CN103778574A - Method for evaluating development coordination of island microgrid - Google Patents

Abstract

The invention discloses a method for evaluating the development coordination of an island microgrid. The method comprises the following contents: 1) analysing from the four aspects of distributed power side, load side, national economy and environment of the island microgrid to construct a development coordination evaluation index system for the island microgrid; 2) carrying developed illustration on coordination indexes with island microgrid characteristics, and establishing a corresponding calculation model; 3) improving a rough set theory by virtue of a multi-interval partition method; 4) determining an index weight matrix by virtue of the improved rough set theory; 5) fusing the weight matrix by virtue of a robust regression algorithm, and determining index weight values; 6) analysing the influence of each index in the island microgrid and the conditions which need to be met during planning and construction according to each index weight, and outputting a development coordination evaluation report for the island microgrid. The evaluation system and evaluation method disclosed by the invention are capable of scientifically and objectively evaluating the development coordination of island microgrids with different structures. Therefore, scientific guidance basis can be provided for the planning and construction for the island microgrids.

Description

The micro-power network development harmony of isolated island evaluation method

Technical field

The present invention relates to micro electric network coordination index system establishment and evaluation method, particularly a kind of isolated island micro-power network development harmony index system and evaluation method.

Background technology

In the face of the challenge of fossil energy consumption and ecological deterioration, Mini Diesel Engine, wind-force, photovoltaic generation distributed generation mode find broad application.But this distributed power generation is subject to the distribution of regenerative resource and the impact of geographical environment.For flexible management and this generation mode of control, scholars have proposed micro-electrical network concept.Micro-electrical network refers to the small-sized electric system of being transported to being made up of distributed power source (DG), energy storage device, relevant load and power electronic equipment.This system can with large electrical network parallel running, also can islet operation.

The micro-electrical network of isolated island (being hereinafter called for short lonely net) has power distribution inequality, generating is intermittent and be subject to load to affect the features such as larger.This has greatly increased electric energy scheduling, has guaranteed the difficulty of power supply reliability and power supply capacity, has also increased the complicacy of Generation Side and load side coordinated operation.And domestic the research of power network development harmony is mainly concentrated on to power distribution network and the large electrical network of tradition.But these indexs can not directly apply in lonely net Development and Harmony Journal of Sex Research.And evaluation method is also based on single objective and subjective synthetic approach, fails to consider subjective and objective factor, and combination enabling legislation can not reduce the impact of exceptional value effectively.Therefore, instruct not good to lonely net Development and Harmony evaluation and planning construction.For correct reliability and the security of evaluating lonely net current situation, improving power supply, also help optimized network structure, improve lonely net economic benefit and environmental benefit, improve lonely net and city planning.Therefore be, vital to furtheing investigate lonely net Development and Harmony appraisement system and method.

Summary of the invention

To achieve these goals, the invention provides a kind of lonely net Development and Harmony evaluation method, can make science, objective appraisal to the lonely net Development and Harmony of different structure, therefore, the guidance foundation of science is provided can to the planning construction of lonely net.

The lonely net Development and Harmony of one provided by the invention evaluation method, its comprehensive evaluation step is as follows:

(1) in conjunction with the feature of existing lonely net, the factor of the lonely net Development and Harmony of analyzing influence, has creatively proposed lonely net Development and Harmony assessment indicator system;

(2), according to the characteristic of orphan's net, launch the computation model of the each index of explanation;

(3) utilize the many partitionings of opinion rating to improve traditional rough set theory, determine index weights matrix;

(4) utilize robustness regression algorithm fusion weight matrix, build rough set robustness regression evaluation method, thereby determine each index weights;

(5), according to each index weights, by analyzing the impact of indices on lonely net Development and Harmony, improve power supply reliability, economy and the power supply quality etc. of lonely net.

One-level evaluation index in described lonely net Development and Harmony assessment indicator system, it comprises distributed power source degrees of coordination, Coordinated degree, national economy degrees of coordination and environmental harmony degree.

