CN105158723A - Error evaluation system and method for digital electric energy metering system - Google Patents

Error evaluation system and method for digital electric energy metering system Download PDF

Info

Publication number
CN105158723A
CN105158723A CN201510458899.6A CN201510458899A CN105158723A CN 105158723 A CN105158723 A CN 105158723A CN 201510458899 A CN201510458899 A CN 201510458899A CN 105158723 A CN105158723 A CN 105158723A
Authority
CN
China
Prior art keywords
electric energy
error
current
energy metering
digital electric
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201510458899.6A
Other languages
Chinese (zh)
Other versions
CN105158723B (en
Inventor
张秋雁
李红斌
程含渺
魏伟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huazhong University of Science and Technology
Guizhou Electric Power Test and Research Institute
Original Assignee
Huazhong University of Science and Technology
Guizhou Electric Power Test and Research Institute
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huazhong University of Science and Technology, Guizhou Electric Power Test and Research Institute filed Critical Huazhong University of Science and Technology
Priority to CN201510458899.6A priority Critical patent/CN105158723B/en
Publication of CN105158723A publication Critical patent/CN105158723A/en
Application granted granted Critical
Publication of CN105158723B publication Critical patent/CN105158723B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Abstract

The invention discloses a digital electric energy metering system error evaluation method based on a multi-parameter degradation model, and belongs to the field of smart power grid equipment on-line state evaluation. Through sensors installed in key positions of a transformer station, operating environment data of the digital electric energy metering device are obtained, and the evaluation method based on the multi-parameter degradation model is adopted to evaluate an error state of the digital electric energy metering system, thereby realizing state monitoring and maintenance of the digital electric energy metering system in the true sense. The error evaluation system provided by the invention includes an environment sensing unit, a current and voltage sensing unit, a computer and evaluation software running on the computer. The error evaluation system and method provide an engineering on-line state monitoring scheme for the digital electric energy metering system, evaluate the error state of the digital electric energy metering system, and provide an on-line monitoring means for engineering application of a digital electric energy metering technology.

