CN103473605A - Method and system for predicting energy consumption - Google Patents

Abstract

The invention relates to a method and a system for predicting energy consumption, which are used for predicting various sub-divisional energy consumption and total energy consumption in a to-be-predicted region; because prediction is carried out by adopting the combination of a time sequence and a structure, the accuracy of a predicting result is improved; furthermore, the method and the system have wide applicability and can provide reference for adjustment of the energy structure.

Description

Energy-consuming Forecasting Methodology and system

Technical field

The present invention relates to the energy-consuming electric powder prediction, particularly relate to a kind of energy-consuming Forecasting Methodology and system.

Background technology

Building a resource-conserving and environment-friendly society is the inevitable requirement of regional development, is a long-term systems engineering, needs government, enterprise and resident's joint efforts and Collaboration.Government department should provide corresponding policy guide and coordination for implementing useful energy-consuming behavior.

If can be predicted the energy-consuming trend of following several years, contributing to provides the data reference for Energy restructuring.But at present, the technology that energy-consuming is predicted only lays particular emphasis on temporal data statistics, and the result of predicting can not reflect the energy-consuming development trend all-sidedly and accurately.

Summary of the invention

Based on above-mentioned situation, the present invention proposes a kind of energy-consuming Forecasting Methodology and system, to improve the accuracy of energy-consuming prediction.

A kind of energy-consuming Forecasting Methodology comprises step:

Obtain the historical data of petroleum consumption, Natural Gas Consumption Using, consumption of coal and the total energy consumption in zone to be predicted;

According to the sequence characteristic of described historical data, set up respectively petroleum consumption forecast model, Natural Gas Consumption Using forecast model, consumption of coal forecast model and total energy consumption forecast model;

According to the structural relation between petroleum consumption, Natural Gas Consumption Using and consumption of coal and total energy consumption, and petroleum consumption, Natural Gas Consumption Using and consumption of coal account for the whole relation of association between the proportion of total energy consumption, adjust petroleum consumption forecast model, Natural Gas Consumption Using forecast model, consumption of coal forecast model and the total energy consumption forecast model set up;

Take out petroleum consumption, Natural Gas Consumption Using, consumption of coal and the data of nearest several years of total energy consumption from described historical data, petroleum consumption prediction model, natural gas consumption forecast model, consumption of coal forecast model and total energy consumption forecast model by the data substitution of nearest several years after corresponding adjustment, prediction obtains the development trend of following several years of zone to be predicted in petroleum consumption, Natural Gas Consumption Using, consumption of coal and total energy consumption.

A kind of energy-consuming prognoses system comprises:

The historical data acquiring unit, for the historical data of petroleum consumption, Natural Gas Consumption Using, consumption of coal and the total energy consumption of obtaining zone to be predicted;

Forecast model is set up unit, for the sequence characteristic according to described historical data, sets up respectively petroleum consumption forecast model, Natural Gas Consumption Using forecast model, consumption of coal forecast model and total energy consumption forecast model;

The forecast model adjustment unit, be used for according to the structural relation between petroleum consumption, Natural Gas Consumption Using and consumption of coal and total energy consumption, and petroleum consumption, Natural Gas Consumption Using and consumption of coal account for the whole relation of association between the proportion of total energy consumption, adjust petroleum consumption forecast model, Natural Gas Consumption Using forecast model, consumption of coal forecast model and the total energy consumption forecast model set up;

The prediction of the development trend unit, for from described historical data, taking out petroleum consumption, Natural Gas Consumption Using, consumption of coal and the data of nearest several years of total energy consumption, petroleum consumption prediction model, natural gas consumption forecast model, consumption of coal forecast model and total energy consumption forecast model by the data substitution of nearest several years after corresponding adjustment, prediction obtains the development trend of following several years of zone to be predicted in petroleum consumption, Natural Gas Consumption Using, consumption of coal and total energy consumption.

Energy-consuming Forecasting Methodology of the present invention and system, treating in estimation range respectively itemize energy consumption and total energy consumption is predicted, because the mode that adopts sequential to combine with structure is predicted, improved the accuracy predicted the outcome, and there is applicability widely, for Energy restructuring provides reference.

