Disclosure of Invention
In view of the above, the present invention aims to provide an electric energy meter state on-line monitoring system based on the related information.
In order to solve the technical problems, the electric energy meter state online monitoring system based on the association information comprises a topology construction subsystem and an abnormality analysis subsystem, wherein the topology construction subsystem is used for constructing an electric energy meter topology relation, the abnormality analysis subsystem is used for analyzing abnormal conditions of the electric energy meter according to the electric energy meter topology relation and feedback information of each electric energy meter, and the system is characterized by further comprising a plurality of virtual load units, wherein the virtual load units are used for generating virtual loads, each virtual load unit is provided with an electric energy meter with a corresponding relation, and the virtual load units and the corresponding electric energy meters are provided with association data;
The topology construction subsystem is configured with a topology update policy comprising
A1, generating a topology updating instruction set according to a preset topology information table, and sending a topology updating instruction in the topology updating instruction set to a corresponding virtual load unit;
A2, receiving collected data fed back by each electric energy meter;
A3, extracting acquisition characteristics in the acquisition data, and matching topology characteristic data in a topology information table according to the acquisition characteristics;
step A4, associating the electric energy meters with the same topological characteristic data, and determining the topological relation according to the acquired data;
step A5, determining association information between the electric energy meters according to collected data corresponding to the electric energy meters with topological relation;
a6, marking topological relation lines among the electric energy meters through the associated information to generate corresponding electric energy topological models;
the abnormality analysis subsystem is configured with an abnormality correction strategy and an abnormality analysis strategy, the abnormality analysis strategy is used for analyzing the abnormal condition of the electric energy meter, and the abnormality correction strategy comprises
Step B1, acquiring acquisition data fed back by an electric energy meter;
step B2, corresponding association information is called according to the topological relation between the electric energy meters;
Step B3, updating the acquired data according to the associated information to obtain new abnormal analysis data;
And step B4, bringing the new anomaly analysis data into a corresponding anomaly analysis strategy.
Further, the abnormality analysis subsystem is further configured with an abnormality verification policy, and after determining an abnormal electric energy meter position, the abnormality analysis policy is executed, where the abnormality verification policy includes
Step C1, generating an abnormal verification instruction set according to the position of the abnormal electric energy meter, and sending an abnormal verification instruction in the abnormal verification instruction set to a corresponding virtual load unit;
Step C2, acquiring acquisition data fed back by the electric energy meter;
step C3, bringing the acquired data into an anomaly analysis strategy again;
step C4, repeating the step C2 until an abnormal analysis value of a preset abnormal verification number is obtained;
and step C5, matching the abnormal analysis value with a preset abnormal type table to determine the abnormal type.
Further, the virtual load unit is configured as a programmable ac load.
Further, the topology information table records the topology series, the topology region and the topology association relation of each electric energy meter, in the step A1, the feature abundant value of each electric energy meter and the feature difference reference between the electric energy meters are calculated according to the topology series and the topology association relation, the feature abundant value is calculated according to the topology region and the topology association relation, a topology instruction update set is generated according to the feature abundant value of each electric energy meter, topology feature data is determined in each topology instruction update set according to the feature difference reference, and the topology feature data reflects the corresponding relation between the topology update instruction and the virtual load unit;
when the virtual load unit receives the topology updating instruction, different virtual loads are generated according to different topology updating instructions, and when the electricity utilization branch circuit where the electric energy meter is located is connected to the virtual loads, virtual electricity utilization data with virtual electricity utilization characteristics are generated, wherein the virtual electricity utilization characteristics correspond to the acquisition characteristics.
Further, in the step A5, the correlation function is calculated by a first correlation algorithm, which includesWherein F (x) is given by reflection numberCorresponding to the association function between child nodes numbered βn, n being the numberCorresponding to the number of child nodes,Is numbered asAn acquisition function of the electric energy meter corresponding to the father node,For the collection function of the electric energy meter corresponding to the child node with the number of beta n, an is the inheritance weight corresponding to the nth child node,
Has the following components[ X2n-1,x2n ] is the value range of the acquisition function of the acquisition characteristic corresponding to the child node with the number of beta n, and the association information is generated according to the association function.
Further, in step B3, a correction algorithm is configured to calculate the anomaly analysis data, the correction algorithm includingWherein χw is a certain abnormal analysis value corresponding to new abnormal analysis data, χs is a certain acquisition value corresponding to acquired data, alphad is a reference trust value, alphas is a type correction parameter, the type correction parameter is determined by inquiring a correction type table according to the acquisition type corresponding to the acquisition value, F (x) is a correlation function in correlation information, [ S2, S1] is a value range of the acquisition type in the correlation function, a value condition is determined by inquiring the correction type table according to the corresponding acquisition type, and the value range is determined according to the value condition;
the correction type table is pre-stored with a value condition and a corresponding type correction parameter, and takes the acquisition type as an index.
