a substation based on a knowledge graph.

[0046] Beneficial effects: the application first classifies the knowledge graph model database, facilitates quick discrimination of substation secondary equipment to query the knowledge graph, then calculates the correlation degree of all secondary equipment of the substation, then calculates the correlation degree of the substation secondary equipment and the knowledge graph model, finally determines the knowledge graph corresponding to the highest coincidence rate value as the knowledge graph corresponding to the substation secondary equipment according to the calculated coincidence rate, selects the highest matching knowledge graph from the knowledge graph model database, completes the modeling of the substation secondary equipment, realizes the reuse of the knowledge graph in the knowledge graph model database, greatly improves the work efficiency, and reduces unnecessary work cost. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 The method flowchart of the application;

[0048] Figure 2 The first network structure schematic diagram of the application;

[0049] Figure 3 The second network structure schematic diagram of the application. DETAILED DESCRIPTION

[0050] The application will be further described and explained in combination with the accompanying drawings.

[0051] As shown in the accompanying Figure 1 A substation secondary equipment modeling method based on a knowledge graph, comprising the following steps:

[0052] Step S1, for the type of secondary equipment, the first classification of the basic database of the substation secondary equipment is carried out, and the first classified database is obtained;

[0053] The secondary devices are divided into T types according to the device type, and a first classification database DATA is constructed. In some embodiments of the present application, the secondary devices can be divided into seven categories, that is, T=7, including: electrical parameter measurement category, insulation monitoring category, signal control category, relay protection category, power supply category, high-frequency wave blocking category, and other comprehensive automation device category. The secondary devices of the electrical parameter measurement category are mainly used for monitoring and measuring the electrical parameters such as current, voltage, and power of the circuit, such as various measuring tables, etc. The secondary devices of the insulation monitoring category are mainly used for monitoring the insulation condition of the AC and DC power grids. The secondary devices of the signal control category are mainly used for closing and tripping the circuit breaker in the power distribution device through manual or automatic operation according to the operation circuit. The secondary devices of the relay protection category are mainly used for monitoring the operating condition of the primary devices, quickly responding to abnormalities and accidents, and protecting and controlling through the circuit breaker, such as relays, automatic devices, etc. The secondary devices of the power supply category are mainly used for providing stable DC power for other secondary devices, such as battery packs, DC generators, silicon rectifier devices, emergency lighting power, backup power, etc. The secondary devices of the high-frequency wave blocking category are mainly used for preventing high-frequency current from leaking to the transformer substation or branch line and reducing high-frequency energy loss, and form a power line high-frequency communication channel with coupling capacitors, combination filters, high-frequency cables, high-frequency communication machines, etc. The secondary devices of the other comprehensive automation device category are automation devices other than the above six types of secondary devices, such as backup automatic switching devices, etc.

[0054] DATA = {data 1 ,data 2 ,...,data t ,...,data T}, 1≤t≤T, data t represents the data set of the tth type of secondary device, 1≤n≤N, N is the total number of the tth type of secondary device; is the nth secondary device set in the tth type of secondary device;

[0055] Step S2, for the data set of each type of secondary device in the first classification database, the characteristic information of the secondary device is processed to obtain the characteristic matrix and the device type matrix of all secondary devices in the transformer substation;

[0056] The characteristic information of the secondary equipment includes ID characteristic attribute, static characteristic attribute, dynamic characteristic attribute, early warning characteristic attribute and control characteristic attribute; the ID characteristic attribute is a characteristic of the secondary equipment for representing its own identity, including name, model, manufacturer and the like; the static characteristic attribute is a characteristic of the secondary equipment that does not change with time, and the number of the static characteristic attribute is S, such as the voltage value of the standby power supply and the like; the dynamic characteristic attribute is a characteristic of the secondary equipment that changes with time, and the number of the dynamic characteristic attribute is D, such as the real-time voltage value and current value of the circuit and the like; the early warning characteristic attribute is whether the secondary equipment can feed back early warning information, if yes, the early warning characteristic attribute value is 1, if not, the early warning characteristic attribute value is 0; the early warning information includes fault reporting, insulation condition reporting and the like; the control characteristic attribute is whether the secondary equipment can control a signal or a circuit, for example, whether the secondary equipment can control the closing and tripping of a circuit breaker, whether the secondary equipment can start a standby power supply and the like, if yes, the control characteristic attribute value is 1, if not, the control characteristic attribute value is 0;