Description

A kind of error evaluation system and method for digitalized electric energy metering system
Technical field
The invention belongs to intelligent grid equipment on-line status monitoring field, more specifically, relate to a kind of digitalized electric energy metering system error evaluation method based on many reference amounts degradation model.
Background technology
Along with the progress of science and technology, electric system is towards digitizing and intelligent direction development, and digital transformer substation is the important component part of intelligent grid.In digital transformer substation, electric energy metrical no longer adopts traditional electric power meter, but adopts totally digitilized electric power meter, and typical three-phase digital electric power meter composition as shown in Figure 1.Be positioned on high-tension side combined electronic transformer 1-3 and be respectively A combined formula electric mutual inductor, B combined formula electric mutual inductor and the combined formula electric mutual inductor of C, wherein, combined electronic transformer comprises the electronic current mutual inductor and electronic type voltage transformer that fit together, synchronous through synchronous clock 4, gather electric current and the voltage of A phase bus 5, B phase bus 6 and C phase bus 7 respectively, the data of collection with certain form as FT3 to send to the merge cells 8 of low-pressure side through optical fiber.The sampled data of reception is sent to digital electric energy meter 9 with IEC61850-9-2/LE frame format by merge cells 8 again.Digital electric energy meter 9 adopts certain electric energy algorithm to calculate electric energy.Digital electric energy meter 9 has RS485 or infrared interface, for communicating with other equipment.
Numeral electric energy metered system is one of key link of intelligent substation, realizes the application of digital electric energy metered system through engineering approaches and repair based on condition of component, meets " strong intelligent grid " requirement.In recent years, have a lot to the theory of digital electric energy metered system and experimental study, fundamental purpose promotes its stability and reliability, advances its through engineering approaches to apply.For digital electric energy metered system, the most important thing is error characteristics, in order to grasp the kinematic error rule of digital electric energy metered system, the R&D institutions such as Jiangsu DianKeYuan and Guangdong DianKeYuan establish digital electric energy metrical monitoring system, running environment parameter and the error information of digital electric energy metered system can be obtained, but do not utilize these data further.The electric energy metered system thermodynamic state verification strategy that Chongqing DianKeYuan proposes, makes the repair based on condition of component of electric energy metered system further again.Although these work achieve achievement, Shortcomings part respectively: (1) does not make full use of the error state of the data assessment metering system that monitoring system obtains; (2) state of empirical evaluation metering system is relied on but not scientific and reasonable mathematical model.
The current state estimation model for most of power transmission and transforming equipment and appraisal procedure all have comparatively perfect research system, but be substantially in blank for digital electric energy metered system assessment models and method, cause electric energy metered system can only periodic verification or occur error in dipping overproof after interruption maintenance, have impact on power supply reliability, too increase O&M cost.The difficult point of numeral electric energy metered system error evaluation is: (1) error effect factor is many, and error in dipping is the result of these combined factors effects, and being difficult to set up single factors affects model to error in dipping; (2) the error performance parameter of electric energy metered system and influence amount data many, and each other without definite relation, be difficult to adopt traditional data disposal route to obtain the useful effect of result (3) each influence factor to error in dipping shifting, make its appraisal procedure significantly be different from the performance estimating method of most equipment, there is no ripe available assessment models.The digitalized electric energy metering system stability problem exposed that is in operation makes it in a large number for engineering practice, and also can not lack effective ways to the error evaluation of digitalized electric energy metering system always.In a word, in order to correctly assess the error of digitalized electric energy metering system, need adopt new assessment models and adopt modern data processing method.
The present invention is directed to digitalized electric energy metering system feature, adopt many reference amounts degeneration error evaluation model, in conjunction with error evaluation constraint condition, reasonable assessment is made to digital electric energy metering systematic error state.Be beneficial to and advance the application of digital Electric Energy Metering Technology through engineering approaches, advance digital electric energy metered system repair based on condition of component, there is most important theories and practice significance.
Summary of the invention
For problems of the prior art, the application provides a kind of error evaluation system and method for digitalized electric energy metering system, wherein by studying the concrete structure of BP neural network and set-up mode thereof and relate to, achieve digital electric energy metering directional error state on_line monitoring.
For achieving the above object, according to one aspect of the present invention, provide a kind of error evaluation system of digitalized electric energy metering system, this error evaluation system is used for carrying out online error state assessment to digital electric energy metering device, and comprise environmentally sensitive unit, current/voltage sensing unit and CPU (central processing unit), it is characterized in that:
Described environmentally sensitive unit comprises temperature sensor, humidity sensor, the second microprocessor, and the second direct supply; Wherein this temperature sensor, humidity sensor are respectively used to perform temperature to environment residing for the digitalized electric energy metering system as monitoring target, humidity measures in real time, and give described second microprocessor by common transport, this second direct supply is then for providing working power to other component units in described;
Described current/voltage sensing unit comprises electromagnetic current transducer, electromagnetic potential transformer, first microprocessor, and is the first direct supply that in current/voltage sensing unit, each device is powered; Wherein said current transformer, described voltage transformer (VT) are used for measuring in real time the frequency of described digitalized electric energy metering system and harmonic wave;
In addition, described CPU (central processing unit) is connected with first, second microprocessor described simultaneously, and sets up in real time according to measurement data and upgrade many reference amounts degradation model, assesses digital electric energy metered system error state in short-term in current and future:
The integrated data analyzing and processing software of described computer run is as state estimation platform, (1) difference normalization preprocess method is adopted to carry out pre-service to described measurement data, (2) artificial neural network is adopted to train degenerate network to be specially again: the measurement data through difference normalized to be input to described degeneration grid, to try to achieve the degeneration grid that differential pair is answered; (3) last and future current according to the described digital electric energy metered system of described measurement data assessment of described each sensing unit again error state in short-term, thus realize the error state monitoring of digitalized electric energy measuring apparatus.
According to another aspect of the present invention, provide a kind of error evaluation method of digitalized electric energy metering system, it is characterized in that, this system comprises the following steps:
(1) under the prerequisite of same employing rate, respectively the factor of the influential system error of described digitalized electric energy metering system and corresponding systematic error are measured; The factor of described influential system error is one or more in environment temperature, ambient humidity, electric parameter, frequency and harmonic wave, electromagnetic field, vibration, communication abnormality, decompression, cutout;