Further, the anomaly verification policy includes a learning modifier policy, where the learning modifier policy obtains a difference between an anomaly analysis value for verification and an anomaly analysis value obtained initially to obtain a trust anomaly difference value, the learning modifier policy includes a learning modifier policy configured with a reference trust range, modifies the reference trust value through a preset first modification algorithm when the trust anomaly difference value is higher than the reference trust range, has an αd＝δaαd, modifies the reference trust value through a preset second modification algorithm when the trust anomaly difference value is lower than the reference trust range, has an αd＝δeαd, wherein δa is a preset trust gain value, δe is a preset trust attenuation value, and has a δa＞1＞δe.
Further, the anomaly analysis strategy is configured as Pearson correlation coefficient analysis and/or quartile range analysis.
Further, in step C1, the method further includes calculating a verification complexity, and determining the abnormal verification instruction set by querying a preset instruction classification table through the verification complexity, where the verification complexity isRq is verification complexity, qf is the topological level number corresponding to the abnormal electric energy meter, qz1 is the number of father nodes of the abnormal electric energy meter and the number of child nodes of qz2, the instruction classification table stores a plurality of abnormal verification instruction sets, each abnormal verification instruction set corresponds to an abnormal verification number, and the abnormal verification instruction sets take verification complexity as an index.
Further, the step C5 further includes vectorizing the anomaly analysis values according to different anomaly verification instructions, obtaining a vector sum of all the obtained anomaly analysis values to obtain an anomaly analysis vector, determining the anomaly type through a preset anomaly vector analysis table, storing a plurality of anomaly types in the anomaly vector index table, wherein the anomaly types are indexed by using the vector conditions, and outputting the corresponding anomaly types when the obtained anomaly analysis vector accords with the corresponding vector conditions.
The method has the advantages that through the arrangement of the virtual load unit, the topology relation can be accurately verified and determined, collected data abnormality caused by topology relation change is avoided, dependence on input topology information is reduced, and the data relation among the electric energy meters can be accurately determined through the association information, so that the accuracy degree of the data is better when the electric energy meters are subjected to abnormality analysis, the action can be automatically completed, the association information can be dynamically adjusted, and a more accurate abnormality analysis result can be obtained by matching with the existing abnormality analysis strategy.
Detailed Description
The following detailed description of the invention is provided in connection with the accompanying drawings to facilitate understanding and grasping of the technical scheme of the invention.
The utility model provides an on-line monitoring system of electric energy meter 1 state based on association information, including topology construction subsystem 100 and unusual analysis subsystem 200, general electric energy meter 1 state on-line monitoring system is all established based on the topological relation between electric energy meter 1, so the core part of the invention also divides into two subsystems, topology construction subsystem 100 is used for constructing ammeter topological relation, unusual analysis subsystem 200 is used for analyzing the unusual condition of ammeter according to ammeter topological relation and the feedback information of each electric energy meter 1, realize on-line monitoring's purpose, still include a plurality of virtual load units 20, virtual load unit 20 is used for producing virtual load, each virtual load unit 20 has electric energy meter 1 of correspondence, and virtual load unit 20 and corresponding electric energy meter 1 have associated data; the virtual load unit 20 can be understood as actually loading in the statistical range corresponding to the electric energy meter 1, but the virtual load unit 20 is different from the common load, the virtual load stability is higher, the output load value has certainty, the influence on the whole power utilization network can be deduced, and the design is completely different from the original electric energy meter 1 monitoring system, because the original electric energy meter 1 monitoring system performs calculation analysis judgment based on the condition that the actual load value is uncertain, thus higher precision can not be provided, the virtual load is characterized in that 1, the virtual load unit 20 can be directly configured in the electric equipment corresponding to the power grid, different loads are realized by controlling the electric equipment of the power grid to work in different powers, the advantages of lower cost and no need of providing additional wiring networking are realized, the disadvantage of no stability of the power of the electric equipment is overcome, the load range of the virtual load which can be generated by the device is limited, the device cannot adapt to various requirements, the type of electric equipment, the working mode and the internal hardware composition are not uniform, and peripheral equipment needs to be added. 2. The dummy load unit 20 is configured as a programmable ac load. The programmable alternating current load can realize the adjustment of different load resistance values, susceptibilities and capacitive loads, has the advantages that various loads can be adjusted through instructions, meanwhile, the loads can be changed in one instruction, the precision is high, the influence of the loads on the electric energy meter 1 can be expected to be accurate, the disadvantage is that the cost is high, additional wiring networking is needed, 3, the loads can be set into the combination of a plurality of charging circuits and discharging circuits, different load conditions are simulated through the combination of the plurality of different charging circuits, and the electric energy stored during the simulation of the loads is used for compensating working waveforms and the like when the power grid system is needed through the discharging circuits, the advantage is that the electric energy is saved in the mode, but the precision is between schemes 2 and 1, because the precision of the working loads of the charging circuits is not high, and the selectable load composition range and the response efficiency to the instructions are also between schemes 1 and 2.
