CN101969197A - Bus load forecasting accuracy comprehensive evaluation method - Google Patents

Abstract

The invention discloses a bus load forecasting accuracy comprehensive evaluation method, which comprises the following steps of: acquiring forecast data and real data of each bus load forecasting object, calculating the relative error of the forecast data and the real data by using different reference data according to different load levels, calculating a bus load importance index according to the sensitivity of each bus load forecasting object to stable cross sections and important branch equipment based on sensitivity analysis, calculating a load fluctuation index according to the load historical fluctuation condition, and calculating a bus load forecasting accuracy comprehensive evaluation index according to the indexes. By the method, a more comprehensive and practical index for bus load forecasting accuracy is obtained; bus load forecasting work is more reasonably evaluated by each-level scheduling; the practical level of the bus load forecasting work is continuously improved; and foundations are laid for making reasonable scheduling plans for the scheduling plan department of a power grid and improving the operation safety and economy of the power grid.

Description

A kind of bus load precision of prediction integrated evaluating method

Technical field

The invention belongs to power system load forecast analysis field, relate to a kind of bus load precision of prediction integrated evaluating method.

Background technology

Along with deepening constantly of power system reform, the whole society improves day by day to electrical network and whole power industry safety, requirement energy-conservation, effective operation.For carrying out the requirement of national energy-saving and emission-reduction policy, dispatching patcher must be taked the power generation dispatching mode of science, and accurate bus load prediction is the operation plan of rationally arranging production, and implements the prerequisite and the guarantee of energy-saving power generation dispatching.As the nucleus module of safe nucleus correcting system a few days ago, improve the bus load precision of prediction and become the prerequisite that realizes the scheduling fine-grained management.

With respect to system loading, bus load research object number is numerous, the user of service can not more demand the precision evaluation technology of scientific and reasonable bus load urgently to the manual operation analysis one by one of all objects, weighs the degree of being practical of bus load prediction work in order to science.

Summary of the invention

Purpose of the invention process is to provide a kind of bus load precision of prediction integrated evaluating method, characteristics at bus load, from demands of applications such as safety check and operation plan establishments, take into full account various factors, the final acquisition index more comprehensive and practical to the bus load precision of prediction.

The present invention is directed to the characteristics of bus load prediction, from bus load size, load importance and a plurality of angles of load fluctuation, the computational methods of design bus load precision of prediction comprehensive evaluation index.

The present invention has mainly solved following problem:

Traditionally the bus load prediction reference system loading prediction and evaluation index, mainly considered the difference size between predicted value and the actual value, ideally, must improve constantly all bus load precision of predictions.But the bus load forecasting object is numerous, and part throttle characteristics and the effect size of electrical network all had nothing in common with each other is ignored these characteristics, and work does not have emphasis, is disadvantageous for continuous propelling bus load prediction work.Power department presses for the practical comprehensively bus load precision of prediction comprehensive evaluation index of a cover.

A kind of bus load precision of prediction integrated evaluating method comprises the steps:

1), obtains the prediction data and the real data of each bus load forecasting object according to the modeling of electrical network network analysis;

2) consider different load level height, adopt and calculate the fiducial error index based on the mode of relative value;

3), each bus load forecasting object is calculated the bus load importance index to stable cross section and important branch equipment based on sensitivity analysis;

4) according to load historical volatility situation, calculated load fluctuation index;

5) at last according to above every index, calculate bus load precision of prediction comprehensive evaluation index, calculate and finish.

Method of the present invention has following characteristics and function:

1) can consider otherness between the different bus loads, form bus load precision of prediction comprehensive evaluation index computational methods from load level, load importance, load fluctuation several respects, help more rational evaluation bus load prediction work of scheduling at different levels, improve constantly the practicability level of bus load prediction work, for dispatching of power netwoks planning authorities works out rational operation plan, the fail safe and the economy that improve operation of power networks lay the first stone.

2) with reference to the result of calculation of bus load importance index, can directly instruct dispatching of power netwoks department to carry out the emphasis of bus load prediction work, improve the specific aim of prediction work.

3) with reference to the result of calculation of bus load fluctuation index, can allow dispatching of power netwoks department that different forecasting objects are produced different precision expectations, understand the load fluctuation situation, take different predicting strategies to improve precision of prediction.