(2) difference normalization preprocess method is adopted to carry out pre-service to described measurement data;
(3) artificial neural network is adopted to train degenerate network again, adopt artificial neural network to carry out learning training, and verify training network, whether training of judgement result is reasonable, as unreasonable then retraining again, until validation error meets setting value;
(4) last again according to the described measurement data of described each sensing unit by the described digital electric energy metered system of described degenerate network assessment current and future error state in short-term, thus the error state realizing digitalized electric energy measuring apparatus is monitored.
Preferably, in described step (3), the measurement data through difference normalized is input to described degeneration grid, tries to achieve the degeneration grid that differential pair is answered, assess the error of electric energy metered system with this.
In general, according to above-mentioned technical conceive of the present invention compared with prior art, following technological merit is mainly possessed:
1, realize digital electric energy metering directional error state on_line monitoring, for digital electric energy metered system provides a kind of engineering grade status monitoring truly and repair method.
Accompanying drawing explanation
Figure 1 shows that typical digitalized electric energy measuring apparatus structural representation in prior art;
Figure 2 shows that single-phase electric energy metering system input and output isoboles;
Figure 3 shows that electric current, voltage measurement error schematic diagram;
Figure 4 shows that the appraisal procedure figure based on the assessment of many reference amounts degradation model error state;
Figure 5 shows that the digitalized electric energy metering system error evaluation method system composition frame chart based on many reference amounts degradation model;
Figure 6 shows that BP neural network structure schematic diagram;
Figure 7 shows that temperature raw data plot figure;
Figure 8 shows that Temperature pre-treatment data and curves figure;
Figure 9 shows that ECT ratio raw data plot figure;
Figure 10 shows that ECT ratio preprocessed data curve map;
Figure 11 shows that training iterations and error relationship figure;
Figure 12 shows that error evaluation result and actual motion error comparison chart.
Embodiment
In order to make object of the present invention, technical scheme and advantage clearly understand, below in conjunction with drawings and Examples, the present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.In addition, if below in described each embodiment of the present invention involved technical characteristic do not form conflict each other and just can mutually combine.
Described assessment models is many reference amounts degradation model.Concrete intension is as follows:
The parameter degradation of an equipment or system generally means hydraulic performance decline.For complex apparatus or system, the parameter characterizing performance often has multiple, makes to need when setting up state estimation model to consider these parameters simultaneously.In transformer state evaluation, such as relate to the parameter such as micro-water in oil dissolved gas, insulation resistance, dielectric loss, oil; Relate to contact electrical wear, divide-shut brake coil current in circuit-breaker status evaluation, cut-off the parameters such as stroke.For Transformer State Assessment, the set that the parameter supposing to characterize certain transformer performance is formed is defined as { P1, P2, ..., Pn}, one is simple and effective Performance Evaluation Model is: as max|Pi| (i=1,2, ..., when n) exceeding threshold value, provide assessment result " a certain performance of this transformer can drop to the equipment of impact and normally work ".
Above-mentioned Performance Evaluation Model is not suitable for digitalized electric energy metering system assessment.Reason mainly contains 3 points: a certain performance parameter of (1) metering system is degenerated, metering system entirety generally can not be caused to occur, and the degeneration of each parameter of operation troubles (2) metering system has directivity, namely parameter degradation not definitely makes error in dipping increase on the impact of electric energy metrical, also may in shifting thus do not affect overall error in dipping, such as voltage transformer (VT) ratio becomes large to positive dirction, current transformer ratio becomes large to negative direction, and metering system still correctly can measure electric energy; (3) under above-mentioned (1) and (2) prerequisite, can not with the error of the weighted sum of maximum amount of degradation or each amount of degradation or the actual digital electric energy metered system of product assessment.Therefore, for digitalized electric energy metering system, need to propose new assessment models.
For simplifying modeling, only consider single-phase electric energy metering.The average power expression formula of single-phase electric energy metering is:
P = 1 T ∫ 0 T u ( t ) i ( t ) d t - - - ( 1 )
Single-phase digital electric energy metered system can be equivalent to a dual input single output system, and it is input as analog voltage and electric current, and export as power, system equivalent schematic as shown in Figure 2.
As seen from Figure 2, electric energy metrical sums up in the point that the end is measuring voltage and electric current.The measuring error of voltage and current directly causes electric energy metering error, and therefore the degeneration parameter of electric energy metered system can be thought following four: voltage measurement ratio, voltage measurement angular difference, current measurement ratio, current measurement angular difference.
Meanwhile, to the system shown in Fig. 2, ratio and angular difference measuring error essence are U (i) and I (i) measuring error, as shown in Figure 3 A and Figure 3 B.
In Fig. 3 A, curve 1 is original input voltage, and curve 2 is for only having the input voltage of ratio, and curve 3 is the input voltage simultaneously having ratio and angular difference.Suppose that the degeneration parameter of single-phase digital electric energy metered system initial time is 0, after putting into operation, they are respectively in respective dimension variation, and change direction is uncertain.Certain moment, four parameters degenerated to f pT, δ pT, f cTand δ cT, now discrete power should be:
P ′ = 1 N Σ i = 0 N U ( i ) ′ ′ I ( i ) ′ ′ = 1 N Σ i = 0 N α i β i U ( i ) I ( i ) - - - ( 2 )
In formula, α iby f pTand δ pTu (i) error coefficient caused, α i=h 1(f pT, δ pT); β iby f cTand δ cTi (i) variation factor caused, β i=h 2(f cT, δ cT), α iand β ispan be :-1≤α≤1 ,-1≤β≤1.
From formula (2), work as α iβ iwhen=1, electric energy metrical accuracy is constant, and reason is that the impact of degeneration on electric energy metrical of each parameter is cancelled out each other, and makes error in dipping be 0.
Based on above analysis, the assessment of numeral electric energy metering error has significant otherness with the Performance Evaluation Model of general device or system, the key distinction is that the parameter degeneration of digital electric energy metered system has directivity, on the impact of error in dipping, arbitrary parameter is degenerated should consider that other parameters current change.Therefore, propose based on many reference amounts degeneration error evaluation model for electric energy metered system herein.
Existing theoretical research is thought, the digital electric energy metered system that transformer station puts into operation affects by many-side, comprises environment temperature, humidity; Electric parameter is as load, frequency and harmonic wave; Electromagnetic field, vibration, communication abnormality; The grid event such as decompression, cutout.Normally work be divided into reliability and stability influence factor according to whether affecting electric energy metered system; Continuation and sporadic influence factor is divided into according to influence time.
Conveniently the present invention describes accurately clear, and further simplify this model, the continuation factor that choosing affects stability carries out modeling, is respectively environment temperature, humidity, mains frequency, harmonic wave.Four influence factor actings in conjunction, in digital electric energy metered system, affect four degeneration parameter f pT, δ pT, f cT, δ cT, available following system of equations describes:
δ C T ( t ) = g 1 ( T e m p ( t ) , H u m ( t ) , F r e q ( t ) , H a r m ( t ) ) f C T ( t ) = g 2 ( T e m p ( t ) , H u m ( t ) , F r e q ( t ) , H a r m ( t ) ) δ P T ( t ) = g 3 ( T e m p ( t ) , H u m ( t ) , F r e q ( t ) , H a r m ( t ) ) f P T ( t ) = g 4 ( T e m p ( t ) , H u m ( t ) , F r e q ( t ) , H a r m ( t ) ) - - - ( 4 )
In above formula, Temp (t), Hum (t), Freq (t), Harm (t), be respectively temperature, humidity, frequency, harmonic wave function over time.Order:
ϵ → = ( δ C T ( t ) , f C T ( t ) , δ P T ( t ) , f P T ( t ) ) - - - ( 5 )
α → = ( T e m p ( t ) , H u m ( t ) , F r e q ( t ) , H a r m ( t ) ) - - - ( 6 )
Solution of equations matrix is designated as A, then equation (4) can be written as:
ϵ → = A α → - - - ( 7 )
Formula (7) is digital electric energy metered system parameter degeneration equation.Wherein, parameter degeneration vector, be degeneration parameter, definition A is the many reference amounts degenerate network of degeneration parameter under effect vector.
The error of electric energy metered system is defined as assessment constraint condition:
η P = P ′ - P P - - - ( 8 )
In above formula, η pfor mean power error.Formula (7) and (8) form digitalized electric energy metering system error evaluation model jointly.
Described error state appraisal procedure is that its concrete intension is as follows based on the assessment of many reference amounts degradation model error state:
The present invention is based on based on the assessment of many reference amounts degradation model error state appraisal procedure as shown in Figure 4, comprise data prediction, training degenerate network and the error evaluation of metering system.
(1) data prediction.Sampling rate unification is carried out to input many reference amounts, then does difference normalized, obtain respectively:
{ΔTemp(1),ΔTemp(2),...,ΔTemp(n)};
{ΔHum(1),ΔHum(2),...,ΔHum(n)};
{ΔFreq(1),ΔFreq(2),...,ΔFreq(n)};
{ΔHarm(1),ΔHarm(2),...,ΔHarm(n)};
{Δδ CT(1),Δδ CT(2),...,Δδ CT(n)};
{Δf CT(1),Δf CT(2),...,Δf CT(n)};
{Δδ PT(1),Δδ PT(2),...,Δδ PT(n)};
{Δf PT(1),Δf PT(2),...,Δf PT(n)}。
Loud by { Δ Temp (i), Δ Hum (i), Δ Freq (i), Δ Harm (i) } formation effect, as the input of degenerate network shown in Fig. 3, { Δ δ cT(i), Δ f cT(i), Δ δ pT(i), Δ f pT(i) } output of pie graph 3.
(2) degenerate network is trained.According to the input and output sample sequence that (1) obtains, adopt BP neural network to carry out learning training, and verify training network, whether training of judgement result is reasonable, as unreasonable then retraining again, until validation error meets setting value;
(3) metering system error evaluation.By the action sequence { Δ Temp, Δ Hum, the Δ Freq that measure, Δ Harm} is input to degenerate network, try to achieve the variable quantity of degeneration parameter, then obtain each state parameter of current electric energy metered system in conjunction with degeneration parameter initial value, finally assess the error current of electric energy metered system.
The sensor of digitalized electric energy metering system error evaluation method based on many reference amounts degradation model provided by the invention by installing at each key position of transformer station, obtain the degradation effects amount of digital electric energy metered system, comprise three bus 5-7 accurately the measured current of three bus 5-7 that records of electric current and voltage and combined electronic transformer and voltage, combined electronic transformer 1-3 ambient temperature and humidity etc., described influence amount is formed the parameter degradation effects vector of digital electric energy metered system; Many reference amounts degradation model is set up again by the history run state of digital electric energy metered system; Finally parameter degradation effects vector is inputed to degradation model, draw digital electric energy metered system error state in short-term in current and future.Thus realize digital electric energy metering directional error state on_line monitoring, for digital electric energy metered system provides a kind of engineering grade status monitoring truly and repair method.
Figure 5 shows that the digitalized electric energy metering system error evaluation block diagram of system based on many reference amounts degradation model of the embodiment of the present invention.As shown in Figure 5, the digitalized electric energy metering system error evaluation method system based on many reference amounts degradation model comprises environmentally sensitive unit 10, current/voltage sensing unit 13 and computing machine 14.
Environmentally sensitive unit 10 comprises temperature sensor 101, humidity sensor 102, first microprocessor 103, and is the first direct supply 104 that in environmentally sensitive unit 10, each device is powered.In embodiments of the present invention, temperature sensor 101 can select Digital Measurement of Temperature chip DS18B20; Humidity sensor 102 can select STTS751; First microprocessor 103 selects the MPS430F149 low-power scm of TI company; SD6-S05A1 unit power supply selected by first direct supply 104.
Current/voltage sensing unit 11 comprises current transformer 131, voltage transformer (VT) 132, the 4th microprocessor 133, and is the 4th direct supply 134 that in current/voltage sensing unit 13, each device is powered.In embodiments of the present invention, current transformer 131 and voltage transformer (VT) 132 require to select the electromagnetic potential transformer and electromagnetic current transducer that meet class of accuracy requirement according to accuracy of measurement, and its measured value can as standard primary voltage and primary current; 4th microprocessor 133 can select the microprocessor MPC860 of PowerPC framework; SD25-S12D1 unit power supply selected by 4th direct supply 134.
Computing machine 14 obtains the measurement data of each sensing module 10,13.Computing machine 14 comprises data preprocessing module, degenerate network training module, error state evaluation module and memory module, adopt the error analysis method based on degeneration thought, first many reference amounts degradation model is set up by digital electric energy metered system history data, training many reference amounts degenerate network, re-use measurements and calculations gained key parameter composition amount of degradation effect vector, described effect vector is input to described many reference amounts degenerate network, thus the error state in short-term in current and future of assessment digitalized electric energy measuring apparatus.Current micro computer hardware configuration is high, and its processor calculating speed is fast, and hard-disk capacity is large, generally can meet computing and memory requirement.In embodiments of the present invention, Hewlett-Packard HPPavilionP6-1480CN microcomputer selected by computing machine 14.
The data obtained based on the digitalized electric energy metering system error evaluation system of many reference amounts degradation model comprise the service data of the many reference amounts of digitalized electric energy measuring apparatus, parameter degeneration amount and digitalized electric energy measuring apparatus itself.Environmentally sensitive unit 10, current/voltage sensing unit 11 form sensing and monitoring system, continuously online many reference amounts and the parameter degeneration amount related data obtaining digital electric energy metering device, comprising: the temperature and humidity etc. of the ratio that measured current accurately on three bus 5-7 recording of electric current and voltage and combined electronic transformer of three bus 5-7 and voltage and the electronic current voltage obtained thus are measured and angular difference, environment.
Illustrate the acquisition methods of each sensing unit mounting means and 6 groups of data below.
1, environmental sensory data: the pedestal place three same environmentally sensitive unit 10 being arranged on respectively combined electronic transformer 1-3, for obtaining temperature, the humidity of combined electronic transformer 1-3 running environment.
2, primary current voltage sensor data accurately: three same current/voltage sensing units 13 being installed on three bus 5-7, for obtaining the accurate primary current of three bus 5-7 and accurate primary voltage, being sent to computing machine 14 by network interface.