Described lonely net Development and Harmony evaluation method, its first class index is decomposed into two-level index, and it is described in detail as follows:

(1) the distributed power source degrees of coordination index of lonely net comprises power supply capacity, power supply reliability and the power supply economics of batch (-type) new forms of energy;

(2) power supply reliability comprises system System average interruption frequency, Suo Xie SAIF index, system System average interruption duration, Suo Xie SAID index and expected loss of energy.

(3) power supply economics comprises the sale of electricity income and the national economy loss that cause due to mains supply deficiency;

(4) the Coordinated degree index of lonely net comprises rate of load condensate, voltage fluctuation rate and variation rate;

(5) the national economy degrees of coordination index of lonely net comprises electricity consumption elasticity coefficient and per GDP power consumption;

(6) the environmental harmony degree index of lonely net comprises clean energy resource occupation rate and waste discharge amount.

Wherein, in step (3), described rough set theory is determined weight matrix, and its calculation procedure is as follows:

(1) described lonely net is carried out to analysis of Influential Factors, extract key index;

(2) Index Establishment of described extraction is there is to the computation model of lonely net feature;

(3) utilize computation model to calculate the initial value of evaluation index;

(4) carry out data processing according to percentage method for making, convert property value to percentage value;

(5) utilize the many partitionings of opinion rating to improve rough set theory, build different index opinion rating tables;

(6) according to different index opinion rating tables, build different Criteria Decision Making tables;

(7) utilize rough set theory parameter weight matrix;

Wherein, in step (4), described robustness regression algorithm is determined index weights, and its calculation procedure is as follows:

(1) select suitable parameter value;

(2) utilize least square method to solve initial expectation value;

(3) solve new residual values according to initial expectation value;

(4) draw residual error weight according to residual values;

(5) expectation value making new advances according to residual error weight calculation;

(6) judge that adjacent twice expectation value error whether in tolerance interval, if so, finishes to calculate, choosing this expectation value is ideal value, otherwise, return to step (3).

The present invention is directed to the feature that the large power network development harmony indicator evaluation system of existing power distribution network and tradition can not be directly applied for lonely net, take into full account the feature of lonely net, consider from many factors such as its distributed electrical source, load side, national economy and environment.Build the Development and Harmony assessment indicator system with lonely net characteristic, and indices has been carried out launching explanation, made it more clear understandable.For the deficiency of existing micro-electrical network evaluation method, adopt different interval values to choose method traditional rough set theory is improved.Make up subjectivity and the unicity of traditional rough set theory in the time determining weight, evaluation result has been had more comprehensive and scientific.Meanwhile, adopt robustness regression algorithm to merge rough set theory weight matrix, it not only can reduce the impact of exceptional value, and can guarantee accuracy and the rationality of weight.Therefore, this evaluation method provides science to instruct reliably foundation can to the development construction of lonely net.

Accompanying drawing explanation

In the present invention, for making user more clear understandable, many places adopt graphic structure, and its concrete meaning is described as follows:

Fig. 1 is the general flowchart of lonely net Development and Harmony evaluation method;

Fig. 2 is lonely net Development and Harmony index system;

Fig. 3 is that lonely net Development and Harmony is evaluated general flowchart;

Fig. 4 is lonely net Development and Harmony evaluation method module map.

Embodiment

For the present invention can extensively be run in engineering practice, can be accepted and adopt by most scholars.Below in conjunction with accompanying drawing, the Some Key Technologies scheme in the present invention is clearly and completely described.

STEP1: lonely net Development and Harmony is evaluated general flowchart as shown in Figure 1.

STEP2: according to the feature of orphan's net, analyze lonely net Development and Harmony influence factor, determine micro-lonely net Development and Harmony index system, see Fig. 2.

STEP3: launching explanation to having the index of lonely net feature, be mainly wherein distributed power source degrees of coordination, Coordinated degree index, and national economy degrees of coordination and the explanation of environmental harmony degree index please refer to power distribution network.