[0057] The characteristic information of all secondary equipment of the substation is coded to obtain a characteristic matrix FM of the secondary equipment e = [fm e0 ,...,fm e1 ,...,fm er ,...,fm eR ], wherein 1≤r≤R, R is the number of characteristic information, fm e0 is the type of the e-th secondary equipment, and fm er is the coding of the r-th characteristic information of the e-th secondary equipment; a device type matrix TYPE is constructed separately for the type of the secondary equipment, TYPE = [fm 10 ,...,fm 20 ,...,fm e0 ,...,fm E0 ], the characteristic information matrix of the e-th secondary equipment is FM' e = [fm e1 ,...,fm er ,...,fm eR ]; wherein 1≤e≤E, E is the total number of all secondary equipment of the substation;

[0058] In step S3, the correlation degree calculation network after training is used to calculate the correlation degree of all secondary equipment of the substation according to the characteristic matrix and the device type matrix of all secondary equipment of the substation;

[0059] The correlation degree calculation network comprises a first network and a second network, the first network comprises an ID feature attribute correlation degree network, and is configured to output an entity in the knowledge graph; and the second network comprises a static feature attribute correlation degree network, a dynamic feature attribute correlation degree network, a pre-warning feature attribute correlation degree network and a control feature attribute correlation degree network, and is configured to output an attribute and a relationship in the knowledge graph.

[0060] The ID feature attribute correlation degree network comprises:

[0061] An input unit, both the input and the output are device type matrices TYPE, TYPE=[fm 10 ,fm 20 ,...,fm e0 ,...,fm E0 ]; the dimension is 1,

[0062] A first hidden unit connected with the input unit and a third hidden unit, the input of the first hidden unit is the device type matrix TYPE, an ID feature attribute in the feature information and a linear conversion output OUT'3 of the third hidden unit, the weight between the first hidden unit and the input unit is 1≤i≤I, I is the dimension of the first hidden unit, and the activation function of the first hidden unit is f 1 (x); the output of the first hidden unit is

[0063] A second hidden unit connected with the first hidden unit and the input unit, the input of the second hidden unit is the device type matrix TYPE and the output OUT1 of the first hidden unit, and the weight between the second hidden unit and the first hidden unit and the input unit is The dimension of the second hidden unit is E, the activation function of the second hidden unit is f 2 (x), and the output of the second hidden unit is

[0064] A third hidden unit connected with the second hidden unit, the input of the third hidden unit is the output OUT2 of the second hidden unit, and the weight between the third hidden unit and the second hidden unit is The activation function of the third hidden unit is f 3 (x), the dimension of the third hidden unit is I, and the output of the third hidden unit is The output threshold interval [δ min ,δ max ] of the preset third hidden unit is set, δ min ,δ maxrespectively, and the output OUT3 of the hidden unit three is input to the output unit if the output OUT3 of the hidden unit three is within the output threshold interval, and the output OUT3 of the hidden unit three is linearly converted to obtain the linearly converted output OUT3' of the hidden unit three if the output OUT3 of the hidden unit three is not within the output threshold interval, the linearly converted output OUT3' of the hidden unit three is returned to the hidden unit one processing;

[0065] an output unit, an input of the output unit being the output OUT3 of the hidden unit three, and a weight between the output unit and the hidden unit three being a dimension of the output unit being E, and an output of the output unit being the correlation degree between the e-th secondary device and other secondary devices obtaining the correlation degree between the secondary devices based on the ID feature attributes as entities in the knowledge graph;

[0066] the early warning feature attribute correlation degree network comprises:

[0067] a first input unit, an input and an output of the first input unit both being the feature information matrix of the e-th secondary device, i.e., FM' e ,...,fm e1 ,...,fm er ,...,fm eR , 1≤r≤R, R being the number of feature information, and a dimension of the first input unit being 1;

[0068] a first hidden unit connected with the first input unit, an input of the first hidden unit being the feature information matrix of the e-th secondary device, and a weight between the first hidden unit and the first input unit being a dimension of the first hidden unit being 1, and an activation function of the first hidden unit being g 1 (x); an output of the first hidden unit being λ1 being a first hidden unit parameter, and 0<λ1<1, by adjusting the value of the first hidden unit parameter λ1, the output of the first hidden unit is within the first early warning feature attribute threshold interval [0, 1], and the output of the first hidden unit is input to the first output unit;

