CN103324840A - Power utilization quality comprehensive evaluation method for power demand side - Google Patents

Abstract

The invention discloses a power utilization quality comprehensive evaluation method for a power demand side. The power utilization quality comprehensive evaluation method comprises a first step of finishing data collection and transmission of power distribution network monitoring points of the power demand side; a second step of establishing power quality grades for power quality indexes, and performing unitization processing on the established grades according to benefit type indexes and cost type indexes; a third step of utilizing a Bayes method to correct the indexes on the basis of obtaining of subjective weights, and obtaining weight optimization values of various indexes through an optimization algorithm with constraints; a fourth step of performing grading evaluation on various power indexes of the monitoring points by means of an attribute recognition model; and a fifth step of obtaining comprehensive evaluation grades of power quality by calculation according to the weight optimization values, and finishing quantitative evaluation on the power quality. Reasonable calculation of the weights of various indexes of power is achieved through the power utilization quality comprehensive evaluation method, reasonability and reliability of power comprehensive evaluation results are improved, and reliable information can be provided for power grid safety.

Description

A kind of electric power demand side power quality comprehensive estimation method

Technical field

The present invention relates to a kind of electric power demand side power quality comprehensive estimation method, be specifically related to a kind of electric energy comprehensive estimation method based on Bayes and attribute Recognition Model, belong to the electrical engineering field.

Background technology

Electric energy is as the indispensable energy of society, and the quality of its quality is related to the safe operation of electrical network and electric power demand side user's electricity consumption level.The desirable quality of power supply is perfect sine wave, but owing to have various complicated factors in the electrical network, thereby cause the quality of power supply deviation to occur, so produced power quality problem.

Electricity quality evaluation is the correlation parameter according to electric energy, provides the quality grade of electric energy according to certain appraisal procedure, checks for power department and user.Traditional electricity quality evaluation mainly concentrates on to be assessed single index, has ignored the impact between each index of electric energy.The quality of power supply standard that China in 2008 has revised to have promulgated, this five indices has the supply voltage deviation, power system frequency deviation, imbalance of three-phase voltage degree, voltage fluctuation and flickering, utility network harmonic wave.In fact, the quality of power supply is a comprehensive concept, it is to be formed by a plurality of index common combinations, therefore, the single index assessment can not have been satisfied the demand of electric energy development, so the comprehensive assessment of electric energy occurred, its main thought is the weight of obtaining each index, thereby utilize weight as a whole a plurality of index comprehensives, realize the comprehensive assessment of user power utilization level.

Patent " a kind of method of electric energy quality synthesis evaluation " [application number: CN201110445808.7, publication number: CN101750561A], a kind of TOPSIS method of grey incidence coefficient matrix of using is disclosed to the method for electric energy quality synthesis evaluation, determine subjective and objective weight by AHP method and entropy power method, utilize the decision matrix of TOPSIS method, finally obtain quality of power supply grade by approach degree.There is artificial subjective factor in various degree in the method, calculates relatively complicated.

Patent " comprehensive assessment normalization " [number of patent application: CN201110051283.9, publication number: CN102339355A], a kind of method of electric energy quality synthesis evaluation normalized is disclosed, the every single index of the electric energy of actual measurement and calculating is carried out normalized, obtain every metewand, then calculate with weights, at last in conjunction with the quality of power supply score situation table of comparisons, judge the general status of the quality of power supply.Also there is artificial subjective factor in various degree in the method, and assessment result has certain uncertainty.

The present invention is directed to above-mentioned deficiency, a kind of electric energy appraisal procedure of the electric power demand side in conjunction with bayesian algorithm and Attribute Recognition is proposed, utilize the subjective weight of Bayes's enabling legislation correction, and then obtain weighted value under each grade of the quality of power supply, and utilize the constrained optimization algorithm, and obtain the optimal value of weight, realize the subjective and objective unification of weight, be affected by human factors when having avoided weight to determine excessive, thereby so that the assessment the result objective, credible.Relend and help attribute Recognition Model, realized final quality of power supply ranking, better to satisfy the demand of Modern Electric Power Quality comprehensive assessment.