1) power supply capacity

Power supply capacity is to weigh a key index of lonely net its construction level and distributed power source coordination generating capacity.Also refer to the ability that in lonely net, power-supply unit maximum can be met consumers' demand.Its computation model is:

$C=\underset{\mathrm{DS}=1}{\overset{n}{\mathrm{\Σ}}}{\left(\mathrm{S\η\θ}\right)}_{\mathrm{DS}}+\underset{\mathrm{WT}=1}{\overset{n}{\mathrm{\Σ}}}{\left(\mathrm{S\η\θ}\right)}_{\mathrm{WT}}+\underset{\mathrm{PV}=1}{\overset{n}{\mathrm{\Σ}}}{\left(\mathrm{S\η\θ}\right)}_{\mathrm{PV}}+\underset{\mathrm{FC}=1}{\overset{n}{\mathrm{\Σ}}}{\left(\mathrm{S\η\θ}\right)}_{\mathrm{FC}}---\left(1\right)$

In formula: C is power supply capacity, n is power-supply unit number of units, and S, η, θ are respectively capacity, utilization factor and the power factor of power equipment.DS, WT, PV, FC are respectively miniature oil-burning machine, wind-force, photovoltaic generator and energy-storage battery.

2) power supply economics

Power supply economics refers to the economic loss can not coordinated operation causing due to distributed power source.The loss of outage that mainly comprises the electric energy loss in distributed power source when generating and cause due to electricity shortage or fault.Wherein loss of outage is again by the electrical network electricity sales amount loss directly causing with because the indirect economic loss bringing to user that has a power failure forms.Its computation model is:

E=α(△P+△S)+△Sβ (2)

In formula: E is economic loss, α is annual sale of electricity average unit price, and △ P, △ S are respectively electric energy loss and the power failure sale of electricity loss amount of generating set, and β is unit loss of outage expense.

3) power supply reliability

Power supply reliability is to weigh the important indicator of operation of power networks level.Power supply reliability computation model is in the present invention consistent with the large electrical network of tradition, mainly from system have a power failure every year frequency, average annual interruption duration and average annual electric weight is not enough considers.

4) voltage fluctuation rate

Voltage fluctuation rate refers within the unit interval because the variation of load causes that supply voltage exceeds the number of times of normal power supply wish degree.Its computation model is:

$\mathrm{\γ}=\frac{N}{T}---\left(3\right)$

In formula: γ, N, T are respectively voltage fluctuation rate, fluctuation number of times and unit interval section.

5) variation rate

Variation rate in the present invention refers to because the variation of load causes.Computation model is:

$\mathrm{\δ}=\frac{U_{\max }-U_{\min }}{U_B}\×100\%---\left(4\right)$

In formula: δ is variation rate, U _max, U _min, U _bbe respectively voltage max, minimum value and ratings.

STEP4: lonely network parameters is set.

1) various distributed generator types are as shown in table 1.

Table 1 distributed generator type

Note: in table, CHP is writing a Chinese character in simplified form of cogeneration of heat and power (combined and power)

2) various energy storage device types and acquisition cost parameter are as shown in table 2.

Table 2 energy storage device type and acquisition cost

Energy storage type	Yearization unit quantity of electricity cost/(＄/(Kw.h))
		Lead-acid battery	25
Lithium ion battery	120
		Sodium-sulphur battery	85
Superconduction magnetic energy	200
		Super carbon level capacitor	85
Low-speed flywheel system	40
		High speed flywheel system	80

Note: in table, data are only reference, according to different producers, energy storage device needs different unit quantity of electricity costs.

3) the fuel cost parameter of various generators is as shown in table 3.

The various fuel cost parameter value of table 3 table

Parameter	Numerical value
		Steam coal price/(unit/t)	700
Rock gas calorific capacity/(KJ/m3)	35160
		Gas Prices/(unit/m3)	1.4
Caloric value/(unit/GJ)	32
		The thermal efficiency/% of CHP	80
Confession ratio of specific heat/% of CHP	35

Note: in table, data are only reference, and according to different areas, fuel cost is not quite similar.

4) the waste discharge amount parameter list of distributed generator is as shown in table 4.

The waste discharge amount data g/ (kW.h) of the various distributed generators of table 4

STEP5: achievement data processing.