[0069] a first output unit connected with the first hidden unit, an input of the first output unit being the output of the first hidden unit, and an output of the first output unit being wherein θ is a preset second early warning feature attribute threshold; the output OUT 2 of the first output unit being the early warning feature attribute of the secondary device, and an output result being 0 or 1;

[0070] the control feature attribute correlation degree network comprises:

[0071] The second input unit inputs and outputs the feature information matrix of the e-th secondary device, i.e., FM e e1 ,...,fm er ,...,fm eR , 1≤r≤R, R is the number of feature information, and the dimension of the second input unit is 1;

[0072] The second hidden unit connected with the second input unit inputs the feature information matrix of the e-th secondary device, and the weight between the second hidden unit and the second input unit is The dimension of the second hidden unit is 1, and the activation function of the second hidden unit is g 2 (x); the output of the second hidden unit is λ2 is the second hidden unit parameter, and 0<λ2<1; by adjusting the value of the second hidden unit parameter λ2, the output of the second hidden unit is in the first control feature attribute threshold interval [0, 1], and the output of the second hidden unit is input to the second output unit;

[0073] The second output unit connected with the second hidden unit inputs the output of the second hidden unit; the output of the second output unit is Wherein, η is a preset second control feature attribute threshold; the output OUT 4 of the second output unit is the control feature attribute of the secondary device, and the output result is 0 or 1;

[0074] The static feature attribute correlation degree network comprises:

[0075] The third input unit connected with the first output unit and the second output unit inputs and outputs the feature information matrix of the e-th secondary device and the output OUT 2 of the first output unit and the output OUT 4 of the second output unit, also denoted as IN 3 , and the dimension of the third input unit is R+2; wherein, IN 3 is represented as:

[0076]

[0077] Wherein, is the j-th value in the output of the third input unit;

[0078] The third hidden unit connected with the third input unit inputs the output of the third input unit and the output OUT4 of the knowledge graph model correlation degree calculation network; the weight between the third hidden unit and the third input unit is ​1≤j≤R+2, R is the number of characteristic information, 1≤e≤E, E is the total number of all secondary devices of the substation; the activation function of the third hidden unit is g 3 (x); the output of the third hidden unit is

[0079] A fourth hidden unit connected with the third hidden unit, the input of the fourth hidden unit is the output OUT 5 of the third hidden unit; the weight between the fourth hidden unit and the third hidden unit is 1≤h≤H, H is the dimension of the fourth hidden unit; the activation function of the fourth hidden unit is g 4 (x); the output of the fourth hidden unit is

[0080] A third output unit connected with the fourth hidden unit, the input of the third output unit is the output OUT 6 of the fourth hidden unit; the weight between the fourth hidden unit and the third output unit is The activation function of the third output unit is g 5 (x); the dimension S of the third output unit is also the number of static characteristic attributes; the output of the third output unit is which is also the static characteristic attribute of the secondary device;

[0081] The dynamic characteristic attribute correlation degree network comprises:

[0082] A fourth input unit connected with the first output unit, the second output unit and the third output unit, the input and the output are all the characteristic information matrix of the e-th secondary device and the output OUT 2 of the first output unit, the output OUT 4 of the second output unit and the output OUT 7 of the third output unit, also recorded as IN 4 The dimension of the fourth input unit is S+R+2; wherein, IN 4 is expressed as:

[0083]

[0084] wherein, is the v-th value in the output of the third input unit; 1≤v≤S+R+2, R is the number of characteristic information, S is the number of static characteristic attributes;

[0085] A fifth hidden unit connected with the fourth input unit, the input of the fifth hidden unit is the output of the fourth input unit and the output OUT4 of the knowledge graph model correlation degree calculation network; the dimension of the fifth hidden unit is E; the weight between the fifth hidden unit and the fourth input unit is 1≤e≤E, E is the total number of all secondary devices of the substation; the activation function of the fifth hidden unit is g 6 (x); the output of the fifth hidden unit is

[0086] A sixth hidden unit connected with the fifth hidden unit, the input of the sixth hidden unit is the output OUT of the fifth hidden unit 8 ; the weight between the sixth hidden unit and the fifth hidden unit is 1≤p≤P, P is the dimension of the sixth hidden unit; the activation function of the sixth hidden unit is g 7 (x); the output of the sixth hidden unit is