As a first core function of the present invention, the topology construction subsystem 100 is configured with topology upgrade policies including
Firstly, the topology information table records the topology level, the topology area and the topology association relation of each electric energy meter 1, the topology level refers to the topology level in the power grid, the topology level of the electric energy meter 1 at the tail end is 1, the topology level of the father node corresponding to the electric energy meter 1 at the tail end is 2, and so on, the topology area reflects the mapping of the topology physical position corresponding to the electric energy meter 1 node in the model, and in general, the electric energy meter 1 at a relatively close distance is more likely to replace, The phenomenon of interleaving, the topological association relation reflects the actual connection relation of the electric energy meter 1, and the generation mode of the topological information meter is recorded, Scanning and the like, and the invention aims to generate a corresponding topology update instruction set according to a known topology information table, preferably, selecting the moment when a power station is idle for topology update or selecting the topology update when abnormality occurs, and updating the corresponding topology information table again according to the topology information after each update is completed, so that the topology change condition can be checked more accurately, in the step A1, the feature abundant value of each electric energy meter 1 and the feature difference standard between the electric energy meters 1 are calculated according to the topology information table, the feature abundant value is calculated according to the topology series and the topology association relation, and the feature abundant value reflects the feature abundant degree of the virtual load, for example, the virtual load needs more frequent change, Greater amplitude variation, richness of load variation in inductive and capacitive amounts, more virtual load variation, related variation amount under single instruction, The more the change types are, the more complex the data change of the electric energy meter 1 caused by the load is, vice versa, the reason why the characteristic value is calculated by the topological progression and the topological association relation is as follows, because the higher the topological progression is, the higher the statistical complexity of the power grid involved is, the more complex the corresponding topological association relation is, the more the quantity is, the less easily the change caused by the load change is recognized, so the higher the characteristic value is required, on the other hand, the characteristic difference reference is calculated according to the topological area and the topological association relation, if two electric meters are subordinate to a father node, then if the virtual loads generated by the virtual load units 20 corresponding to the two electric meters are similar at this time, the father node is difficult to distinguish the load generated by which child node, so the concept of the characteristic difference reference is introduced, namely the virtual load is restrained when the virtual load is generated, if two electric meters are closer to each other or are subordinate to the same father node, or the node relationship is closer, the corresponding characteristic difference reference is larger, that is, two similar virtual loads cannot be configured on the two corresponding electric meters, the specific method is that firstly, a topological instruction update set is generated according to the characteristic abundant value of each electric meter 1, an instruction characteristic database is configured, the instruction characteristic database stores a plurality of instruction characteristics, each instruction characteristic has an instruction value, the combination of the plurality of instruction characteristics is selected through the characteristic abundant value, the sum of the instruction values is required to be more than the characteristic abundant value, then the topological update instruction corresponding to the instruction characteristic combination of the current electric meter is determined, the instruction similar value is configured between the instruction characteristics, so that a topological update instruction set can be obtained, and topological characteristic data is determined according to the characteristic difference reference in each topological instruction update set, through exchanging the matching relation between the topology updating instruction and the ammeter under the condition of meeting the feature value until the sum of the similarity values of the instructions is smaller than the feature difference standard, if any matching fails to meet, the instruction feature is required to be searched again according to the feature value, a new topology updating instruction is generated, the topology feature data reflects the corresponding relation between the topology updating instruction and the virtual load unit 20, the topology updating instruction is associated with the virtual load unit 20 which is correspondingly executed, when the virtual load unit 20 receives the topology updating instruction, different virtual loads are generated according to different topology updating instructions, and when the electricity utilization branch of the electric energy meter 1 is accessed to the virtual load, virtual electricity utilization data with virtual electricity utilization features are generated, and the virtual electricity utilization features correspond to the acquisition features.