Evaluation index can be divided quantitative target and qualitative index according to attribute; Wherein quantitative target refers to and can directly measure, and qualitative index can not directly be measured, and can only infer according to previous experiences or partial data.Therefore,, in evaluation problem, wish that as much as possible qualitative index being changed into quantitative target calculates.Can be divided into very big type, minimal type and ideal value type according to type; Greatly type refers to that index parameter is the bigger the better, and minimal type refers to that index parameter is the smaller the better, and it is better that ideal type refers to that index parameter more approaches certain value.

Although and quantitative target can directly be measured data, comprise discrete data and continuous data.And rough set theory is only applicable to discrete data, therefore, carrying out before index weights determines, need to carrying out the discrete processes of data.In order to eliminate the impact of dimension, the present invention adopts percentage method for making to carry out nondimensionalization processing to data simultaneously.Its processing procedure is as follows:

In the time that evaluation index is minimal type:

$y_{\mathrm{ij}}=\frac{x_{i\max }-x_{\mathrm{ij}}}{x_{i\max }-x_{i\min }}\×100\%---\left(5\right)$

In formula: x _imax, x _iminbe respectively maximum, the minimum value of evaluation index, x _ijrepresent the property value of i object under j index, y _ijrepresent that property value converts the fractional value after centesimal system to.

In the time that evaluation index is very big type:

$y_{\mathrm{ij}}=\frac{x_{\mathrm{ij}}-x_{i\min }}{x_{i\max }-x_{i\min }}\×100\%---\left(6\right)$

In the time that evaluation index will be stabilized in certain ideal value:

$y_{\mathrm{ij}}=\left\{\begin{array}{cc}\frac{x_{\mathrm{ij}}-x_{i\min }}{x_i-x_{i\min }}\&times;100\%,& x_{\mathrm{ij}}<x_i\\ \frac{x_{i\max }-x_{\mathrm{ij}}}{x_{i\max }-x_i}\&times;100\%,& x_{\mathrm{ij}}\&GreaterEqual;x_i\end{array}\right.---\left(7\right)$

In formula: x _iit is the ideal value of i index.

STEP6: rough set theory is determined index weights matrix

This patent utilizes between multi-region partitioning to determine that to traditional rough set theory weight method improves, and reduces the subjectivity of traditional rough set theory in the time setting up index opinion rating.

Rough set (Rough Sets) theory is that Polish scholar Pawlak teaches a kind of theoretical method of studying imperfect, uncertain knowledge and data proposing the eighties in 20th century.Keeping, under the prerequisite that classification capacity is constant, carrying out the yojan of knowledge, filter out key index, build evaluation rank and decision table, thereby draw the importance degree of each index and the weight of index.But rough set has certain subjectivity in the time that data are carried out to grade classification, and different grade classification also has certain influence to weight, and go back now the neither one criteria for classifying.Therefore adopt the combination of many group different demarcations to reduce the impact of this subjectivity herein.It is as follows that it calculates weight step:

1) achievement data processing

2) set up evaluation rank collection

Index opinion rating determines that method is as follows:

$x=\frac{y_{\mathrm{ij}}-60}{L}---\left(8\right)$

i=[x]+1 （9）

In formula: y _ijrepresent index percentage value.I represents index grade, and [x] represents to be not more than the maximum integer of x, and L is grade classification interval value.

3) set up Criteria Decision Making table

Set up Criteria Decision Making table according to the opinion rating of index, as shown in table 5.

Table 5 rough set Criteria Decision Making table

{ x in table 5 ₁l x _iexpression evaluation object collection, { C ₁l C _iexpression index set, v _ijrepresent the grade of j index of i object.

4) index weight is determined

If S={U, A, V, F} is an infosystem, respectively to property set A and certain attribute A of removal _iequivalence class divide expression formula be:

U/Ind(A)={X ₁L X _n} （10）

U/Ind(A-A _i)={Y ₁L Y _n} （11）

A _iimportance degree may be defined as:

$I(A-A_i)=\frac{1}{{\left|U\right|}^2}(\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\left|Y_i\right|}^2-\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\left|X_i\right|}^2)---\left(12\right)$

In formula: | X _i|, | Y _i| representation class X _i, Y _iradix, I (A-A _i) expression index A _iimportance degree.