The accompanying drawing explanation

Fig. 1 is energy-consuming proportion prediction schematic diagram;

Fig. 2 is the energy consumption structure schematic diagram that predicts the outcome;

The schematic flow sheet that Fig. 3 is energy-consuming Forecasting Methodology of the present invention;

The structural representation that Fig. 4 is energy-consuming prognoses system of the present invention.

Embodiment

The analytic process of following paper energy forecast.

1, quantitative forecast basic model

Quantitative forecast is basically based on following One-step Extrapolation prognostic equation

$Y_t=C+\underset{i=1}{\overset{p}{\mathrm{\Σ}}}{A}_i{Y}_{t-i}+\underset{i=1}{\overset{q}{\mathrm{\Σ}}}{B}_i{X}_{t-i}---\left(1\right)$

Wherein n (>=1) ties up target column vector Y _tcan grey iterative generation at time span of forecast, m (>=1) dimensional vector exogenous variable X _tneed to produce by suitable method; At sample phase, Y _tstochastic variable, X _tit is the determinacy variable.Work as Y _tand X _twhile being stationary sequence, can directly adopt classical way to be obtained the estimated value of parameter by sample.

Work as Y _twhen being the unit root sequence, estimate following EC-VAR model

${\mathrm{\ΔY}}_t=C^{\′}+{\mathrm{\α\θ}}_t+\underset{i=1}{\overset{p-1}{\mathrm{\Σ}}}{\mathrm{\Φ}}_i{\mathrm{\ΔY}}_{t-i}+\underset{i=1}{\overset{q-1}{\mathrm{\Σ}}}{\mathrm{\Ψ}}_i{\mathrm{\ΔX}}_{t-i}+{\mathrm{\ξ}}_t---\left(2\right)$

Wherein α is n * r matrix, ξ _tzero-mean independent same distribution process,

${\mathrm{\θ}}_t={\mathrm{\β}}_0+{\mathrm{\β}}_1{Y}_t+{\mathrm{\β}}_2{X}_t=({\mathrm{\β}}_0,{\mathrm{\β}}_1,{\mathrm{\β}}_2)\left(\begin{array}{c}1\\ Y_t\\ X_t\end{array}\right)---\left(3\right)$

(1), the pass of (2) two formula parameters is

$\left\{\begin{array}{c}C=C^{\′}+{\mathrm{\α\β}}_0\\ A_1=I+{\mathrm{\Φ}}_1+{\mathrm{\α\β}}_1\\ A_i={\mathrm{\Φ}}_i-{\mathrm{\Φ}}_{i-1},(i=2,\·\·\·,p-1)\\ A_p=-{\mathrm{\Φ}}_{p-1}\\ B_1={\mathrm{\α\β}}_2+{\mathrm{\Ψ}}_1\\ B_i={\mathrm{\Ψ}}_i-{\mathrm{\Ψ}}_{i-1},(i=2,\·\·\·,q-1)\\ B_q=-{\mathrm{\Ψ}}_{q-1}\end{array}\right.---\left(4\right)$

2, total energy consumption and consumption of coal model

The historical data of China's total energy consumption and consumption of coal (taking the logarithm) is the unit root sequence, by estimating the parameter of (2), obtain corresponding to the total energy consumption prediction model parameters of (1) be

$\left\{\begin{array}{c}C=1.2148,A_1=1.4972,A_2=-0.6793\\ B_1=(-\mathrm{0.8516,0.2316,0.6324},-0.0393)\\ B_2=(0.9695,-0.1482,-\mathrm{0.7012,0.0035})\\ B_3=(\mathrm{0,0,0},-0.0035)\end{array}\right.---\left(5\right)$

The consumption of coal prediction model parameters is

Exogenous variable X _t=(x, x ₁, x ₂, x ₃) be respectively the nominal GDP of China and tertiary industries name added value to Number Sequence, at time span of forecast, adopt scenario analysis to generate.Wherein the generative process of nominal GDP rate of growth is

r _t＝0.5176r _t-1-0.2289r _t-3-0.2979r _t-5+λ _i (6)

λ wherein _ivalue be (λ ₁, λ ₂, λ ₃, λ ₄)=(9.9765,3.8,15.2,7.5), the be generated as x of nominal GDP to Number Sequence _t=x _t-1+ r _t.

