Summary of the invention
For in prior art for the problem optimized based on comprehensive diversity load efficiency and the total solution of intelligent demand response lacks, the object of the invention is to the energy management method and the system that propose a kind of distribution garden, can to optimize based on comprehensive diversity load efficiency and the total solution of intelligent demand response for distribution garden provides.
Based on above-mentioned purpose, technical scheme provided by the invention is as follows:
According to an aspect of the present invention, provide a kind of energy management method of distribution garden, comprising:
According to the characteristic of power load each in distribution garden, all power loads in distribution garden are divided into storage-type load, removable load, interruptible load, power-adjustable cooling and heating load and uncontrollable load;
At prediction day proxima luce (prox. luc), obtain temperature prediction curve next day, uncontrollable demand history data and power-adjustable cooling and heating load historical data respectively, and according to next day temperature prediction curve, uncontrollable demand history power data and power-adjustable cooling and heating load historical power data, calculate and obtain uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day;
At prediction day proxima luce (prox. luc), obtain photovoltaic generating system time daily output prediction curve respectively, interruptible load work program next day data, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, and according to photovoltaic generating system time daily output prediction curve, interruptible load work program next day data, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day, calculate and obtain removable load operation curve next day and storage-type load operation curve next day,
Predicting same day day, the power obtained respectively under the comfort level sacrifice that the Demand Side Response information issued on Utilities Electric Co. same day and power-adjustable cooling and heating load allowed the same day reduces amplitude, and according to photovoltaic generating system when daily output prediction curve, interruptible load work program on same day data, uncontrollable load prediction curve on the same day, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, power under the Demand Side Response information that Utilities Electric Co. same day is issued and the comfort level sacrifice that power-adjustable cooling and heating load allowed the same day reduces amplitude, calculate and obtain the start working time period value of removable load under Demand Side Response, the beginning power stage time period value of storage-type load under Demand Side Response, and interruptible load load decreasing value under Demand Side Response, and result of calculation is exported with curve form,
After prediction day, obtain photovoltaic generating system actual power curve on the same day, interruptible load actual electricity consumption on same day curve, removable load actual electricity consumption on same day curve, storage-type load actual operation curve on the same day, uncontrollable load actual electricity consumption on same day curve, power-adjustable cooling and heating load actual electricity consumption curve on the same day, and by all actual electricity consumption curve, actual power curve, actual operation curve according to optimisation strategy execution generation and calculate without the comparing property of electricity consumption situation under optimisation strategy;
After prediction day, the electricity consumption situation under the actual electricity consumption curve of the interruptible load produced according to Demand Side Response strategy execution according to all, removable load, storage-type load and power-adjustable cooling and heating load and actual operation curve and non-reference requirement respond compares comparative calculating.
Wherein, storage-type load comprises electric energy storage load and cold and hot energy storage load, and wherein, cold and hot energy storage load comprises cold energy storage load and hot energy storage load; Power-adjustable cooling and heating load comprises power-adjustable refrigeration duty and power-adjustable thermal load; At prediction day proxima luce (prox. luc), calculate and obtain power-adjustable cooling and heating load prediction curve next day, for calculating power-adjustable refrigeration duty prediction curve next day and power-adjustable thermal load prediction curve next day respectively; At prediction day proxima luce (prox. luc), calculating and obtain storage-type load operation curve next day, for calculating cold and hot energy storage load operation curve next day, comprising cold energy storage load operation curve next day and hot energy storage load operation curve next day; Predicting same day day, calculating and obtain the beginning power stage time period value of storage-type load under Demand Side Response, for calculating electric energy storage load, cold energy storage load and the hot energy storage load beginning power stage time period value under Demand Side Response respectively.