Figure 4 shows that the computing machine 14 of the embodiment of the present invention carries out the process flow diagram of error state assessment.Parameter degeneration amount and parameter are carried out data processing by computing machine 14 difference normalization data preprocess method, carry out based on BP neural metwork training with the parameter degenerate network of historical data to digital electric energy metered system, finally parameter action is inputed to degenerate network, obtain digital electric energy metered system error state in short-term in current and future.Concrete, in embodiments of the present invention, what belong to digital electric energy metered system degeneration parameter is ECT measuring error, EVT measuring error etc.; What belong to degeneration parameter action is that environment temperature, humidity, electric parameter are as harmonic wave, frequency etc.
The pretreated concrete intension of described difference normalization data and method as follows:
The parameter degeneration vector of degradation model comprises temperature, humidity, frequency and harmonic wave, adopt different sample frequency to cause data volume different when obtaining these data, such as humiture is that sampling should be carried out in every three minutes, and frequency, harmonic wave, ratio, angular difference sampling per minute (calculating) once, then temperature, wetness action amount and other amount between sampling rate differ 180 times; The radix of Various types of data and dimension are also inconsistent, and such as temperature radix is 25, and frequency radix is 50.First, in order to synchrodata, should unify the sampling rate of Various types of data.Secondly, the radix difference of each input quantity can affect the correctness of subsequent degradation network training result, and input quantity little for radix can be flooded by the input quantity that radix is large, needs to be normalized Various types of data, so that uniform variable weights and dimension.
Because epidemic disaster can not be undergone mutation, first-order linear method of interpolation therefore can be used to carry out data volume expansion to epidemic disaster data, to reach and other.Be treated to example with temperature data, suppose i-th and j point sampling temperature be T (i) and T (j), data volume needs to be increased to N doubly, then, after carrying out data extending, the data sequence between two temperature sampling values is:
T ( i + k ) = T ( i ) + T ( j ) - T ( i ) N · k k = 1 , 2 , ... , N - 1 T ( j ) k = N - - - ( 9 )
In like manner, after adopting (9) formula process to humidity, humiture data sampling rate is identical with its dependent variable.
Re-use difference normalization to process Various types of data, suppose Variables Sequence for V1, V2 ..., Vn, Vn+1}, its variable gradient is defined as
ΔV=max{ΔV i}(i=2,3,...,n,n+1)(10)
Then following difference normalized is carried out to Variables Sequence:
V i = V i + 1 - V i Δ V - - - ( 11 )
Thus, equation (7) is written as normalized form and is:
Δ ϵ → = B Δ α → - - - ( 12 )
In above formula, for difference effect vector, for difference degeneration parameter, B is corresponding difference degenerate network.
: the degenerate network of training with the normalization data of non-differential that, when carrying out case verification, the degenerate network of result display training gained correctly can not assess the error of metering system to the reason that data have first done difference processing before above-mentioned normalization.Further data characteristic discover, the rate of change of Various types of data is different with variable quantity, and the such as change such as temperature, humidity is slow and variable quantity is comparatively large, and frequency variation is then quite small.But obviously, the error effect of small frequency jitter to electric energy metered system is larger, and namely the action of little weights has the influence degree of large weights, therefore proposes to use difference normalization to carry out pre-service, amplify the weights of subtle change.
Concrete intension and the method for described many reference amounts degenerate network training are as follows:
The essence solving many reference amounts degenerate network is the matrix of coefficients B according to action and degeneration parameter solving equation (12), it is described that the effect of multi-factor comprehensive impact to degeneration parameter, but the element of matrix B does not have resolvability, can not describe with constant or elementary function.In this case, artificial intelligent learning method should be used to train degenerate network, the input-output characteristic of approaching to reality matrix of coefficients B.
Example of the present invention is for BP feedforward neural network.Feedforward neural network is the one of artificial neural network, and each neuron, from input layer, receives previous stage input, then is input to next stage, until output layer.BP neural network is the one of feedforward neural network, and the mode of learning of its network structure and two-way propagation makes this neural network have following features: 1) distributed information storage mode; 2) mode of learning parallel processing; 3) self study and adaptivity; 4) stronger robustness and fault-tolerance.
For the training of many reference amounts degenerate network, its learning network structural representation as shown in Figure 6.I, j, k be corresponding input layer, hidden layer, output layer respectively.For the degenerate network of training, be input as action herein export as amount of degradation the network of training is to the matrix of coefficients B in the change action approximant (12) of input.
Further, in order to the general implementation step of the appraisal procedure of error state described in the present invention is described, describe in detail with concrete example below.The present invention analyzes using the digital electric energy metrical on-line monitoring system fetched data being installed on certain 110kV transformer station as sample.Concrete analysis step is as follows:
(1) intercept the data of the morning 10 on April 7th, 2015 up to afternoon 5 time as analytic target, first use difference normalization to carry out pre-service to data.Action is for temperature, and amount of degradation is for electronic current mutual inductor ratio, and its pre-processed results is as shown in Fig. 7, Fig. 8, Fig. 9 and Figure 10.
Be illustrated in figure 7 temperature raw data, the temperature variation in a day is mild; Fig. 8 is the data that temperature carries out after difference normalized; Be illustrated in figure 9 ECT ratio raw data; Figure 10 is the data that ECT ratio carries out after difference normalized.As can be seen from Fig. 7, Fig. 8, Fig. 9 and Figure 10, after difference normalized, value, all between-1 to 1, is conducive to the correctness ensureing training degenerate network.
(2) use BP neural network to train to data, arranging training error is 0.01%, obtains degenerate network, and tests to degenerate network, and the training result obtained as shown in figure 11.
As seen from Figure 11, after carrying out 50 iteration, error is less than 0.01%, thinks that study terminates, then adopts cross check method to test to training result.
According to the degenerate network of above-mentioned training, in conjunction with amount of degradation initial value, assess digital electric energy metered system April 7 day afternoon 5 time after in short-term in error state, and analyze the data that on-line monitoring system obtains in the same period.Result as shown in figure 12.Correlation curve 1 and curve 2, can find out, based on the assessment result of many reference amounts degradation model, substantially meet with the actual running results trend, assessment errors is compared with actual error in short-term, and its absolute error is not more than 0.2%.
As follows to the deeper understanding of the present invention: to the digital electric energy metered system in a certain transformer station, final only needs arranges the sensor measuring degeneration parameter action, the environmentally sensitive unit 10 namely shown in Fig. 5, but kind of sensor is not limited to listed by Fig. 5; And without the need to arranging the sensor measuring digital electric energy metrical degeneration parameter, the circuit current sensing unit 11 namely shown in Fig. 5.By degeneration parameter action is inputted well-drilled degenerate network, digital electric energy metered system error state can be obtained.
Those skilled in the art will readily understand; the foregoing is only preferred embodiment of the present invention; not in order to limit the present invention, all any amendments done within the spirit and principles in the present invention, equivalent replacement and improvement etc., all should be included within protection scope of the present invention.