The proportion sequence w of tertiary industries _{1, t}, w _{2, t}, w _{3, t}the VAR model.

$W_t=\underset{i=1}{\overset{p}{\mathrm{\Σ}}}{W}_{t-i}{A}_i+{\mathrm{\ϵ}}_t---\left(7\right)$

W wherein _t=(w _1t, w _2t, w _3t), A is 3 rank square formations, sample in 1978 to 2011 is VAR (1), parameter estimation:

$\hat{A}=\left(\begin{array}{ccc}.8379& .0763& .0858\\ .2149& .8699& -.0848\\ -.2074& .1283& 1.0791\end{array}\right)$

3, oil consumption individual event prediction

Notice that the oil consumption sequence is stably, we have set up the oil consumption behavior model, and the regression equation that is converted to t=0 in 1980 is

y _t＝3.3207+0.6294y _t-1+0.0236t

The test statistics explanation, this simple autoregressive model has been described the Changing Pattern of oil consumption in sample well, and can be for prediction under the condition that does not need other any exogenous variable.

If take this model as according to doing petroleum consumption prediction, be equivalent to and take as follows really qualitative function as according to giving a forecast.

y _T+t＝C+Bα ^t+At,(t＝1,2,…,K) (8)

Wherein T is effective modeling sample number (t=0 in 1980), α=0.6294th, autoregressive coefficient

$\left\{\begin{array}{c}A=\mathrm{\β}{(1-\mathrm{\α})}^{-1}\\ C=\mathrm{AT}+({\mathrm{\α}}_0-\mathrm{\αA}){(1-\mathrm{\α})}^{-1}\\ B=y_T-C\end{array}\right.$

α ₀＝3.3207，β＝0.0236。Get T=31 (2011) in above formula, y _t=11.0558, its concrete expression formula is

y _T+t＝10.8237+0.104963α ^t+0.063564t

α wherein ^t0 the speed of converging to is very fast, and from 2012 to the year two thousand fifty, the predicted value of oil consumption logarithm almost is linear growth.

On the other hand, y " _t=B (ln α) ²α ^t>0, y ' _t=Bln (α) α ^t+ A; Wherein the latter is the annual growth of petroleum consumption, it in time t increase progressively, and y ' _t<A.Again because y ' ₂₀₁₁=0.06356, if the formula of utilization (8) as the foundation of China's oil Consumption forecast, is equivalent to suppose that the China's oil consumption figure is to be greater than 6.356%, to be less than 6.3564% speed increment, irrelevant with other any factor.Its " prediction " result shows, to China's oil consumption in 2045, will reach 56.2 hundred million tons of standard coal equivalents, approaches a supply of world oil in 2009.The data demonstration, the growth rate of a supply of world oil is roughly in 1.5% left and right.So time trend stationary sequence (8) is not suitable for the direct foundation as the long-term forecasting of China's oil consumption figure, is only suitable for doing short-term forecasting.Investigate oil consumption rate of growth sequence for this reason.

Rate of growth coefficient A in oil consumption model (8) can reduce along with the enhancing of technical progress and national energy-conserving and environment-protective consciousness.So oil consumption rate of growth coefficient A should descend along with the increase of time t generally.Consider the dull rate of growth function f (t) descended=r _t+t=C+B λ ^t, C wherein, B, λ is undetermined parameter.Lim _{t → ∞}f (t)=0, so C=0; F ' (t)=ln λ B λ ^t<0, so B>0,1>λ>0.Before supposing the year two thousand fifty, the China's oil consumption increase rate is not more than 6%, and average growth rate 2.6% (growth levels of 2011), get B=6, λ=0.95.

Get A=0.95 in formula (8) ^t, obtain the numerical simulation result of petroleum consumption prediction and fiducial interval thereof as Fig. 1, wherein mean that curve between the prediction center is that the error of rate of growth function is got the normal distribution that average is zero, variance is t.