And, at prediction day proxima luce (prox. luc), according to next day temperature prediction curve, uncontrollable demand history power data and power-adjustable cooling and heating load historical power data, calculate and obtain uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day and be: according to next day temperature prediction curve with uncontrollable demand history power data, use uncontrollable load prediction curve next day of Forecasting Methodology calculating acquisition based on gray model, and according to temperature prediction curve and power-adjustable cooling and heating load historical power data next day, use the Forecasting Methodology based on gray model and neural network ensemble to calculate power-adjustable refrigeration duty prediction curve next day and power-adjustable thermal load prediction curve next day respectively, at prediction day proxima luce (prox. luc), according to photovoltaic generating system time daily output prediction curve, interruptible load work program next day data, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day, calculating obtains removable load operation curve next day and storage-type load operation curve next day is: according to photovoltaic generating system time daily output prediction curve, interruptible load work program next day data, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day, use the model based on genetic algorithm, calculate removable load operation curve next day respectively, cold energy storage load operation curve next day and hot energy storage load operation curve next day, predicting same day day, according to photovoltaic generating system when daily output prediction curve, interruptible load work program on same day data, uncontrollable load prediction curve on the same day, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, power under the comfort level sacrifice that the Demand Side Response information issued on Utilities Electric Co. same day and power-adjustable cooling and heating load allowed the same day reduces amplitude, the start working time period value of the calculating removable load of acquisition under Demand Side Response, the beginning power stage time period value of storage-type load under Demand Side Response, and interruptible load load decreasing value under Demand Side Response, and by result of calculation with curve form output be: according to photovoltaic generating system when daily output prediction curve, interruptible load work program on same day data, uncontrollable load prediction curve on the same day, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, the Demand Side Response information that Utilities Electric Co. issued the same day, use the model based on genetic algorithm, and add and reduce with the power under the comfort level sacrifice allowed on power-adjustable cooling and heating load same day the correction that amplitude is penalty function and affect, calculate the start working time period value of the removable load of acquisition under Demand Side Response, the beginning power stage time period value of storage-type load under Demand Side Response, and interruptible load load decreasing value under Demand Side Response, and result of calculation is exported with curve form.
Further, based on the model of genetic algorithm be the following fitness function model of genetic algorithm:
Wherein, fit (z) is fitness function, and z is decision variable, and C (z) is the purchases strategies under decision variable, C
0for optimizing front purchases strategies;
Correction impact based on the penalty function of the model of genetic algorithm is following penalty function model:
Wherein, f (z)=weight × Σ price
i× loadbeyond
i, wherein, z is decision variable, and weight is weight coefficient, price
ifor i period electricity price, loadbeyond
ifor i period out-of-limit value.
Above-mentioned use based on gray model Forecasting Methodology, use based on gray model and neural network ensemble Forecasting Methodology, use based on the model of genetic algorithm and affiliated use based on the model of genetic algorithm and the correction that the power added under the comfort level sacrifice allowed the same day with power-adjustable cooling and heating load reduces the determination fitness function that amplitude is penalty function affect and calculate, for calculating for final goal so that purchases strategies is minimum.
Above-mentioned all number of effective points certificates calculating the curve obtained are spaced apart 15 minutes.
According to another aspect of the present invention, provide a kind of energy management system of distribution garden, comprising:
Load classification module, all power loads in distribution garden, according to the characteristic of power load each in distribution garden, are divided into storage-type load, removable load, interruptible load, power-adjustable cooling and heating load and uncontrollable load by load classification module;
Prediction module, prediction module is at prediction day proxima luce (prox. luc), obtain temperature prediction curve next day, uncontrollable demand history data and power-adjustable cooling and heating load historical data respectively, and according to next day temperature prediction curve, uncontrollable demand history power data and power-adjustable cooling and heating load historical power data, calculate and obtain uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day;
Plan optimization module a few days ago, plan optimization module is at prediction day proxima luce (prox. luc) a few days ago, obtain photovoltaic generating system time daily output prediction curve respectively, interruptible load work program next day data, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, and according to photovoltaic generating system time daily output prediction curve, interruptible load work program next day data, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day, calculate and obtain removable load operation curve next day and storage-type load operation curve next day,
In a few days demand response module, in a few days demand response module is predicting same day day, the power obtained respectively under the comfort level sacrifice that the Demand Side Response information issued on Utilities Electric Co. same day and power-adjustable cooling and heating load allowed the same day reduces amplitude, and according to photovoltaic generating system when daily output prediction curve, interruptible load work program on same day data, uncontrollable load prediction curve on the same day, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, power under the Demand Side Response information that Utilities Electric Co. same day is issued and the comfort level sacrifice that power-adjustable cooling and heating load allowed the same day reduces amplitude, calculate and obtain the start working time period value of removable load under Demand Side Response, the beginning power stage time period value of storage-type load under Demand Side Response, and interruptible load load decreasing value under Demand Side Response, and result of calculation is exported with curve form,
The horizontal assessment module of planning optimization a few days ago, the horizontal assessment module of planning optimization is after prediction day a few days ago, obtain photovoltaic generating system actual power curve on the same day, interruptible load actual electricity consumption curve on the same day, removable load actual electricity consumption curve on the same day, storage-type load actual operation curve on the same day, uncontrollable load actual electricity consumption curve on the same day, power-adjustable cooling and heating load actual electricity consumption curve on the same day, and perform all the actual electricity consumption curve produced according to optimisation strategy, actual power curve, actual operation curve with calculate without the comparing property of electricity consumption situation under optimisation strategy,
The in a few days horizontal assessment module of demand response, in a few days the horizontal assessment module of demand response is after prediction day, and the electricity consumption situation under the actual electricity consumption curve of the interruptible load produced according to Demand Side Response strategy execution according to all, removable load, storage-type load and power-adjustable cooling and heating load and actual operation curve and non-reference requirement respond compares comparative calculating.