[0087] A fourth output unit connected with the sixth hidden unit, the input of the fourth output unit is the output OUT of the sixth hidden unit 9 , the weight between the sixth hidden unit and the fourth output unit is The activation function of the fourth output unit is g 8 (x); the dimension D of the fourth output unit is also the number of dynamic characteristic attributes; the output of the fourth output unit is , which is also the dynamic characteristic attribute of the secondary device;

[0088] The output of the second network is the correlation degree relate between the secondary devices of the substation, relate=[OUT 2 , OUT 4 , OUT 7 , OUT 10 ]; used as attributes and relations in the knowledge graph;

[0089] The first network and the second network are trained; samples are constructed for the first network and the second network and training is performed, to obtain the trained correlation degree calculation network. Network training is a conventional technical means in the art, and will not be repeated here.

[0090] Step S4, constructing a knowledge graph according to the entities, attributes and relations obtained in step S3, completing the modeling of the secondary devices of the substation.

[0091] The application first classifies the secondary devices, extracts features from the classified data set, calculates the correlation degree for each feature, obtains entities, attributes and relations, and then constructs a knowledge graph according to the entities, attributes and relations, thereby improving the working efficiency and the accuracy of modeling.

[0092] The application further discloses an electronic device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the modeling method of the substation secondary equipment based on the knowledge graph when executing the program. The memory can be various types of memories, such as random access memory, read-only memory, flash memory, etc. The processor can be various types of processors, such as central processing units, microprocessors, digital signal processors or image processors, etc.

[0093] A computer readable storage medium stores computer executable instructions for executing the modeling method of the substation secondary equipment based on the knowledge graph. The storage medium comprises various types of storage media that can store program codes, such as U disks, mobile hard disks, ROMs, RAMs, magnetic or optical disks, etc.

[0094] The above description is only preferred embodiments of the application, and it should be pointed out that, for those skilled in the art, some improvements and refinements can be made without departing from the principles of the application, and these improvements and refinements should also be considered as the protection scope of the application.

Claims

1. A substation secondary equipment modeling method based on a knowledge graph, characterized in that, The method comprises the following steps: Step S1, first classifying the basic database of the secondary equipment of the substation according to the types of the secondary equipment, to obtain a first classified database; Step S2, processing the data set of each type of secondary equipment in the first classified database according to the characteristic information of the secondary equipment, to obtain a characteristic matrix and a device type matrix of all secondary equipment of the substation; the characteristic information comprises ID characteristic attributes, static characteristic attributes, dynamic characteristic attributes, early warning characteristic attributes and control characteristic attributes; the ID characteristic attributes are characteristics used by the secondary equipment to represent its own identity, the static characteristic attributes are characteristics in the secondary equipment that do not change with time, the dynamic characteristic attributes are characteristics in the secondary equipment that change with time, the early warning characteristic attributes are whether the secondary equipment can feedback early warning information, and the control characteristic attributes are whether the secondary equipment can control signals or loops; Step S3, calculating the correlation degree of all secondary equipment of the substation through the trained correlation degree calculation network according to the characteristic matrix and the device type matrix of all secondary equipment of the substation; the correlation degree calculation network comprises a first network and a second network; the first network comprises an ID characteristic attribute correlation degree network, which is used to output entities in the knowledge graph; the second network comprises a static characteristic attribute correlation degree network, a dynamic characteristic attribute correlation degree network, an early warning characteristic attribute correlation degree network and a control characteristic attribute correlation degree network, which are used to output attributes and relationships in the knowledge graph; Step S4, constructing a knowledge graph according to the entities, attributes and relationships obtained in step S3, to complete the modeling of the secondary equipment of the substation. 2.The substation secondary equipment modeling method based on a knowledge graph according to claim 1, characterized in that: In step S1, the secondary devices are divided into T types according to the device type, and a first classification database DATA is constructed, where DATA = {data 1 ,data 2 ,...,data t ,...,data T}, 1≤t≤T, data t represents a data set of the tth type of secondary device, 1≤n≤N, N is the total number of the tth type of secondary device; is the nth secondary device set in the tth type of secondary device.