The invention has the beneficial effects as follows at the otherness between the different bus loads, provide from the bus load precision of prediction comprehensive evaluation index computational methods of load level, load importance, load fluctuation several respects.No matter be the back assessment that self bus load precision of prediction is carried out in scheduling at the corresponding levels, still the examination of bus load prediction practicability carried out in subordinate's scheduling, can practical computational methods of the present invention carry out analytical calculation.

The evaluation index computational methods of considering load level, importance and fluctuation are found out in research of bus load precision of prediction comprehensive evaluation index and trial that this method is carried out under the actual electric network data.This method is comprehensive by many index, final formation to the overall assessment of bus load prediction work comprehensive and reasonable more, help to improve constantly the practicability level of bus load prediction work, in the dispatching of power netwoks planning, obtain basic data more accurately, work out rational operation plan, improve the fail safe and the economy of operation of power networks.

Embodiment

The invention discloses a kind of computational methods of bus load precision of prediction comprehensive evaluation index.Be a preferred case study on implementation of the present invention below, comprised employing the inventive method in electrical network bus load precision of prediction evaluation analysis process a few days ago, its feature, purpose and advantage can be found out from the explanation of embodiment.

Electrical network bus load prediction a few days ago be dispatching of power netwoks department carry the previous day in the electrical network all bus loads predict, for its precision is estimated, need contrast bus load real data and electrical network actual operating, analyze electrical network bus load accuracy of predicting weak spot.

Bus load precision of prediction comprehensive evaluation index of the present invention mainly based on to the bus load characteristic analysis, by analysis load level, load importance and load fluctuation sub-indicator, finally forms comprehensive evaluation index.

(1) according to the modeling of electrical network network analysis, determines the bus load forecasting object, obtain the prediction data and the real data of each bus load forecasting object.

(2) consider different load level height, adopt and calculate the fiducial error index based on the mode of relative value.

(3), each bus load forecasting object is calculated the bus load importance index to stable cross section and important branch equipment based on sensitivity analysis.

(4) according to load historical volatility situation, calculated load fluctuation index

(5) at last according to above every index, calculate bus load precision of prediction comprehensive evaluation index, calculate and finish.

1) bus load importance index

In tidal current analysis, the meritorious general expression formula of trend of circuit ij is:

$P_{\mathrm{ij}}{=V}_i^{2}g-V_i{V}_{j}g{\cos \mathrm{\θ}}_{\mathrm{ij}}-V_i{V}_{j}b\sin {\mathrm{\θ}}_{\mathrm{ij}}$ (formula 1)

In the formula, P _IjBe the meritorious trend in circuit ij top; V _iV _jBe circuit ij two ends busbar voltage amplitude; θ _IjBe circuit ij two ends busbar voltage phase angle difference; G, b is that electricity is led and susceptance for circuit ij.

When the bus variable power, the meritorious and busbar voltage of circuit ij can change, and following formula is carried out Taylor expansion near initial point, ignores the above high-order term of single order and obtains

$\mathrm{\Δ}{P}_{\mathrm{ij}}=\left[\frac{\∂P_{\mathrm{ij}}}{\∂\mathrm{\θ}}\frac{\∂P_{\mathrm{ij}}}{\∂V}\right]\left[\begin{array}{c}\mathrm{\Δ\θ}\\ \mathrm{\ΔV}\end{array}\right]$ (formula 2)

In the formula

$\frac{\∂P_{\mathrm{ij}}}{\∂\mathrm{\θ}}=\left[\begin{array}{ccccc}0...0& \frac{\∂p_{\mathrm{ij}}}{\∂{\mathrm{\θ}}_j}& 0...0& \frac{\∂P_{\mathrm{ij}}}{\∂{\mathrm{\θ}}_j}& 0...0\end{array}\right]$

$\frac{\∂P_{\mathrm{ij}}}{\∂V}=\left[\begin{array}{ccccc}0...0& \frac{\∂P_{\mathrm{ij}}}{\∂V_j}& 0...0& \frac{\∂P_{\mathrm{ij}}}{\∂V_j}& 0...0\end{array}\right]$

Power flow equation is write as

$\left[\begin{array}{c}\mathrm{\Δ\θ}\\ \mathrm{\ΔV}\end{array}\right]={\left[\begin{array}{cc}\frac{\frac{\∂P}{\∂\mathrm{\θ}}}{\frac{\∂P}{\∂V}}& \frac{\frac{\∂Q}{\∂\mathrm{\θ}}}{\frac{\∂Q}{\∂V}}\end{array}\right]}^{-1}\left[\begin{array}{c}\mathrm{\ΔP}\\ \mathrm{\ΔQ}\end{array}\right]$ (formula 4)