4) index weights is determined

Index weight is normalized and gets final product to such an extent that weighted value is:

$w_i=\frac{r_i}{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{r}_i}---\left(13\right)$

In formula: w _ibe the weight of i index, r _irepresent the importance degree of i index.

STEP7: robustness regression merges weight

Robustness regression algorithm is to utilize Huber estimation function as influence function, linear regression algorithm to be improved, thereby reduces the impact of exceptional value, improves the precision of stationary value.Based on this feature, robustness regression algorithm can reduce rough set theory determining when weight the impact bringing owing to choosing different interval values, makes weighted value more realistic.

Traditional linear regression is to utilize differential solving method to make residual sum of squares (RSS) Q reach minimum value.

$Q=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{(x_i-x)}^2---\left(14\right)$

In formula: x _i, x is respectively actual value and expectation value, for making formula (14) reach minimum, the solving equation for the treatment of above can be turned to general form, as follows:

$\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\mathrm{\&psi;}\left(\frac{e_i}{\mathrm{\&delta;}}\right)=0---\left(15\right)$

In formula: ψ is sane estimation function, e _i, δ is respectively residual error, overall mean square deviation, wherein:

δ=1.5med(e _i) （16）

$\mathrm{\&psi;}\left(\frac{e_i}{\mathrm{\&delta;}}\right)=\mathrm{\&psi;}\left(e_i^{*}\right)=\left\{\begin{array}{cc}-\mathrm{\&beta;},& e_i^{*}<-\mathrm{\&beta;}\\ e_i^{*},& \left|e_i^{*}\right|\&le;\mathrm{\&beta;}\\ \mathrm{\&beta;},& e_i^{*}>\mathrm{\&beta;}\end{array}\right.---\left(17\right)$

In formula:

for standard variance, select suitable β, can effectively reduce like this | e _i ^*| the impact of > β exceptional value, thus make regression equation there is robustness, improve accuracy.

Residual error weighted value is:

$w_i=\left\{\begin{array}{cc}\frac{\mathrm{\&beta;}}{{e_i}^{*}}& \left|{e_i}^{*}\right|>\mathrm{\&beta;}\\ 1& \left|{e_i}^{*}\right|\&le;\mathrm{\&beta;}\end{array}\right.---\left(18\right)$

Formula (17) can change into:

$\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{w}_i\&CenterDot;{e_i}^{*}=0---\left(19\right)$

Namely weighted least-squares method of formula (19), is equivalent to solve solution while reaching minimum value.

Expectation value is judged:

x _i-x _(i-1)<ε （20）

ε is maximum error limit value, if above formula establishment, x _ifor final robustness regression value.Otherwise, continue circulation until meet above formula.

To sum up, it is as follows that robustness regression merges weight solution procedure:

1) select suitable β, ε value;

2) utilize least square method to solve initial expectation value x ₀;

3) according to x ₀the residual values e that calculating makes new advances and δ;

4) calculate residual error weight w according to formula (18) _i;

5) judge formula x _i-x _(i-1)whether < ε sets up, if set up, finishes to calculate x _ifor robustness regression value; Otherwise return to (2), continue to calculate, until equation meets.

STEP8: analyze lonely net Development and Harmony

After determining index weights according to lonely net Development and Harmony index system and the evaluation method based on rough set robustness regression algorithm, according to index weights, analyze the role and influence of each index to lonely net Development and Harmony.Determine the principal element that lonely net should be considered in planning construction, thereby improve the performance index such as lonely economy of netting, power supply reliability, the quality of power supply.Therefore, the present invention provides science to instruct reliably foundation can to lonely net planning construction.

Those skilled in the art can improve on basis of the present invention.If but within improvement of the present invention still belongs to the scope of the claims in the present invention and equivalent technologies thereof, still within protection domain of the present invention.