We are derived China in 1980 to 2011 " average growth rate oil consumption function " (8) of describing with the petroleum consumption logarithm by (petroleum consumption is to number variable) regression model of China's oil consumer behavior.But " the permanent rate of growth " of this function is unsuitable for being directly used in the oil consumption long-term forecasting.For this reason, we design according to increasing law and the possibility situation of oil consumption the rate of growth function (9) successively decreased in time.Form the petroleum consumption prediction model in conjunction with two functions.

3, natural gas consumption prediction

Identical with petroleum consumption prediction, we can derive " the natural gas consumption behavior model " of the formula of being similar to (8)

y _T+t＝C+Bα ^t+A(T′+t) ²,(t＝1,2,…)， (9)

A=β (1-α) wherein ^-1, B=y _t-AT '-C, C=a (1-α) ^-1, T '=T-α (1-α) ^-1.

It should be noted that (9) are based on the natural gas consumption characteristic model of setting up in t=0 (nineteen eighty-two) sample.The numerical simulation demonstration, when t gets positive integer (T+t is in sample), the error of (9) is very little, and the error of the different same samples of T value is differed to very little; But if t gets negative integer, error is along with the absolute value of t increases and increases.That is to say, (9) are feed forward models.

Get T=29 (2011), y _t=9.6708, calculate China's natural gas consumption feed forward prediction function

y _T+t＝7.3937+0.0542×α ^t+0.02952(t+27.4407) ² (10)

The predicted value that generates the China's natural gas consumption of 2012 to the year two thousand fifty by it becomes the high-order exponential increase.

Similar with above-mentioned petroleum consumption prediction, (9) can be used as foundation (for example 3 to 8 years) of brachymedial phase prediction, and can not be directly used in the medium-and long-term forecasting prediction, but on the basis of modification rate of growth coefficient as basis for forecasting.

Ln (α) B α in formula (9) ^t+ AT '>0, y " _t+t=[ln (α)] ²b α ^t+ 2A>0, so the rate of growth y ' of natural gas consumption _t+t=ln (α) B α ^t+ A (T '+t) singly increase.Due to α ^tclose to zero, so A (t) (T '+t) A ≈ y ' _t+t.We only need to arrange A (t) sight changed occurs: establish T _i(>0) be that the moment changed occurs time span of forecast natural gas consumption rate of growth, for t≤T _inatural gas consumption changes according to the rule of (10), as t>T _ithe time A (t) be time τ=t-T _ithe subtraction function A of (enough large) _i(τ), and meet

$A\left(T_i\right)=A_i\left(0\right),A^{\&prime;}\left(T_i\right)=A_i^{\&prime;}\left(0\right),\underset{t\&RightArrow;\&infin;}{\lim }t{A}_i\left(t\right)=A_0>0.---\left(11\right)$

To any λ>0, get A _i(t, λ)=(T '+T _i+ t) A (1+ λ t ²) ^-1, A (T is arranged _i)=A _i(λ, 0)=(T '+T _i) A, A ' (T _i)=A ' _i(λ, 0)=A ¹, lim _{t → ∞}a _i(λ, t) t=A λ ^-1.As t>T _ithe time, (11.20) become

$y_{T+t}=C+{\mathrm{B\&alpha;}}^t+\frac{A{(T^{\&prime;}+t)}^2}{1+\mathrm{\&lambda;}{(t-T_i)}^2}\underset{t\&RightArrow;\&infin;}{\&DoubleRightArrow;}C+\frac{A}{\mathrm{\&lambda;}}.$

That is to say, natural gas consumption will be tending towards fixing constant.For example, get λ _i=(T '+T _i) ^-2, y _t+tbe tending towards

" recurrence " is to close to initial value; Get λ _i=A, y _t+tbe tending towards C+1.

Using the value of parameter lambda as ultimate limit state, with time T _ias the sight that natural gas consumption rate of growth variation time point is set, arrange as follows; The logarithm sequence prediction of corresponding sight the results are shown in Figure 3.