And A2, receiving the acquired data fed back by each electric energy meter 1, wherein the acquisition mode of the acquired data is not specially improved, so that the details are not repeated.
And A3, extracting the acquisition characteristics in the acquisition data, and matching the topology characteristic data in the topology information table according to the acquisition characteristics, wherein the virtual load is known, so that the acquisition characteristics which can be generated by the virtual load are also known, the corresponding topology characteristic data can be analyzed by extracting the acquisition characteristics, and the child nodes have the same topology characteristic data if the child nodes access the virtual load, and the acquisition characteristics are in one-to-one correspondence with the instruction characteristics, for example, whether the acquisition data has a certain waveform or not is judged, and the acquisition characteristics can be determined, so that the topology characteristic data is determined.
Step A4, associating the electric energy meters 1 with the same topological characteristic data, and determining the topological relation according to the acquired data; the method for determining the topological relation according to the acquired data is that the acquired data of the father node is the acquired data and the superposition of the child nodes, so that the corresponding topological relation can be analyzed through the acquired data.
Step A5, determining association information between the electric energy meters according to collected data corresponding to the electric energy meters with topological relation, wherein in the step A5, an association function is calculated through a first association algorithm, and the association function comprises the following steps ofWherein F (x) is given by reflection numberCorresponding to the association function between child nodes numbered βn, n being the numberCorresponding to the number of child nodes,Is numbered asGβn (x) is the collection function of the electric energy meter corresponding to the child node with the number of beta n, an is the inheritance weight corresponding to the nth child node,
Has the following components[ X2n-1,x2n ] is the value range of the acquisition function of the acquisition characteristic corresponding to the child node with the number of beta n, and the association information is generated according to the association function. The first correlation algorithm can calculate a corresponding correlation function, and aims at that if each correlation information is determined, if the relation of the corresponding nodes is to be calculated, the correlation function can be directly called to eliminate errors, the two correlation functions adopt the formula, firstly, because the acquisition function in the acquired data is known, the collection function of the parent node is also known, so that the sum of the collection functions of the child nodes and the deviation of the collection function of the parent node are regarded as errors, the deviation amount of each child node to the errors is inherited weight, and the inherited weight is calculated through wallpaper of integral of the range corresponding to the collection characteristic corresponding to the child node and accounting for the total range.
And A6, marking the topological relation lines among the electric energy meters through the associated information to generate a corresponding electric energy topological model, and marking the corresponding topological relation lines through the associated information. The labeled model is defined as a power topology model.
The abnormality analysis subsystem is configured with an abnormality correction strategy and an abnormality analysis strategy, the abnormality analysis strategy is used for analyzing the abnormal condition of the electric energy meter, and the abnormality analysis strategy is configured into Pearson correlation coefficient analysis and/or quartile range analysis. The abnormality analysis strategy can also be realized by other analysis methods, only the analysis method takes the collected data of the electric energy meter as the analysis basis, and the analysis result is characterized by a numerical form, so that the analysis method can be suitable for the system. The purpose of the anomaly correction strategy is to make the collected numerical values more accurate in analysis and accurately eliminate errors caused by loss, and the anomaly correction strategy comprises
Step B1, acquiring acquisition data fed back by an electric energy meter;
step B2, corresponding association information is called according to the topological relation between the electric energy meters;
Step B3, updating the acquired data according to the associated information to obtain new abnormal analysis data, wherein in step B3, a correction algorithm is configured to calculate the abnormal analysis data, and the correction algorithm comprisesWherein χw is a certain abnormal analysis value corresponding to new abnormal analysis data, χs is a certain acquisition value corresponding to the acquired data, alphad is a reference trust value, alphas is a type correction parameter, the type correction parameter is determined by inquiring a correction type table according to the acquisition type corresponding to the acquisition value, F (x) is a correlation function in correlation information, [ S2, S1] is a value range of the acquisition type in the correlation function, a value condition is determined by inquiring the correction type table according to the corresponding acquisition type, and the value range is determined according to the value condition. The preset correction type table judges corresponding value conditions according to the acquisition types, because the correction parameters corresponding to different acquisition types are different, for example, the acquisition voltage data, the corresponding correction parameters and the correction parameters corresponding to the acquisition current data are different, meanwhile, the value conditions reflect that the corresponding value ranges are determined according to the stability requirements of a certain specific type within a certain acquisition characteristic range, and particularly for different acquisition parameters of different acquisition types, because a section where one acquisition waveform is located is a combination of a plurality of instruction features, each instruction feature has independent stability except for a recognized function, for example, a certain instruction feature has stability for representing current data, and other instruction features have stability for representing voltage data, the corresponding value ranges in the functions are different, so that the corresponding value ranges can be determined according to the range of the corresponding instruction feature, and the abnormal analysis data can be determined as accurately as possible. The correction type table is pre-stored with a value condition and a corresponding type correction parameter, and the collection type is used as an index.