Claims (6)

1. the micro-power network development harmony of an isolated island evaluation method, is characterized in that, comprises the steps:

(1) in conjunction with the feature of the micro-electrical network of existing isolated island, the factor of the micro-power network development harmony of analyzing influence isolated island, has creatively proposed the micro-power network development harmony of isolated island assessment indicator system;

(2), according to the characteristic of the micro-electrical network of isolated island, launch the computation model of the each index of explanation;

(3) by the many partitionings of opinion rating, traditional rough set theory is improved, determine index weights matrix;

(4) utilize robustness regression algorithm fusion weight matrix, build rough set robustness regression evaluation method, thereby determine each index weights;

(5), according to each index weights, by analyzing the impact of indices on the micro-power network development harmony of isolated island, improve power supply reliability, economy and the power supply quality etc. of the micro-electrical network of isolated island.

2. the micro-power network development harmony of isolated island according to claim 1 evaluation method, it is characterized in that, one-level evaluation index in the micro-power network development harmony of described isolated island assessment indicator system, it comprises distributed power source degrees of coordination, Coordinated degree, national economy degrees of coordination and environmental harmony degree.

3. the micro-power network development harmony of isolated island according to claim 2 evaluation method, is characterized in that, all first class index are decomposed into two-level index, and it is described in detail as follows:

(1) the distributed power source degrees of coordination index of the micro-electrical network of isolated island comprises power supply capacity, power supply reliability and the power supply economics of batch (-type) new forms of energy;

(2) power supply reliability comprises system System average interruption frequency, Suo Xie SAIF index, system System average interruption duration, Suo Xie SAID index and expected loss of energy;

(3) power supply economics comprises the sale of electricity income and the national economy loss that cause due to mains supply deficiency;

(4) the Coordinated degree index of the micro-electrical network of isolated island comprises rate of load condensate, voltage fluctuation rate and variation rate;

(5) the national economy degrees of coordination index of the micro-electrical network of isolated island comprises electricity consumption elasticity coefficient and per GDP power consumption;

(6) the environmental harmony degree index of the micro-electrical network of isolated island comprises clean energy resource occupation rate and waste discharge amount.

4. according to claim 1 the micro-power network development harmony of isolated island evaluation method be is characterized in that, wherein, in step (3), described rough set theory is determined weight matrix, and its calculation procedure is as follows:

(1) the micro-electrical network of described isolated island is carried out to analysis of Influential Factors, extract key index;

(2) Index Establishment of described extraction is there is to the computation model of the micro-electrical network feature of isolated island;

(3) utilize computation model to calculate the initial value of evaluation index;

(4) carry out data processing according to percentage method for making, convert property value to percentage value;

(5) utilize the many partitionings of opinion rating to improve rough set theory, build different index opinion rating tables;

(6) according to different index opinion rating tables, build different Criteria Decision Making tables;

(7) utilize rough set theory parameter weight matrix.

5. according to claim 1 the micro-power network development harmony of isolated island evaluation method be is characterized in that, wherein, in step (4), described robustness regression algorithm is determined index weights, and its calculation procedure is as follows:

(1) select suitable parameter value;

(2) utilize least square method to solve initial expectation value;

(3) solve new residual values according to initial expectation value;

(4) draw residual error weight according to residual values;

(5) expectation value making new advances according to residual error weight calculation;

(6) judge that adjacent twice expectation value error whether in tolerance interval, if so, finishes to calculate, choosing this expectation value is ideal value, otherwise, return to step (3).

6. the realization to the micro-power network development harmony of isolated island evaluation method according to claim 1, is characterized in that, comprising:

(1) load module, comprises that parameter is worth the intrinsic parameter of the micro-electrical network of required isolated island, determines opinion rating division and robustness regression setting parameter that weight is required;

(2) quantification of targets module, it,, for nondimensionalization that the actual value of index computation model gained is classified, obtains evaluating the normalized quantized value of required indices;

(3) index weights determination module, it is for adopting rough set theory to determine single evaluation grading index weight;

(4) weight Fusion Module, the weight matrix definite to rough set theory merges, and reduces the impact of exceptional value, determines the final weighted value of index;

(5) output module, according to the weighted value of index, the micro-power network development harmony of isolated island appraisal report is analyzed in output.

Images