Sight I---λ _i,I=(T '+T _i) ^-2;

Sight II---λ _{i, II}=A;

Sight 1:T ₁=5, China's natural gas consumption Changing Pattern (10) remains to 2016, and growth rate is followed A afterwards _i(t) descend;

Sight 2:T ₂=9, i.e. China's natural gas consumption rises to the level of oil consumption in 2011 with the rule of sample phase, and (the year two thousand twenty) growth rate is followed A afterwards _i(t) descend;

Sight 3:T ₃=13, i.e. China's natural gas consumption rises to the level of consumption of coal in 2008 with the rule of sample phase, and (2025) growth rate is followed A afterwards _i(t) descend.

Due to (T '+T _i) ^-2<A, for same time point sight i (=1,2,3), the predicted value of sight I is greater than sight II.

4, integrated forecasting

The consumption variable of energy-consuming has structural relation:

Y _e＝Y _c+Y _p+Y _g+Y _o （12）

Y wherein _etotal energy consumption, Y _c, Y _pand Y _grespectively coal, oil and natural gas consumption figure, Y _oit is the consumption figure of water, core, wind and other biomass power generation.

Up to the present Y _oaccounting very little, generally can be at Y _e-(Y _c+ Y _p+ Y _gunder the constraint condition of)>0, adopt vectorial autoregression to set up Y _e, Y _c, Y _pand Y _gthe Changing Pattern of four variablees and forecast model.But as previously mentioned, these four variablees have the statistical nature that three classes are different, make us set up respectively in front the forecast model of relevant variable.Yet aforementioned predicting the outcome can not guarantee the establishment of (12) formula, and its Comprehensive Model problem below is discussed.

-----------------

¹A′ _I(t)＝A(1+λ _It ²) ^-2[1-λ _It ²-2λ _It(T′+T _I)]

(1) comprehensive selection that subitem predicts the outcome

As previously mentioned, consumption of coal and total energy consumption sequence have essentially identical statistical property, forecasting mechanism and essentially identical forecast result.But owing to also having other energy kind, so two performances of sequence in the sample phase have certain difference, cause take the sample phase to consume certain difference that predicts the outcome that the sequence behavior model is foundation, the sight difference that for example combined prediction is corresponding, and may occur that the consumption of coal of prediction is greater than the situation of total energy consumption.That is to say, predicting the outcome of corresponding sight can not guarantee the establishment of (12) formula.So need to do for predicting the outcome corresponding adjustment and combination.

At first, according to total energy consumption and the coal of front, predict the outcome, the coal proportion of " most probable result " is between 0.4 to 0.83 accordingly, wherein to the proportion before 2035 on 70%, the proportion of latter 15 years drops to 40.84% of whole year the year two thousand fifty year by year.By contrast, the predicted value of the consumption of coal upper limit is excessive, the situation that exists proportion to be greater than 1.So we only get middle predicted value and do comprehensive prediction and analysis.

Petroleum consumption prediction only has a result, although predicted value has experienced from being incremented to the read procedure that successively decreases, its proportion that accounts for corresponding total energy consumption is from 17.79% monotone decreasing 6.43% year by year.Consider energy consumption structure identical relation (12), the natural gas consumption prediction can only coordinate sight II1.

Because the main determining factor of natural gas consumption is source of the gas and supply and consumption facility, and in the cost of supply and consumption facility, fixed investment accounts for larger proportion, so the possibility of the medium-term and long-term trend decline of natural gas consumption is very little.For this reason, need to revise the model prediction result: under note (12) condition, the model predication value of (9) formula is G _t, the rock gas predicted value in integrated forecasting is

On the other hand, from 2021 to 2031, consumption of coal predicted value accounting reaches between 75% to 82%, causes coal, three sums of oil and natural gas to be greater than total energy consumption.In coming few decades, the consumption figure of the non-fossil energy outside above-mentioned three energy kinds extremely proportion all will be improved to some extent.So, we take " proportion of consumption of coal is no more than 76% " revise predicting the outcome of consumption of coal as principle, Fig. 1 is shown in by the energy consumption structure of the energy-consuming prediction obtained, wherein regional a. is the difference of total energy consumption and coal, oil, three kinds of main fossil energy consumption figures of rock gas, be mainly derived from the energy consumption of the generating of the primary energy such as water, core, wind, sun power and other biomass energy or other available energy mode, be referred to as " the non-carbon emission energy " or non-fossil energy.