Wherein, storage-type load comprises electric energy storage load and cold and hot energy storage load, and wherein, cold and hot energy storage load comprises cold energy storage load and hot energy storage load; Power-adjustable cooling and heating load comprises power-adjustable refrigeration duty and power-adjustable thermal load; At prediction day proxima luce (prox. luc), calculate and obtain power-adjustable cooling and heating load prediction curve next day, for calculating power-adjustable refrigeration duty prediction curve next day and power-adjustable thermal load prediction curve next day respectively; At prediction day proxima luce (prox. luc), calculating and obtain storage-type load operation curve next day, for calculating cold and hot energy storage load operation curve next day, comprising cold energy storage load operation curve next day and hot energy storage load operation curve next day; Predicting same day day, calculating and obtain the beginning power stage time period value of storage-type load under Demand Side Response, for calculating electric energy storage load, cold energy storage load and the hot energy storage load beginning power stage time period value under Demand Side Response respectively.
And, at prediction day proxima luce (prox. luc), according to next day temperature prediction curve, uncontrollable demand history power data and power-adjustable cooling and heating load historical power data, calculate and obtain uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day and be: according to next day temperature prediction curve with uncontrollable demand history power data, use uncontrollable load prediction curve next day of Forecasting Methodology calculating acquisition based on gray model, and according to temperature prediction curve and power-adjustable cooling and heating load historical power data next day, use the Forecasting Methodology based on gray model and neural network ensemble to calculate power-adjustable refrigeration duty prediction curve next day and power-adjustable thermal load prediction curve next day respectively, at prediction day proxima luce (prox. luc), according to photovoltaic generating system time daily output prediction curve, interruptible load work program next day data, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day, calculating obtains removable load operation curve next day and storage-type load operation curve next day is: according to photovoltaic generating system time daily output prediction curve, interruptible load work program next day data, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day, use the model based on genetic algorithm, calculate removable load operation curve next day respectively, cold energy storage load operation curve next day and hot energy storage load operation curve next day, predicting same day day, according to photovoltaic generating system when daily output prediction curve, interruptible load work program on same day data, uncontrollable load prediction curve on the same day, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, power under the comfort level sacrifice that the Demand Side Response information issued on Utilities Electric Co. same day and power-adjustable cooling and heating load allowed the same day reduces amplitude, the start working time period value of the calculating removable load of acquisition under Demand Side Response, the beginning power stage time period value of storage-type load under Demand Side Response, and interruptible load load decreasing value under Demand Side Response, and by result of calculation with curve form output be: according to photovoltaic generating system when daily output prediction curve, interruptible load work program on same day data, uncontrollable load prediction curve on the same day, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, the Demand Side Response information that Utilities Electric Co. issued the same day, use the model based on genetic algorithm, and add and reduce with the power under the comfort level sacrifice allowed on power-adjustable cooling and heating load same day the correction that amplitude is penalty function and affect, calculate the start working time period value of the removable load of acquisition under Demand Side Response, the beginning power stage time period value of storage-type load under Demand Side Response, and interruptible load load decreasing value under Demand Side Response, and result of calculation is exported with curve form.
Further, based on the model of genetic algorithm be the following fitness function model of genetic algorithm:
Wherein, fit (z) is fitness function, and z is decision variable, and C (z) is the purchases strategies under decision variable, C
0for optimizing front purchases strategies;
Correction impact based on the penalty function of the model of genetic algorithm is following penalty function model:
Wherein, f (z)=weight × Σ price
i× loadbeyond
i, wherein, z is decision variable, and weight is weight coefficient, price
ifor i period electricity price, loadbeyond
ifor i period out-of-limit value.
Above-mentioned use based on gray model Forecasting Methodology, use based on gray model and neural network ensemble Forecasting Methodology, use based on the model of genetic algorithm and affiliated use based on the model of genetic algorithm and the correction that the power added under the comfort level sacrifice allowed the same day with power-adjustable cooling and heating load reduces the determination fitness function that amplitude is penalty function affect and calculate, for calculating for final goal so that purchases strategies is minimum.
