Summary of the invention
The object of the invention is to Quantitative Study power market transaction evaluation criterion measuring method, for the market operation, supervision department provide decision-making.For achieving the above object, the implementation step of a kind of power market transaction evaluation criterion measuring method based on hierarchical clustering of the present invention's proposition is as follows:
S1. before power market transaction, warning index is chosen
Market supply and demand and market structure are the two large essential characteristics in market, are also the most basic factors determining market conclusion of the business result.They not with the will of participant in the market for transfer: market structure is determined by the number of the power plant that bids in market and scale, and when not having new power plant to add market, market structure is relatively stable; Market equilibrium is subject to that installed capacity increases, the impact of social electricity consumption growth and seasonal factor, and in the specific day of trade, market supply and demand situation can be predicted out in advance; Market structure and market supply and demand are two large essential characteristics of any commodity market, are also the most basic factors determining market conclusion of the business result.
(1) market structure class index
Market structure class is the fundamental nature in market.In microeconomics, market structure has four kinds of fundamental types: perfect competition, Monopolistic competition, oligopoly and complete monopoly.This class index is the formation situation by participant in market, and reflect that market may by the degree monopolized, it is index the most frequently used in various countries' Electricity market analysis, comprises Top-m share index and HHI index.
1) Top-m share index
Top-m share refers to the market share shared by m maximum in a market supplier.Conventional Top-4 index in general industry field, namely get m=4, Top-4 index >65% shows that market has the character of oligopoly.This index is larger, shows that market concentration degree is higher.
2) HHI index (Herfindahl-Hirshman Index)
Measure by the quadratic sum of the market share shared by each market supply person:
It is wherein the market share of i-th market supply person.
The HHI that corners the market is 10000, and perfect competition market HHI is tending towards 0.Generally, the market of HHI<1800, should be regarded as competition more abundant.
HHI index has following characteristic:
1) HHI index depends primarily on the size of participant in the market's number and the market share: market member less, market resource distribute more concentrated, then HHI is larger, shows that the possibility abusing monopolization power in market is larger;
The market share of maximum participant in market, maximum to HHI Index Influence: if maximum supplier accounts for 80%, then no matter HHI> (100*80%) 2=6400, also have the supplier that how many are little;
2) HHI index is generally more stable, and the impact by the dynamic change of the quotation strategy of market supply and demand, supplier is little, as long as participant in the market's installation, scale, transmitting capacity of the electric wire netting do not change, HHI is substantially constant.
(2) market supply and demand class index
Market supply-demand ratio (Supply-Demand Ratio), is defined as:
Wherein Q
dthe market aggregate demand of prediction, Q
sit is aggregate supply.
When Γ → 1 or when being less than 1, supply falls short of demand in market, and power plant has monopolization power, can the left and right market price.Therefore, this desired value is less, and market more levels off to and corners the market; Γ is larger, then market supply is more abundant, and competitiveness is better, Γ → ∞ under perfect competition market.
In above formula, " always supply " and " aggregate demand " can need to choose different amounts according to evaluating, and as declared supply and demand ratio=always declare electricity/aggregate demand, the supply and demand ratio in actual market, just refer to and declare supply and demand ratio, supply declares electricity to calculate according to actual.
S2. by the above-mentioned segmentation index that can reflect market structure and market supply and demand determined, power market transaction, without retraining average knock-down price, Top4 index, declaring HHI, is declared supply and demand ratio and is adopted chromatography cluster determination sample data clusters number.
S3. according to the sample data clusters number that chromatography clustering method determines, utilize K Mean Method obtain the final class equation of the ecentre between each classification apart from and class off-centring degree, set up electricity market supply and demand Early-warning Model.
Compared to existing technology, beneficial effect of the present invention is:
1, the present invention chooses the essential characteristic that market supply and demand and market structure describe market, from the most basic factor determining market conclusion of the business result, set up power market transaction without retraining average knock-down price, Top4 index, declaring HHI, declare supply and demand ratio four characteristic indexs, evaluate market after electricity transaction, they not with the will of participant in the market for transfer.Wherein market structure is determined by the number of the power plant that bids in market and scale, and when not having new power plant to add market, market structure is relatively stable; Market equilibrium is subject to that installed capacity increases, the impact of social electricity consumption growth and seasonal factor, and in the specific day of trade, market supply and demand situation can be predicted out in advance.