(2) energy-consuming integrated forecasting

The difference that is limited to energy-consuming sequence statistical property, we predict the outcome and have constructed the energy-consuming integrated forecasting based on subitem.But from above-mentioned selection course, can find out, this integrated approach lacks systematicness.Similar above about the prediction of GDP and tertiary industries added value, we can construct the system prediction scheme of total energy consumption and structure thereof.

The proportion that coal, oil and natural gas account for China's total energy consumption is designated as respectively W _c, W _pand W _g, the ADF check shows that they are all unit root sequence (assay slightly).The consumption proportion W of other energy such as water power, nuclear power _n=100-(W _c+ W _p+ W _g), the ADF check shows that the significance probability of its refusal unit root hypothesis is 0.0566, to suppose close to 95% confidence level refusal unit root.Can think that it is the stationary sequence with linear time trend; Can guess sequence W _c, W _pand W _gthere is the whole relation of association.Johanson assists whole check to show, they have 1 whole relation of association, and assist in whole relation equation linear time trend is arranged.

Application Johanson association adjusting technique obtains assisting whole pass to be

θ _t＝W _c,t+1.0347W _p,t+1.4224W _g,t+0.0394t-100.4 (13)

T=1 in 1978 wherein.On the other hand, by W _cto W _pand W _grecurrence obtains regression equation

W _c＝-0.9935W _p-1.1542W _g-0.0496t+99.08

The confidence level of its regression residuals ADF check refusal unit root hypothesis is greater than 99.9%.So, assist whole relation (13) to set up, corresponding to the parameter p of the EC-VAR model of (2) formula=4, q=0, other major parameter estimated value is

$\mathrm{\&alpha;}=\left(\begin{array}{c}9.8860\\ -10.6406\\ 0\end{array}\right),c=\begin{array}{c}\left(\begin{array}{c}-0.6584\\ 0.5679\\ 0\end{array}\right),\end{array}$

Coal, oil and natural gas account for the proportion W of total energy consumption _c, W _pand W _gby following EC-VAR model, produced

${\mathrm{\&Delta;y}}_t=c+{\mathrm{\&alpha;\&theta;}}_{t-1}^{\&prime;}+\underset{k=1}{\overset{p}{\mathrm{\&Sigma;}}}{\mathrm{\&beta;}}_k{\mathrm{\&Delta;y}}_{t-k}+\underset{k=1}{\overset{q}{\mathrm{\&Sigma;}}}{\mathrm{\&Phi;}}_k{\mathrm{\&Delta;X}}_{t-k}^{\&prime;}+{\mathrm{\&epsiv;}}_t---\left(14\right)$

Wherein with

the t statistic only have 0.0207 and 0.5449, can not refuse null null hypothesis, so they are set to 0.Impulse response is checked and is shown, the system that regression model is described has stability.

By parametric regression value substitution regression model, the demonstration that predicts the outcome obtained (is shown in Fig. 1, gas means the rock gas zone, petrol means petroleum region, coal means the coal zone), the proportion of coal in China and oil consumption can descend generally, but will remain on respectively more than 70% and 10% to the year two thousand fifty; Gas gravity keeps rising tendency, reaches 10% left and right.The proportion of other non-fossil energy consumption also improves year by year, reaches 7% to the year two thousand fifty.

After obtaining the predicting the outcome of main fossil energy consumption proportion, for given total energy consumption, can predicting the outcome in the hope of three kinds of main fossil energy levels of consumption.For " most probable predicts the outcome ", (gas means the rock gas zone as Fig. 2 to calculate the result of China's energy consumption structure by the prediction proportion of three kinds of fossil energies, petrol means petroleum region, and coal means the coal zone, and no-carbon means non-carbon emission energy zone).As seen from the figure, although non-fossil energy total quantity consumed minimum, the growth rate of its proportion and absolute number is the fastest, to the year two thousand fifty its consumption figure will reach 14.74 hundred million tons of standard coal equivalents, be equivalent to 2 times of petroleum consumption in 2011; Next is the growth rate of rock gas, the growth rate minimum be oil consumption.