Above-mentioned all number of effective points certificates calculating the curve obtained are spaced apart 15 minutes.
As can be seen from above, technical scheme provided by the invention with in practical application typical case intelligence adapted electricity garden for research object, by carrying out analyzing to part throttle characteristics all kinds of in garden and concluding, propose a kind of based on gray model, the energy management method of neural network and combinations genetic algorithms, the method can the optimal operation model of each type load in garden under reasonable construction dsm condition, and by genetic algorithm, problem is solved, thus realize peak load shifting, strengthen demand response analysis and Control and efficiency coordination optimization, reach the object reducing electric cost, for distribution garden provides the total solution based on comprehensive diversity load efficiency optimization and intelligent demand response.
Embodiment
Clearly understand for making the object, technical solutions and advantages of the present invention, below in conjunction with the accompanying drawing in the embodiment of the present invention, to the technical scheme in the embodiment of the present invention carry out further clear, complete, describe in detail, obviously, described embodiment is only the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, the every other embodiment that those of ordinary skill in the art obtain, all belongs to the scope of protection of the invention.
According to one embodiment of present invention, a kind of energy management method of distribution garden is provided.
As shown in Figure 1, the energy management method of the distribution garden provided according to the embodiment of the present invention comprises:
All power loads in distribution garden, according to the characteristic of power load each in distribution garden, are divided into storage-type load, removable load, interruptible load, power-adjustable cooling and heating load and uncontrollable load by step S101;
Step S103, at prediction day proxima luce (prox. luc), obtain temperature prediction curve next day, uncontrollable demand history data and power-adjustable cooling and heating load historical data respectively, and according to next day temperature prediction curve, uncontrollable demand history power data and power-adjustable cooling and heating load historical power data, calculate and obtain uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day;
Step S105, at prediction day proxima luce (prox. luc), obtain photovoltaic generating system time daily output prediction curve respectively, interruptible load work program next day data, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, and according to photovoltaic generating system time daily output prediction curve, interruptible load work program next day data, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day, calculate and obtain removable load operation curve next day and storage-type load operation curve next day,
Step S107, predicting same day day, the power obtained respectively under the comfort level sacrifice that the Demand Side Response information issued on Utilities Electric Co. same day and power-adjustable cooling and heating load allowed the same day reduces amplitude, and according to photovoltaic generating system when daily output prediction curve, interruptible load work program on same day data, uncontrollable load prediction curve on the same day, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, power under the Demand Side Response information that Utilities Electric Co. same day is issued and the comfort level sacrifice that power-adjustable cooling and heating load allowed the same day reduces amplitude, calculate and obtain the start working time period value of removable load under Demand Side Response, the beginning power stage time period value of storage-type load under Demand Side Response, and interruptible load load decreasing value under Demand Side Response, and result of calculation is exported with curve form,
Step S109, after prediction day, obtain photovoltaic generating system actual power curve on the same day, interruptible load actual electricity consumption on same day curve, removable load actual electricity consumption on same day curve, storage-type load actual operation curve on the same day, uncontrollable load actual electricity consumption on same day curve, power-adjustable cooling and heating load actual electricity consumption curve on the same day, and by all actual electricity consumption curve, actual power curve, actual operation curve according to optimisation strategy execution generation and calculate without the comparing property of electricity consumption situation under optimisation strategy;
Step S111, after prediction day, the electricity consumption situation under the actual electricity consumption curve of the interruptible load produced according to Demand Side Response strategy execution according to all, removable load, storage-type load and power-adjustable cooling and heating load and actual operation curve and non-reference requirement respond compares comparative calculating.
Wherein, storage-type load comprises electric energy storage load and cold and hot energy storage load, and wherein, cold and hot energy storage load comprises cold energy storage load and hot energy storage load; Power-adjustable cooling and heating load comprises power-adjustable refrigeration duty and power-adjustable thermal load; At prediction day proxima luce (prox. luc), calculate and obtain power-adjustable cooling and heating load prediction curve next day, for calculating power-adjustable refrigeration duty prediction curve next day and power-adjustable thermal load prediction curve next day respectively; At prediction day proxima luce (prox. luc), calculating and obtain storage-type load operation curve next day, for calculating cold and hot energy storage load operation curve next day, comprising cold energy storage load operation curve next day and hot energy storage load operation curve next day; Predicting same day day, calculating and obtain the beginning power stage time period value of storage-type load under Demand Side Response, for calculating electric energy storage load, cold energy storage load and the hot energy storage load beginning power stage time period value under Demand Side Response respectively.