And step B4, bringing the new anomaly analysis data into a corresponding anomaly analysis strategy. The anomaly analysis strategy can yield more accurate results.
As another functional point of the present invention, the abnormality analysis subsystem is further configured with an abnormality verification policy, which is executed after determining an abnormal electric energy meter position, the abnormality verification policy including
Step C1, generating an abnormal verification instruction set according to the position of the abnormal electric energy meter, and sending an abnormal verification instruction in the abnormal verification instruction set to a corresponding virtual load unit, wherein the step C1 further comprises the steps of calculating verification impurity degree and inquiring a preset instruction classification table through the verification impurity degree, wherein the verification impurity degree is as followsRq is verification complexity, qf is the topological level number corresponding to the abnormal electric energy meter, qz1 is the number of father nodes of the abnormal electric energy meter and the number of child nodes of qz2, the instruction classification table stores a plurality of abnormal verification instruction sets, each abnormal verification instruction set corresponds to an abnormal verification number, and the abnormal verification instruction sets take verification complexity as an index. By calculating the verification complexity, the complexity of the instruction required by the verification of the abnormal electric energy meter can be determined, the corresponding abnormal verification instruction set is determined by calculating the verification complexity, then the corresponding abnormal verification instruction is sent to the virtual load to wait for the next acquisition, and the explanation is that at the moment, each instruction of the abnormal verification instructions is resent for a plurality of times under different logics, the purpose of this is to determine, by means of different virtual loads, the abnormal situation, for example, that the electric energy meter loses the metering effect in the case of capacitive loads or inductive loads, until an abnormal analysis value with an abnormal verification number is obtained, and if this conclusion is to be obtained, it is necessary to carry out a change test on the load, so as to obtain the corresponding conclusion of the abnormal type.
Step C2, acquiring acquisition data fed back by the electric energy meter;
And C3, bringing the acquired data into an anomaly analysis strategy again, wherein the anomaly verification strategy comprises a learning correction sub-strategy which obtains an anomaly analysis value for verification and an anomaly analysis value obtained initially to obtain a trust anomaly difference value, the learning correction sub-strategy comprises a learning correction strategy which is configured with a reference trust range, when the trust anomaly difference value is higher than the reference trust range, the reference trust value is corrected through a preset first correction algorithm and is alphad＝δaαd, when the trust anomaly difference value is lower than the reference trust range, the reference trust value is corrected through a preset second correction algorithm and is alphad＝δeαd, wherein deltaa is a preset trust gain value, deltae is a preset trust attenuation value and is deltaa＞1＞δe. By continuously bringing the anomaly analysis sub-strategy, the obtained result can be corrected, if the anomaly analysis value corresponding to the Pearson correlation coefficient analysis is the correlation value P, and if the anomaly analysis value corresponding to the quarter-bit-distance rule is the voltage judgment score G, the method can be extended to other anomaly analysis strategies according to the anomaly analysis value, then the anomaly situation is judged by the learning correction strategy, the reference trust value can be adjusted, the corresponding influence is regulated, and preferably, deltaa is set to be 1.01, and deltae is set to be 0.96. The whole system is enabled to be continuously true under verification.
Step C4, repeating the step C2 until an abnormal analysis value of a preset abnormal verification number is obtained;
And step C5, matching the abnormal analysis value with a preset abnormal type table to determine the abnormal type. And step C5, vectorizing the abnormal analysis values according to different abnormal verification instructions, solving the vector sum of all the obtained abnormal analysis values to obtain an abnormal analysis vector, determining the abnormal type through a preset abnormal vector analysis table, storing a plurality of abnormal types in an abnormal vector index table, taking vector conditions as indexes, and outputting the corresponding abnormal type when the obtained abnormal analysis vector accords with the corresponding vector conditions. Since different abnormality types may exist and a plurality of abnormality analysis values may occur in the verification process, by constructing a vectorized design, different areas of the coordinate system are divided to correspond to different abnormality types, and the corresponding abnormality types can be determined according to vectorized results. And is convenient to maintain or replace.
Of course, the above is only a typical example of the invention, and other embodiments of the invention are also possible, and all technical solutions formed by equivalent substitution or equivalent transformation fall within the scope of the invention claimed.