According to above analysis, sum up energy-consuming Forecasting Methodology of the present invention and system.The step of following paper energy-consuming Forecasting Methodology of the present invention, as shown in Figure 3:

The historical data of step S101, the petroleum consumption that obtains zone to be predicted, Natural Gas Consumption Using, consumption of coal and total energy consumption;

Step S102, according to the sequence characteristic of described historical data, set up respectively petroleum consumption forecast model, Natural Gas Consumption Using forecast model, consumption of coal forecast model and total energy consumption forecast model;

Step S103, according to the structural relation between petroleum consumption, Natural Gas Consumption Using and consumption of coal and total energy consumption, and petroleum consumption, Natural Gas Consumption Using and consumption of coal account for the whole relation of association between the proportion of total energy consumption, adjust petroleum consumption forecast model, Natural Gas Consumption Using forecast model, consumption of coal forecast model and the total energy consumption forecast model set up;

Step S104, take out petroleum consumption, Natural Gas Consumption Using, consumption of coal and the data of nearest several years of total energy consumption from described historical data, petroleum consumption prediction model, natural gas consumption forecast model, consumption of coal forecast model and total energy consumption forecast model by the data substitution of nearest several years after corresponding adjustment, prediction obtains the development trend of following several years of zone to be predicted in petroleum consumption, Natural Gas Consumption Using, consumption of coal and total energy consumption.

Energy-consuming prognoses system of the present invention is the system corresponding with said method, as shown in Figure 4, comprising:

The historical data acquiring unit, for the historical data of petroleum consumption, Natural Gas Consumption Using, consumption of coal and the total energy consumption of obtaining zone to be predicted;

Forecast model is set up unit, for the sequence characteristic according to described historical data, sets up respectively petroleum consumption forecast model, Natural Gas Consumption Using forecast model, consumption of coal forecast model and total energy consumption forecast model;

The forecast model adjustment unit, be used for according to the structural relation between petroleum consumption, Natural Gas Consumption Using and consumption of coal and total energy consumption, and petroleum consumption, Natural Gas Consumption Using and consumption of coal account for the whole relation of association between the proportion of total energy consumption, adjust petroleum consumption forecast model, Natural Gas Consumption Using forecast model, consumption of coal forecast model and the total energy consumption forecast model set up;

The prediction of the development trend unit, for from described historical data, taking out petroleum consumption, Natural Gas Consumption Using, consumption of coal and the data of nearest several years of total energy consumption, petroleum consumption prediction model, natural gas consumption forecast model, consumption of coal forecast model and total energy consumption forecast model by the data substitution of nearest several years after corresponding adjustment, prediction obtains the development trend of following several years of zone to be predicted in petroleum consumption, Natural Gas Consumption Using, consumption of coal and total energy consumption.

To sum up, energy-consuming Forecasting Methodology of the present invention and system, treating in estimation range respectively itemize energy consumption and total energy consumption is predicted, because the mode that adopts sequential to combine with structure is predicted, improved the accuracy predicted the outcome, and there is applicability widely, for Energy restructuring provides reference.

The above embodiment has only expressed several embodiment of the present invention, and it describes comparatively concrete and detailed, but can not therefore be interpreted as the restriction to the scope of the claims of the present invention.It should be pointed out that for the person of ordinary skill of the art, without departing from the inventive concept of the premise, can also make some distortion and improvement, these all belong to protection scope of the present invention.Therefore, the protection domain of patent of the present invention should be as the criterion with claims.