3. The substation secondary equipment modeling method based on a knowledge graph according to claim 2, characterized in that: In step S2, The feature information of the secondary device is included in the message. Encoding the characteristic information of all secondary devices in the substation to obtain a characteristic matrix FM of the secondary devices e = [fm e0 , fm e1 ,..., fm er ,..., fm eR ], where 1≤r≤R, R is the number of characteristic information, fm e0 is the type of the e-th secondary device, fm er is the encoding of the r-th characteristic information of the e-th secondary device, a device type matrix TYPE is separately constructed for the type of the secondary device, TYPE = [fm 10 , fm 20 ,..., fm e0 ,..., fm E0 ], and the characteristic information matrix of the e-th secondary device is FM' e = [fm e1 ,..., fm er ,..., fm eR ]; where 1≤e≤E, E is the total number of all secondary devices in the substation.

4. The substation secondary equipment modeling method based on a knowledge graph according to claim 3, characterized in that: The ID characteristic attribute correlation degree network comprises: Input unit, input and output are both device type matrix TYPE, TYPE = [fm 10 ,fm 20 ,...,fm e0 ,...,fm E0 ]; dimension is 1, The hidden unit one connected with the input unit, the hidden unit three, the input of the hidden unit one is the device type matrix TYPE, the ID feature attribute in the feature information and the linear conversion output OUT'3 of the hidden unit three; the weight between the hidden unit one and the input unit is 1≤i≤I, I is the dimension of the hidden unit one, the activation function of the hidden unit one is f 1 (x); the output of the hidden unit one is a hidden unit two connected to the hidden unit one and the input unit, an input of the hidden unit two being the device type matrix TYPE and the output OUT1 of the hidden unit one, a weight between the hidden unit two and the hidden unit one and the input unit being a dimension of the hidden unit two being E, an activation function of the hidden unit two being f 2 (x), an output of the hidden unit two being a third hidden unit connected to the second hidden unit, an input of the third hidden unit being the output OUT2 of the second hidden unit, a weight between the third hidden unit and the second hidden unit being an activation function of the third hidden unit being f 3 (x), a dimension of the third hidden unit being I, an output of the third hidden unit being a preset output threshold interval [δ min ,δ max ] of the third hidden unit, δ min ,δ max being respectively an output threshold minimum value and an output threshold maximum value; if the output OUT3 of the third hidden unit is within the output threshold interval, the output OUT3 of the third hidden unit is input to the output unit, if the output OUT3 of the third hidden unit is not within the output threshold interval, the output OUT3 of the third hidden unit is linearly converted to obtain a linearly converted output OUT3' of the third hidden unit, the linearly converted output OUT3' of the third hidden unit returning to the first hidden unit for processing. an output unit, the input of the output unit being the output OUT3 of the hidden unit three, the weight between the output unit and the hidden unit three being the dimension of the output unit being E, the output being the relevance between the e-th secondary device and other secondary devices obtaining the relevance between the secondary devices based on the ID feature attributes as entities in the knowledge graph.