Summary of the invention

The invention provides a kind of electric power demand side power quality comprehensive estimation method based on Bayes and Attribute Recognition Theory, can realize the reasonable computation of each index weights of electric energy, improve electric energy comprehensive assessment result's rationality and reliability, for the security of electrical network provides authentic communication, and satisfy the demand that the electric power demand side user knows electric electricity consumption.

According to purpose of the present invention, a kind of electric power demand side electricity quality evaluation system is proposed, it comprises an electric power demand side power distribution network, be installed in the electrical energy parameter monitor of some monitoring points, one is used for transmitting the GPRS network of electrical energy parameter monitor Monitoring Data, the database server of one storage of electrical energy assessment data, one for assessment of the monitoring and evaluation center of calculating.

Wherein, the power distribution network of electric power demand side is arbitrarily power consumer distribution point;

Wherein, the electrical energy parameter monitor mainly detects and the power quality index that calculates the electric power demand side user, comprising: voltage deviation, voltage fluctuation, three-phase imbalance, frequency departure, harmonic content, power off time etc.;

Wherein, GPRS network is for being used for a kind of mobile data services of mobile phone user, combining wireless communication and Internet, the wireless transmission of realization power quality index data;

Wherein, database server is used for all power quality index data of reception and the transmission of storage of electrical energy parameter monitor;

Wherein, the monitoring and evaluation center is used for the chart demonstration of power quality index data and the calculating of assessment result.

Comprehensive estimation method of the present invention comprises the steps:

(1) finishes data acquisition and the transmission of electric power demand side power distribution network monitoring point;

(2) power quality index is set up quality of power supply grade, and according to " benefit type " index and " cost type " index the grade of setting up is carried out normalized, wherein " cost type " index is namely: the property value little better index that heals, " benefit type " index namely: the larger index better of property value;

(3) obtaining on the basis of subjective weight, utilizing bayes method that it is revised, and obtaining the weight optimization value of each index with the constrained optimization algorithm;

(4) by means of attribute Recognition Model, every power index of monitoring point is carried out classified estimation;

(5) in conjunction with the weight optimization value, calculate the comprehensive evaluation grade of the quality of power supply, thereby finish the qualitative assessment to the quality of power supply.

The present invention's useful achievement compared with the prior art is: the present invention takes into full account the relation between each index of the quality of power supply, utilize bayesian theory, subjective weight is revised, rather than simply subjective and objective weight is made up, thereby the index weights that obtains more meets the practice situation, thereby makes assessment result more reasonable.The utilization of attribute Recognition Model is conducive to obtain fast the evaluation grade value, makes assessment result more vivid objective.

Description of drawings

Fig. 1 is the Organization Chart of electric power demand side power quality integrated estimation system;

Fig. 2 is based on the schematic flow sheet of the energy quality comprehensive assessment method of Bayes and Attribute Recognition Theory.

Embodiment

See also the 1st figure, it is used for the Organization Chart of electric power demand side power quality integrated estimation system for the present invention, as shown in the figure, the present invention includes an electric power demand side power distribution network, be installed in the electrical energy parameter monitor of some monitoring points, one is used for transmitting the GPRS network of electrical energy parameter monitor Monitoring Data, the database server of a storage of electrical energy quality index data, and one for assessment of the monitoring and evaluation center of calculating.

Wherein, the electrical energy parameter monitor mainly detects the power quality index data of each monitoring point of electric power demand side power distribution network, comprising: voltage deviation, voltage fluctuation, three-phase imbalance, frequency departure, harmonic content, power off time etc.;

Wherein, whole computation processes of comprehensive estimation method proposed by the invention are finished at the monitoring and evaluation center, detailed step as shown in Figure 2:

The 1st step: Criterion index value matrix.Make up each the power index value matrix Y under five grades:

$Y={\left(y_{i,j}\right)}_{c\×m}=\left[\begin{array}{cccc}{y}_{\mathrm{1,1}}& y_{\mathrm{1,2}}& \·\·\·& y_{1,m}\\ y_{\mathrm{2,1}}& y_{\mathrm{2,2}}& \·\·\·& y_{2,m}\\ \·& \·& \·& \·\\ \·& \·& \·& \·\\ \·& \·& \·& \·\\ y_{c,1}& y_{c,2}& \·\·\·& y_{c,m}\end{array}\right]---\left(1\right)$

Wherein, m is the power quality index number, and c is evaluation grade number (c=5 here), y _{I, j}Power quality index data for the detection of electrical energy parameter monitor.Again each achievement data is carried out standardization, be normalized to same dimension, obtain the standard index value matrix

Its standardization processing method is:

$r_{i,j}=\left\{\begin{array}{c}\frac{\min {\left\{y_{i,j}\right\}}_{i=1}^c}{y_{i,j}},y_{i,j}\∈R\\ \frac{y_{i,j}}{\max {\left\{y_{i,j}\right\}}_{i=1}^c},y_{i,j}\∈A\end{array}\right.---\left(2\right)$

In the formula, R is " cost type " index, that is: the property value little better index that heals; A is " benefit type " index, that is: the larger index better of property value.The power quality index that relates to has voltage deviation, frequency departure, harmonic content, three imbalances, voltage fluctuation, these six indexs of power off time;

The 2nd step: bayes method is determined weight.Determining the subjective weight of index under each grade

The basis on, adopt bayes method, with the standard index value under each grade subjective weight is revised.If subjective weight is u ₁, u ₂..., u _m, i.e. index I _jProbability be p (I _j)=u _j, and at index I _jLower, grade G _iThe probability that occurs is p (G _i| I _j)=r _{I, j}, utilize bayes method, at grade G _iLower, the weight of j index (posterior probability) is:

$p\left(I_j\right|G_i)=\frac{p\left(I_j\right)p\left(G_i\right|I_j)}{{\mathrm{\Σ}}_{j=1}^{m}p\left(I_j\right)p\left(G_i\right|I_j)}=\frac{{\mathrm{\μ}}_j{r}_{i,j}}{{\mathrm{\Σ}}_{j=1}^m{\mathrm{\μ}}_j{r}_{i,j}}=w_j^{\left(i\right)}---\left(3\right)$

The establishing target Optimized model:

$J=\min \left(\underset{i=1}{\overset{c}{\mathrm{\Σ}}}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{(w_j-w_j^{\left(i\right)})}^2{r}_{i,j}^2\right)$

$s.t.\left\{\begin{array}{c}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{w}_j=1\\ 1>w_j>0\end{array}\right.$

In the formula,

Be subjective weight, w _jBe the weight modified value, m is the power quality index number, and c is the evaluation grade number, r _{I, j}Element for the standard index value matrix.

Utilize the nonlinear iteration optimizing algorithm to obtain: the whole weight of the quality of power supply

w _j∈ (0,1).

The 3rd step: set up attribute measure matrix.If the m item achievement data of certain monitoring point is

Be cutting apart in order of certain generic attribute space of X, and satisfy q ₁Q ₂... q _c, can write out the criteria for classification matrix by the categorised demarcation line of each index and be $f=\left[\begin{array}{cccc}{a}_{\mathrm{1,1}}& a_{\mathrm{1,2}}& \·\·\·& a_{1,c}\\ a_{\mathrm{2,1}}& a_{\mathrm{2,2}}& \·\·\·& a_{2,c}\\ \·& \·& \·& \·\\ \·& \·& \·& \·\\ \·& \·& \·& \·\\ a_{m,1}& a_{m,2}& \·\·\·& a_{m,c}\end{array}\right],$ If u _{I, j}Can be expressed as x _iBelong to attribute q _jQuantitative description (that is: x _i∈ q _j) Attribute Measure, then the Attribute Measure of monitoring point X is U={u _{I, j}, i=1,2 ..., m; J=1,2 ..., c:

(1) works as x _i≤ a _{I, 1}The time,

u _i,1=1，u _i,2=…=u _i,c=0 （5）

(2) work as x _iA _{I, c}The time,

u _i,c=1，u _i,1=…=u _i,c-1=0 （6）

(3) work as a _{I, j}＜x _i≤ a _{I, j+1}The time,

$u_{i,j}=\frac{|x_i-a_{i,j+1}|}{|a_{i,j}-a_{i,j+1}|},u_{i,j+1}=\frac{|x_i-a_{i,j}|}{|a_{i,j}-a_{i,j+1}|},---\left(7\right)$

The 4th step: ask quality of power supply grade in conjunction with weight and Attribute Measure.The overall target of monitoring point X belongs to the (v for V=that estimates of each generic attribute ₁, v ₂..., v _c),

$v_j=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{w}_i\·u_{i,j}---\left(8\right)$

Thus quality of power supply grade finally is $k_0=\min \{k:\underset{j=1}{\overset{k}{\mathrm{\Σ}}}{v}_j\≥\mathrm{\λ},1\≤k\≤c\}.$

Grade point k ₀Less, represent the electric energy deviation less, the quality of power supply is better.In the comprehensive assessment of electric power demand side power quality, by the quality of power supply is carried out classification, and by calculating the residing grade of comparative assessment result, can hold intuitively the quality condition of electric energy.Be in the situation of same grade for some result, can be estimated above the degree of Reliability Code by its synthesized attribute and judge, thereby can compare in the same grade situation, which quality of power supply is better.

Claims (3)

1. electric power demand side power quality comprehensive estimation method, the system of its assessment comprises an electric power demand side power distribution network, is installed in the electrical energy parameter monitor of some monitoring points, the database server and of a GPRS network that is used for transmitting electrical energy parameter monitor Monitoring Data, a storage of electrical energy assessment data is for assessment of the monitoring and evaluation center of calculating, wherein, the electric power demand side power distribution network is arbitrarily power consumer distribution point; The electrical energy parameter monitor is used for detecting and calculating electric power demand side user's power quality index; The communication of GPRS network combining wireless and Internet, the wireless transmission of realization power quality index data; Database server is used for all power quality index data of reception and the transmission of storage of electrical energy parameter monitor; The monitoring and evaluation center is used for the chart demonstration of power quality index data and the calculating of assessment result; It is characterized in that, comprehensive estimation method comprises the steps:

(1) described system finishes data acquisition and the transmission of electric power demand side power distribution network monitoring point;

(2) power quality index is set up quality of power supply grade, and according to " benefit type " index and " cost type " index the grade of setting up is carried out normalized, wherein " cost type " index is namely: the property value little better index that heals, " benefit type " index namely: the larger index better of property value;

(3) obtaining on the basis of subjective weight, utilizing bayes method that it is revised, and obtaining the weight optimization value of each index with the constrained optimization algorithm;

(4) by means of attribute Recognition Model, every power index of monitoring point is carried out classified estimation;

(5) in conjunction with the weight optimization value, calculate the comprehensive evaluation grade of the quality of power supply, thereby finish the qualitative assessment to the quality of power supply.

2. a kind of electric power demand side power quality comprehensive estimation method according to claim 1 is characterized in that, described power quality index comprises voltage deviation, voltage fluctuation, three-phase imbalance, frequency departure, harmonic content and power off time.

3. a kind of electric power demand side power quality comprehensive estimation method according to claim 1 and 2, it is characterized in that, the weight modification method of described step (3) is the iteration optimizing, i.e. the objective function of constitution optimization, and solution corresponding to its minimum value is the weight modified value:

$J=\min \left(\underset{i=1}{\overset{c}{\mathrm{\Σ}}}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{(w_j-w_j^{\left(i\right)})}^2{r}_{i,j}^2\right)$

$s.t.\left\{\begin{array}{c}\underset{j=1}{\overset{m}{\mathrm{\Σ}}}{w}_j=1\\ 1>w_j>0\end{array}\right.$

In the formula,

Be subjective weight, w _jBe the weight modified value, m is the power quality index number, and c is the evaluation grade number, r _{I, j}Element for the standard index value matrix.

Images