Claims (3)

1. the error evaluation system of a digitalized electric energy metering system, this error evaluation system is used for carrying out online error state assessment to digital electric energy metering device, and comprise environmentally sensitive unit, current/voltage sensing unit and CPU (central processing unit), it is characterized in that:
Described environmentally sensitive unit comprises temperature sensor, humidity sensor, the second microprocessor, and the second direct supply; Wherein this temperature sensor, humidity sensor are respectively used to perform temperature to environment residing for the digitalized electric energy metering system as monitoring target, humidity measures in real time, and give described second microprocessor by common transport, this second direct supply is then for providing working power to other component units in described;
Described current/voltage sensing unit comprises electromagnetic current transducer, electromagnetic potential transformer, first microprocessor, and is the first direct supply that in current/voltage sensing unit, each device is powered; Wherein said current transformer, described voltage transformer (VT) are used for measuring in real time the frequency of described digitalized electric energy metering system and harmonic wave;
In addition, described CPU (central processing unit) is connected with first, second microprocessor described simultaneously, and sets up in real time according to measurement data and upgrade many reference amounts degradation model, assesses digital electric energy metered system error state in short-term in current and future:
The integrated data analyzing and processing software of described computer run is as state estimation platform, (1) difference normalization preprocess method is adopted to carry out pre-service to described measurement data, (2) artificial neural network is adopted to train degenerate network to be specially again: the measurement data through difference normalized to be input to described degeneration grid, to try to achieve the degeneration grid that differential pair is answered; (3) last and future current according to the described digital electric energy metered system of described measurement data assessment of described each sensing unit again error state in short-term, thus realize the error state monitoring of digitalized electric energy measuring apparatus.
2. an error evaluation method for digitalized electric energy metering system, is characterized in that, this system comprises the following steps:
(1) under the prerequisite of uniform sampling rate, respectively the factor of the influential system error of described digitalized electric energy metering system and corresponding systematic error are measured; The factor of described influential system error is one or more in environment temperature, ambient humidity, electric parameter, frequency and harmonic wave, electromagnetic field, vibration, communication abnormality, decompression, cutout;
(2) difference normalization preprocess method is adopted to carry out pre-service to described measurement data;
(3) artificial neural network is adopted to train degenerate network again, adopt artificial neural network to carry out learning training, and verify training network, whether training of judgement result is reasonable, as unreasonable then retraining again, until validation error meets setting value;
(4) last again according to the described measurement data of described each sensing unit by the described digital electric energy metered system of described degenerate network assessment current and future error state in short-term, thus the error state realizing digitalized electric energy measuring apparatus is monitored.
3. appraisal procedure as claimed in claim 2, it is characterized in that, in described step (3), the measurement data through difference normalized is input to described degeneration grid, try to achieve the degeneration grid that differential pair is answered, assess the error of electric energy metered system with this.
CN201510458899.6A 2015-07-30 2015-07-30 A kind of error evaluation system and method for digitalized electric energy metering system Active CN105158723B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510458899.6A CN105158723B (en) 2015-07-30 2015-07-30 A kind of error evaluation system and method for digitalized electric energy metering system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510458899.6A CN105158723B (en) 2015-07-30 2015-07-30 A kind of error evaluation system and method for digitalized electric energy metering system