$\mathrm{\Δ}{P}_{\mathrm{ij}}=\left[\begin{array}{cc}\frac{\∂P_{\mathrm{ij}}}{\∂\mathrm{\θ}}& \frac{\∂P_{\mathrm{ij}}}{\∂V}\end{array}\right]{\left[\begin{array}{cc}\frac{\frac{\∂P}{\∂\mathrm{\θ}}}{\frac{\∂Q}{\∂V}}& \frac{\frac{\∂Q}{\∂\mathrm{\θ}}}{\frac{\∂Q}{\∂V}}\end{array}\right]}^{-1}\left[\begin{array}{c}\mathrm{\ΔP}\\ \mathrm{\ΔQ}\end{array}\right]$ (formula 5)

P, Q, V, θ are respectively the meritorious injection rate of each bus nodes, idle injection rate and node voltage amplitude and phase angle.So obtain

$\left[S_{\mathrm{ij}}\right]=\left[\begin{array}{cc}\frac{\∂P_{\mathrm{ij}}}{\∂\mathrm{\θ}}& \frac{\∂P_{\mathrm{ij}}}{\∂V}\end{array}\right]{\left[\begin{array}{cc}\frac{\frac{\∂P}{\∂\mathrm{\θ}}}{\frac{\∂P}{\∂V}}& \frac{\frac{\∂Q}{\∂\mathrm{\θ}}}{\frac{\∂Q}{\∂V}}\end{array}\right]}^{-1}$ (formula 6)

Matrix S _IjBe sensitivity matrix, the sensitivity that its element representation is gained merit to bus for meritorious trend.P, Q, V, θ are respectively the meritorious injection rate of each bus nodes, idle injection rate and node voltage amplitude and phase angle.

The stable cross section that need check consider in safety in the operation plan compilation process be the important monitored object of electrical network, can be with bus load to the sensitivity of stable cross section as this bus load importance step-index parameter.

Under certain operational mode, establish bus load object b _iTo stable cross section W _jSensitivity be S (b _i, W _j), can make up bus i and be one group of significance level index of paying close attention to stable cross section:

η _i=max{S (b _i, W _j) (formula 7)

2) bus load fluctuation index

Historical data according to load counts in M days the fluctuation mean square deviation of period k bus load i

${\mathrm{\σ}}_i=\frac{1}{N}\underset{k=1}{\overset{N}{\mathrm{\Σ}}}\sqrt{\underset{j=1}{\overset{M}{\mathrm{\Σ}}}\frac{{(X_{j,k}-E_k)}^2}{M}}$ (formula 8)

X in the formula _{J, k}For bus load i in the j historical data of day k period, N is the time hop count of every day statistics,

Be the mean value of bus load i in the k period.With σ _iFluctuation index as this bus load.

3) bus load precision of prediction comprehensive evaluation index

The fiducial error of single busbar connection load i period k

$E_{i,k}=\frac{|F-L|}{\mathrm{Base}}\×100\%$ (formula 9)

Wherein: F is the bus load predicted value, and L is the bus load actual value, and Base is the load fiducial value.Fiducial value can be set according to different electrical network features, and the fiducial value that load level is high is big, and the fiducial value that load level is low is little.Usually can adopt electric pressure to divide.

In conjunction with load fluctuation to crucial branch road trend influence, payload influence, load fluctuation amplitude and aspect factor, we form final practicability accuracy rate error analysis index:

$\hat{Z}=(1-\sqrt{\frac{1}{N}\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{A}_k^{\′2}})\×100\%$ (formula 10)

In the formula:

Be comprehensive accuracy rate index

Be the composition error of all buses at period k, M is the fate of statistics, and N is the time hop count of statistics every day.

This system comprises the back evaluation function of scheduling at the corresponding levels and subordinate's scheduling bus load prediction, is after finishing prediction work, uses that the present invention carries out finishing on the basis that bus load prediction and evaluation index calculates.