2, the present invention adopts hierarchy clustering method, quantitative test by index dynamic cataloging be excellent, in, differ from three classes, make things character close be classified as a class, overcome the ambiguity of qualitative analysis, have that recognition speed is fast, classification accuracy is high, classification effectiveness is high, simple operation and other advantages.
3, the cluster result of the present invention's acquisition, can as power market transaction evaluation measuring and calculating criterion score, namely excellent, in, differ from the standard value of three classes, to the assessment of electricity market, there is very strong directiveness, so similar solution can be taked, to a large amount of index calculate standard values that may occur in the future, follow-up electricity market evaluation procedure is made more to embody high efficiency and operability.
Embodiment
As shown in Figure 1, the case study on implementation concrete steps of a kind of power market transaction evaluation criterion measuring method based on hierarchical clustering of the present invention's proposition are as follows:
(1) this patent data Regional Electric Market, the year quotation of certain year and point moon quote data, and according to determining the segmentation index that can reflect market structure and market supply and demand above, be respectively power market transaction without retraining average knock-down price, Top4 index, declaring HHI, declare supply and demand ratio.Collect the collection of historical data and carry out pre-service.
Market supply and demand in certain year, structure index table
Title |
Without the average knock-down price of constraint |
Top4 index |
Declare HHI |
Declare supply and demand ratio |
Year 1 takes turns |
185.79 |
47.5% |
474.97 |
2.61 |
Year 2 takes turns |
202.31 |
43.81% |
609.94 |
1.78 |
May |
270.09 |
65.25% |
997.24 |
2.08 |
June |
267.39 |
84.85% |
1447.53 |
1.63 |
July |
261.37 |
85.29% |
1434.26 |
1.65 |
August |
260.64 |
83.71% |
1398.37 |
1.45 |
September |
260.64 |
83.71% |
1398.37 |
1.45 |
October |
248.55 |
80.63% |
954.52 |
2.30 |
November |
263.96 |
91.01% |
950.62 |
2.72 |
Dec |
263.05 |
73.07% |
807.89 |
2.42 |
(2) by the above-mentioned segmentation index that can reflect market structure and market supply and demand determined, power market transaction is without retraining average knock-down price, Top4 index, declaring HHI, declare supply and demand ratio and adopt chromatography cluster determination sample data clusters number, result as shown in Figure 2.
Above-mentioned classification results shows, and year 1 takes turns, take turns quotation year 2 for the first kind, and June, July, August, September are Equations of The Second Kind, and May, October, November, Dec are the 3rd class
(3) according to the feature at each class center, in conjunction with the meaning of each index, these four indexs can be divided into " excellent, in, poor " Three Estate, set up electricity market supply and demand Early-warning Model.
Final cluster centre
As can be seen here:
(1) take turns in year 1, take turns in year 2 and belong to the 1st class, illustrate the market structure of annual trade market and market supply and demand higher than grade, compare monthly trade market then not as annual trade market, cause price high.
In electricity market, cause the real time price of electric power often to fluctuate because electric power cannot store, the shortage of Demand-side also exacerbates this problem to a certain extent, in order to evade the short period price very easily fluctuated, makes manufacturer tend to sign relatively long-term annual contract.And under the constant condition of electricity needs, the existence in annual contract market will inevitably reduce the total demand in monthly market, make part power plant year trading volume little, part power plant strikes a bargain and causes residual capacity few more in year, exacerbates the imbalance of each power plant in monthly market bidding space.
Therefore, reasonably distribute annual contract electricity, then according to load prediction curve, the Factor Decompositions such as unit maintenance scheduling and machine set technology restriction are to each moon, reasonably distribute monthly plan again, according to the deviation of running result and plan, the annual contract power energy allocation in adjustment unit residue month, and jointly form the monthly generation scheduling of unit with monthly newly-increased electricity, each day is decomposed according to typical day load curve, with the homogeneity of contract of guarantee electricity on room and time, improve the market structure of next stage transaction, avoid the rapid rise of price that next stage is concluded the business.
(2) can also intuitively find out from form, market supply and demand and market structure have material impact to the market price, and market supply and demand is than index and market price inverse change, and market structure HHI index and the market price change in the same way.Therefore, power exchange the index such as computing market supply and demand, structure can be predicted transaction results before each competitive bidding, can look-ahead price trend, utilizes supervision department's " tangible hand " to carry out control, improves stability and the economic benefit in market.