Publications (2)

Publication Number Publication Date
CN105158723A true CN105158723A (en) 2015-12-16
CN105158723B CN105158723B (en) 2019-02-15

Family

ID=54799641

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510458899.6A Active CN105158723B (en) 2015-07-30 2015-07-30 A kind of error evaluation system and method for digitalized electric energy metering system

Country Status (1)

Country Link
CN (1) CN105158723B (en)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107490778A (en) * 2017-08-25 2017-12-19 成都思晗科技股份有限公司 A kind of multifactor quantitative analysis method of error in dipping influence factor
CN107861089A (en) * 2017-09-26 2018-03-30 国网江西省电力公司电力科学研究院 A kind of intelligent electric energy meter method of testing for being adapted to exceptional operating conditions
CN108088469A (en) * 2017-11-14 2018-05-29 中国航空工业集团公司西安飞机设计研究所 A kind of long endurance airplane inertial navigation error compensation method
CN108445435A (en) * 2018-02-05 2018-08-24 国电南瑞科技股份有限公司 A kind of electric energy meter calibrating apparatus error online evaluation method
CN108693496A (en) * 2018-05-07 2018-10-23 国网山东省电力公司电力科学研究院 A kind of intelligent electric energy meter error predictor method based on parameter degeneration equation
CN109188341A (en) * 2018-09-18 2019-01-11 南方电网科学研究院有限责任公司 Live kinematic error appraisal procedure, device, equipment and the storage medium of electric energy meter
CN109597014A (en) * 2018-11-30 2019-04-09 国网上海市电力公司 A kind of electric energy meter error diagnostic method based on artificial intelligence technology
CN110658487A (en) * 2019-10-18 2020-01-07 国网福建省电力有限公司 Meter box and system capable of achieving intelligent electric meter error online estimation
CN110907885A (en) * 2019-12-06 2020-03-24 国网湖北省电力有限公司计量中心 Method and system for evaluating field operation state of digital electric energy metering system

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1172261A (en) * 1997-08-27 1998-02-04 武汉国测电器公司 Method and device for checking electric power mutual inductor
CN201069470Y (en) * 2007-07-20 2008-06-04 太原市优特奥科电子科技有限公司 Real time online monitoring device for integrated error of high-voltage power calculation device
CN102707256A (en) * 2012-06-20 2012-10-03 北京航空航天大学 Fault diagnosis method based on BP-Ada Boost nerve network for electric energy meter
CN202758062U (en) * 2012-07-05 2013-02-27 广东电网公司电力科学研究院 Overall metering error detection system of digital transformer station electric energy metering device
CN202794496U (en) * 2012-05-22 2013-03-13 鹤岗电业局 Real-time online monitoring instrument for power metering device
CN104076319A (en) * 2014-05-04 2014-10-01 贵州电力试验研究院 Online error analysis system of digitized electric energy metering device