The evaluation index computational methods of considering load level, importance and fluctuation are found out in research of bus load precision of prediction comprehensive evaluation index and trial that this method is carried out under the actual electric network data.This method is comprehensive by many index, final formation the overall assessment of bus load prediction work comprehensive and reasonable more.No matter be the back assessment that self bus load precision of prediction is carried out in scheduling at the corresponding levels, still the examination of bus load prediction practicability carried out in subordinate's scheduling, can practical computational methods of the present invention carry out analytical calculation.It is achievement and the deficiency that helps the prediction of scientific evaluation bus load that one of the present invention significantly acts on, improve constantly the practicability level of bus load prediction work, in the dispatching of power netwoks planning, obtain basic data more accurately, work out rational operation plan, improve the fail safe and the economy of operation of power networks.

According to specific exemplary case study on implementation the present invention has been described herein.Do not break away to one skilled in the art and carry out suitable replacement under the scope of the invention or modification is conspicuous.Exemplary case study on implementation only is illustrative, rather than to the restriction of scope of the present invention, scope of the present invention is by affiliated claim definition.

Claims (5)

1. a bus load precision of prediction integrated evaluating method is characterized in that, may further comprise the steps:

1), obtains the prediction data and the real data of each bus load forecasting object according to the modeling of electrical network network analysis;

2) consider different load level height, adopt and calculate the fiducial error index based on the mode of relative value;

3), each bus load forecasting object is calculated the bus load importance index to stable cross section and important branch equipment based on sensitivity analysis;

4) according to load historical volatility situation, calculated load fluctuation index;

5) at last according to above every index, calculate bus load precision of prediction comprehensive evaluation index, calculate and finish.

2. bus load precision of prediction integrated evaluating method according to claim 1 is characterized in that: in described step 3), bus load importance index computational methods are as follows:

With bus load to the sensitivity of stable cross section as this bus load importance step-index parameter:

$\left[S_{\mathrm{ij}}\right]=\left[\begin{array}{cc}\frac{\∂P_{\mathrm{ij}}}{\∂\mathrm{\θ}}& \frac{\∂P_{\mathrm{ij}}}{\∂V}\end{array}\right]{\left[\begin{array}{cc}\frac{\frac{\∂P}{\∂\mathrm{\θ}}}{\frac{\∂P}{\∂V}}& \frac{\frac{\∂Q}{\∂\mathrm{\θ}}}{\frac{\∂Q}{\∂V}}\end{array}\right]}^{-1}$

Wherein, matrix S _IjBe sensitivity matrix, the sensitivity that its element representation is gained merit to bus for meritorious trend, P _IjBe the meritorious trend in circuit ij top, P, Q, V, θ are respectively the meritorious injection rate of each bus nodes, idle injection rate and node voltage amplitude and phase angle;

Stable cross section is the branch road group, is the meritorious summation of one group of circuit, under certain operational mode, establishes bus load object b _iTo stable cross section W _jSensitivity be S (b _i, W _j), structure bus i to one group of significance level index of paying close attention to stable cross section is:

η _i＝max{S(b _i，W _j)}。

3. bus load precision of prediction integrated evaluating method according to claim 1 is characterized in that: in described step 4), bus load fluctuation index calculating method is as follows:

Historical data according to load counts in M days the fluctuation mean square deviation of bus load i

${\mathrm{\σ}}_i=\frac{1}{N}\underset{k=1}{\overset{N}{\mathrm{\Σ}}}\sqrt{\underset{j=1}{\overset{M}{\mathrm{\Σ}}}\frac{{(X_{j,k}-E_k)}^2}{M}}$

X in the formula _{J, k}For bus load i in the j historical data of day k period, N is the time hop count of every day statistics, For the mean value of bus load i, with σ in the k period _iFluctuation index as this bus load.

4. bus load precision of prediction integrated evaluating method according to claim 1 is characterized in that: in described step 5), bus load precision of prediction comprehensive evaluation index computational methods are as follows:

The fiducial error of single busbar connection load i period k:

$E_{i,k}=\frac{|F-L|}{\mathrm{Base}}\×100\%$

Wherein: F is the bus load predicted value, and L is the bus load actual value, and Base is the load fiducial value, and accuracy rate error analysis index is:

$\hat{Z}=(1-\sqrt{\frac{1}{N}\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{A}_k^{\′2}})\×100\%$

In the formula:

Be comprehensive accuracy rate index,

Be the composition error of all buses at period k, M is the fate of statistics, and N is the time hop count of statistics every day.

5. bus load precision of prediction integrated evaluating method according to claim 4, it is characterized in that: described fiducial value is set according to different electrical network features, and the fiducial value that load level is high is big, and the fiducial value that load level is low is little, adopts electric pressure to divide.