Claims (6)

1. an energy-consuming Forecasting Methodology, is characterized in that, comprises step:

Obtain the historical data of petroleum consumption, Natural Gas Consumption Using, consumption of coal and the total energy consumption in zone to be predicted;

According to the sequence characteristic of described historical data, set up respectively petroleum consumption forecast model, Natural Gas Consumption Using forecast model, consumption of coal forecast model and total energy consumption forecast model;

According to the structural relation between petroleum consumption, Natural Gas Consumption Using and consumption of coal and total energy consumption, and petroleum consumption, Natural Gas Consumption Using and consumption of coal account for the whole relation of association between the proportion of total energy consumption, adjust petroleum consumption forecast model, Natural Gas Consumption Using forecast model, consumption of coal forecast model and the total energy consumption forecast model set up;

Take out petroleum consumption, Natural Gas Consumption Using, consumption of coal and the data of nearest several years of total energy consumption from described historical data, petroleum consumption prediction model, natural gas consumption forecast model, consumption of coal forecast model and total energy consumption forecast model by the data substitution of nearest several years after corresponding adjustment, prediction obtains the development trend of following several years of zone to be predicted in petroleum consumption, Natural Gas Consumption Using, consumption of coal and total energy consumption.

2. energy-consuming Forecasting Methodology according to claim 1, is characterized in that,

Structural relation between petroleum consumption, Natural Gas Consumption Using and consumption of coal and total energy consumption is:

Y _e＝Y _c+Y _p+Y _g+Y _o

In above formula, Y _etotal energy consumption, Y _c, Y _pand Y _grespectively coal, oil and natural gas consumption figure, Y _oit is the consumption figure of water, core, wind and other biomass power generation.

3. energy-consuming Forecasting Methodology according to claim 1 and 2, is characterized in that,

The whole pass of association that petroleum consumption, Natural Gas Consumption Using and consumption of coal account between the proportion of total energy consumption is:

θ _t＝W _c,t+1.0347W _p,t+1.4224W _g,t+0.0394t-100.4

In above formula, t expression of years sequence number, θ _tthe whole relation of association that means t, W _c,t, W _p,t, W _g,tmean that respectively t consumption of coal rate, petroleum consumption, Natural Gas Consumption Using account for the proportion of total energy consumption.

4. an energy-consuming prognoses system, is characterized in that, comprising:

The historical data acquiring unit, for the historical data of petroleum consumption, Natural Gas Consumption Using, consumption of coal and the total energy consumption of obtaining zone to be predicted;

Forecast model is set up unit, for the sequence characteristic according to described historical data, sets up respectively petroleum consumption forecast model, Natural Gas Consumption Using forecast model, consumption of coal forecast model and total energy consumption forecast model;

The forecast model adjustment unit, be used for according to the structural relation between petroleum consumption, Natural Gas Consumption Using and consumption of coal and total energy consumption, and petroleum consumption, Natural Gas Consumption Using and consumption of coal account for the whole relation of association between the proportion of total energy consumption, adjust petroleum consumption forecast model, Natural Gas Consumption Using forecast model, consumption of coal forecast model and the total energy consumption forecast model set up;

The prediction of the development trend unit, for from described historical data, taking out petroleum consumption, Natural Gas Consumption Using, consumption of coal and the data of nearest several years of total energy consumption, petroleum consumption prediction model, natural gas consumption forecast model, consumption of coal forecast model and total energy consumption forecast model by the data substitution of nearest several years after corresponding adjustment, prediction obtains the development trend of following several years of zone to be predicted in petroleum consumption, Natural Gas Consumption Using, consumption of coal and total energy consumption.

5. energy-consuming prognoses system according to claim 4, is characterized in that,

Structural relation between petroleum consumption, Natural Gas Consumption Using and consumption of coal and total energy consumption is:

Y _e＝Y _c+Y _p+Y _g+Y _o

In above formula, Y _etotal energy consumption, Y _c, Y _pand Y _grespectively coal, oil and natural gas consumption figure, Y _oit is the consumption figure of water, core, wind and other biomass power generation.

6. according to the described energy-consuming prognoses system of claim 4 or 5, it is characterized in that,

The whole pass of association that petroleum consumption, Natural Gas Consumption Using and consumption of coal account between the proportion of total energy consumption is:

θ _t＝W _c,t+1.0347W _p,t+1.4224W _g,t+0.0394t-100.4

In above formula, t expression of years sequence number, θ _tthe whole relation of association that means t, W _c,t, W _p,t, W _g,tmean that respectively t consumption of coal rate, petroleum consumption, Natural Gas Consumption Using account for the proportion of total energy consumption.

Images