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1172261A (en) * 1997-08-27 1998-02-04 武汉国测电器公司 Method and device for checking electric power mutual inductor
CN201069470Y (en) * 2007-07-20 2008-06-04 太原市优特奥科电子科技有限公司 Real time online monitoring device for integrated error of high-voltage power calculation device
CN202794496U (en) * 2012-05-22 2013-03-13 鹤岗电业局 Real-time online monitoring instrument for power metering device
CN102707256A (en) * 2012-06-20 2012-10-03 北京航空航天大学 Fault diagnosis method based on BP-Ada Boost nerve network for electric energy meter
CN202758062U (en) * 2012-07-05 2013-02-27 广东电网公司电力科学研究院 Overall metering error detection system of digital transformer station electric energy metering device
CN104076319A (en) * 2014-05-04 2014-10-01 贵州电力试验研究院 Online error analysis system of digitized electric energy metering device

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
陶锐: "人工神经网络在电能计量综合误差分析中的应用", 《内蒙古科技与经济》 *

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107490778A (en) * 2017-08-25 2017-12-19 成都思晗科技股份有限公司 A kind of multifactor quantitative analysis method of error in dipping influence factor
CN107490778B (en) * 2017-08-25 2019-07-02 成都思晗科技股份有限公司 A kind of multifactor quantitative analysis method of measurement error influence factor
CN107861089A (en) * 2017-09-26 2018-03-30 国网江西省电力公司电力科学研究院 A kind of intelligent electric energy meter method of testing for being adapted to exceptional operating conditions
CN108088469A (en) * 2017-11-14 2018-05-29 中国航空工业集团公司西安飞机设计研究所 A kind of long endurance airplane inertial navigation error compensation method
CN108445435A (en) * 2018-02-05 2018-08-24 国电南瑞科技股份有限公司 A kind of electric energy meter calibrating apparatus error online evaluation method
CN108445435B (en) * 2018-02-05 2020-08-25 国电南瑞科技股份有限公司 Online error evaluation method for electric energy meter calibrating device
CN108693496A (en) * 2018-05-07 2018-10-23 国网山东省电力公司电力科学研究院 A kind of intelligent electric energy meter error predictor method based on parameter degeneration equation
CN109188341A (en) * 2018-09-18 2019-01-11 南方电网科学研究院有限责任公司 Live kinematic error appraisal procedure, device, equipment and the storage medium of electric energy meter
CN109597014A (en) * 2018-11-30 2019-04-09 国网上海市电力公司 A kind of electric energy meter error diagnostic method based on artificial intelligence technology
CN110658487A (en) * 2019-10-18 2020-01-07 国网福建省电力有限公司 Meter box and system capable of achieving intelligent electric meter error online estimation
CN110907885A (en) * 2019-12-06 2020-03-24 国网湖北省电力有限公司计量中心 Method and system for evaluating field operation state of digital electric energy metering system
CN110907885B (en) * 2019-12-06 2021-10-01 国网湖北省电力有限公司计量中心 Method and system for evaluating field operation state of digital electric energy metering system

Also Published As

Publication number Publication date
CN105158723B (en) 2019-02-15

Similar Documents

Publication Publication Date Title
CN105158723A (en) Error evaluation system and method for digital electric energy metering system
Bolognani et al. Fast power system analysis via implicit linearization of the power flow manifold
CN103268519B (en) Based on the power-system short-term load forecasting method and the device that improve Lyapunov index
Zhai A new method for short-term load forecasting based on fractal interpretation and wavelet analysis
CN102377180B (en) Power system load modeling method based on electric energy quality monitoring system
CN103838959A (en) Method for applying partial least squares regression to power distribution network harmonic source positioning and detecting
CN104076319A (en) Online error analysis system of digitized electric energy metering device
CN106443285A (en) Multi-harmonic-source harmonic contribution quantitative analysis method based on total-least square method
CN102868160B (en) Macrozone load modeling method in intelligent power system
CN106786567B (en) A kind of online load modeling method based on PMU noise like data
Almutairi et al. Use of MCMC to incorporate a wind power model for the evaluation of generating capacity adequacy
CN107741578B (en) Original meter reading data processing method for remote calibration of running error of intelligent electric energy meter
CN105044649A (en) Current transformer error horizontal dynamic detection method and system
Zhang et al. Power system sensitivity identification—Inherent system properties and data quality
CN106599417A (en) Method for identifying urban power grid feeder load based on artificial neural network
CN105488335A (en) Lyapunov exponent based power system load prediction method and apparatus
Cui et al. HVDC transmission line fault localization base on RBF neural network with wavelet packet decomposition
Kaplan et al. A novel method based on Weibull distribution for short-term wind speed prediction
CN105259398A (en) Method for analyzing random error characteristic of optical fiber current transformer based on total variance
CN103412171B (en) A kind of mains by harmonics voltage signal detection method based on extreme learning machine
Wang et al. Using AMR data for load estimation for distribution system analysis
CN102545218A (en) Online load modeling parallel computing method based on electric energy quality monitoring system
CN104091047A (en) Traffic flow missing data estimation system and method based on traffic time-spatial information
CN106443253A (en) Power transmission line parameter identification method based on PMU (phasor measurement unit) data
Liu et al. Asynchronous harmonic analysis based on out-of-sequence measurement for large-scale residential power network

Legal Events

Date Code Title Description
PB01 Publication
C06 Publication
SE01 Entry into force of request for substantive examination
C10 Entry into substantive examination
GR01 Patent grant
GR01 Patent grant