5. The substation secondary equipment modeling method based on a knowledge graph according to claim 3, characterized in that: The early warning characteristic attribute correlation degree network comprises: The first input unit inputs and outputs a feature information matrix of the e-th secondary device, i.e., FM e = [fm e1 ,...,fm er ,...,fm eR ], 1≤r≤R, R is the number of feature information, and the dimension of the first input unit is 1; a first hidden unit connected with the first input unit, an input of the first hidden unit being a feature information matrix of the e-th secondary device, and a weight between the first hidden unit and the first input unit being a dimension of the first hidden unit being 1, and an activation function of the first hidden unit being g 1 (x); an output of the first hidden unit being λ1 being a first hidden unit parameter, and 0 < λ1 < 1, by adjusting a value of the first hidden unit parameter λ1, the output of the first hidden unit is within a first pre-warning feature attribute threshold interval [0, 1], and the output of the first hidden unit is input to the first output unit. a first output unit connected with the first hidden unit, an input of the first output unit being an output of the first hidden unit, and an output of the first output unit being wherein θ is a preset second early warning feature attribute threshold value; the output OUT of the first output unit 2 is an early warning feature attribute of the secondary device, and the output result is 0 or 1; The control characteristic attribute correlation degree network comprises: The second input unit inputs and outputs a feature information matrix of the e-th secondary device, i.e., FM e = [fm e1 ,...,fm er ,...,fm eR ], 1≤r≤R, R is the number of feature information, and the dimension of the second input unit is 1; a second hidden unit connected with the second input unit, an input of the second hidden unit being a feature information matrix of the e-th secondary device, and a weight between the second hidden unit and the second input unit being a dimension of the second hidden unit being 1, and an activation function of the second hidden unit being g 2 (x); an output of the second hidden unit being λ2 being a second hidden unit parameter, and 0 < λ2 < 1, by adjusting a value of the second hidden unit parameter λ2, the output of the second hidden unit is within a first control feature attribute threshold interval [0, 1], and the output of the second hidden unit is input to the second output unit. a second output unit connected with the second hidden unit, an input of the second output unit being an output of the second hidden unit; an output of the second output unit being wherein η is a preset second control feature attribute threshold value; the output OUT of the second output unit 4 a control feature attribute of the secondary device, the output result being 0 or 1; The static characteristic attribute correlation degree network comprises: a third input unit connected with the first output unit and the second output unit, inputting and outputting a characteristic information matrix of the e-th secondary device and the output OUT of the first output unit 2 , the output OUT of the second output unit 4 , also denoted as IN 3 , the dimension of the third input unit is R+2; wherein IN 3 is expressed as: wherein, is the jth value in the output of the third input unit; a third hidden unit connected with the third input unit, an input of the third hidden unit being an output of the third input unit and the output OUT4 of the knowledge graph model correlation degree calculation network; a weight between the third hidden unit and the third input unit being 1≤j≤R+2, R being a quantity of characteristic information, 1≤e≤E, E being a total number of all secondary devices of the transformer substation; an activation function of the third hidden unit being g 3 (x); an output of the third hidden unit being a fourth hidden unit connected to the third hidden unit, an input of the fourth hidden unit being an output OUT of the third hidden unit 5 a weight between the fourth hidden unit and the third hidden unit being 1≤h≤H, H being a dimension of the fourth hidden unit; an activation function of the fourth hidden unit being g 4 (x); an output of the fourth hidden unit being a third output unit connected to the fourth hidden unit, an input of the third output unit being an output OUT of the fourth hidden unit 6 a weight between the fourth hidden unit and the third output unit being an activation function of the third output unit being g 5 (x); a dimension S of the third output unit, also a number of static feature attributes; an output of the third output unit being also a static feature attribute of the secondary device; The dynamic characteristic attribute correlation degree network comprises: a fourth input unit connected to the first output unit, the second output unit, and the third output unit, and inputting and outputting a characteristic information matrix of the e-th secondary device and the output OUT of the first output unit 2 , the output OUT of the second output unit 4 , the output OUT of the third output unit 7 , also denoted as IN 4 The dimension of the fourth input unit is S+R+2; wherein IN 4 is represented as: wherein, is the vth value in the output of the third input unit; 1≤v≤S+R+2, R is the number of characteristic information, and S is the number of static characteristic attributes. a fifth hidden unit connected with the fourth input unit, an input of the fifth hidden unit being the output of the fourth input unit and the output OUT4 of the knowledge graph model correlation degree calculation network; a dimension of the fifth hidden unit being E; a weight between the fifth hidden unit and the fourth input unit being 1≤e≤E, E being a total number of all secondary devices of the transformer substation; an activation function of the fifth hidden unit being g 6 (x); an output of the fifth hidden unit being a sixth hidden unit connected to the fifth hidden unit, the input of the sixth hidden unit being the output OUT of the fifth hidden unit 8 ; the weight between the sixth hidden unit and the fifth hidden unit being 1≤p≤P, P being the dimension of the sixth hidden unit; the activation function of the sixth hidden unit being g 7 (x); the output of the sixth hidden unit being The fourth output unit connected to the sixth hidden unit, the input of the fourth output unit is the output OUT of the sixth hidden unit 9 , the weight between the sixth hidden unit and the fourth output unit is The activation function of the fourth output unit is g 8 (x); the dimension D of the fourth output unit is also the number of dynamic feature attributes; the output of the fourth output unit is It is also the dynamic characteristic attribute of the secondary equipment; The output of the second network is the correlation degree relate between the substation secondary devices, relate = [OUT 2 , OUT 4 , OUT 7 , OUT 10 ]; used as attributes and relationships in the knowledge graph.

6. An electronic device, comprising: The computer program stored in the memory and executable on the processor comprises a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor executes the program to implement the method of claim 1-5.

7. A computer-readable storage medium, characterized in that, The computer executable instructions are stored in the memory, and the computer executable instructions are used to execute the method of claim 1-5.

Images