For ease of Taxonomic discussion, can be classified according to the characteristic difference of load.All loads can be divided into storage-type load and non-memory type load two class.
Wherein, storage-type load comprises electric energy storage load, cold energy storage load and hot energy storage load, and the general character of storage-type load comprises two kinds of duties, i.e. energy storing state and fault offset state.Electricity energy storage load storage of electrical energy, exporting also is electric energy; Electric energy conversion is interior can also storage by hot energy storage load, exports as interior energy; Electric energy conversion is cold and is stored by cold energy storage load, exports as cold.
Non-memory type load can be classified as follows:
Removable, non-adjustable large Smaller load (being called for short removable load), is classified as this type of by electric automobile, laundry, not considered critical working time equipment (as VMC (Ventilation Mechanical Control System) partial power equipment) etc. under study for action.The feature of this type load is that the shape size of load (on lasting time and each time point) of working curve is fixing, initial working time adjustable;
Irremovable, interruptible price, non-adjustable large Smaller load (abbreviation interruptible load), be classified as this type of by small processing factory in research.The feature of this type load can cut down part or all of load in certain short period length, but will the load of cutting down is supplemented by certain period afterwards.
Irremovable, not interruptible price, fine-tuning large Smaller load (be called for short power-adjustable cooling and heating load), be classified as this type of by air-conditioning system (comprising refrigeration duty and power-equipment load), hot-water heating system (comprising thermal load and power-equipment load) in research.The feature of this type load can cut down a small amount of load in certain short period length, and the load of reduction is follow-up without the need to supplementing.
Finally, the illumination of office buildings, VMC (Ventilation Mechanical Control System) partial power equipment, residential building (comprising illumination, air-conditioning, refrigerator, TV etc.) are classified as this type of in research by irremovable, not interruptible price, non-adjustable large Smaller load (being called for short uncontrollable load).The feature of this type load unconditionally to meet, without the space of any adjustment.
And, at prediction day proxima luce (prox. luc), according to next day temperature prediction curve, uncontrollable demand history power data and power-adjustable cooling and heating load historical power data, calculate and obtain uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day and be: according to next day temperature prediction curve with uncontrollable demand history power data, use uncontrollable load prediction curve next day of Forecasting Methodology calculating acquisition based on gray model, and according to temperature prediction curve and power-adjustable cooling and heating load historical power data next day, use the Forecasting Methodology based on gray model and neural network ensemble to calculate power-adjustable refrigeration duty prediction curve next day and power-adjustable thermal load prediction curve next day respectively, at prediction day proxima luce (prox. luc), according to photovoltaic generating system time daily output prediction curve, interruptible load work program next day data, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day, calculating obtains removable load operation curve next day and storage-type load operation curve next day is: according to photovoltaic generating system time daily output prediction curve, interruptible load work program next day data, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day, use the model based on genetic algorithm, calculate removable load operation curve next day respectively, cold energy storage load operation curve next day and hot energy storage load operation curve next day, predicting same day day, according to photovoltaic generating system when daily output prediction curve, interruptible load work program on same day data, uncontrollable load prediction curve on the same day, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, power under the comfort level sacrifice that the Demand Side Response information issued on Utilities Electric Co. same day and power-adjustable cooling and heating load allowed the same day reduces amplitude, the start working time period value of the calculating removable load of acquisition under Demand Side Response, the beginning power stage time period value of storage-type load under Demand Side Response, and interruptible load load decreasing value under Demand Side Response, and by result of calculation with curve form output be: according to photovoltaic generating system when daily output prediction curve, interruptible load work program on same day data, uncontrollable load prediction curve on the same day, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, the Demand Side Response information that Utilities Electric Co. issued the same day, use the model based on genetic algorithm, and add and reduce with the power under the comfort level sacrifice allowed on power-adjustable cooling and heating load same day the correction that amplitude is penalty function and affect, calculate the start working time period value of the removable load of acquisition under Demand Side Response, the beginning power stage time period value of storage-type load under Demand Side Response, and interruptible load load decreasing value under Demand Side Response, and result of calculation is exported with curve form.
We claim, and be planning optimization a few days ago in the work of prediction day proxima luce (prox. luc), and the work done same day day in prediction is in a few days demand response.
Planning optimization refers to a few days ago to predicting that the operation action of the in a few days administrative each load of research object or device is optimized a few days ago, seeks meeting the operational plan that under Electrical Safety prerequisite, electric cost is minimum.Under the condition performing time-of-use tariffs, crest segment electricity price is apparently higher than paddy section electricity price, therefore can predict, the operational plan that electric cost is minimum must be achieve power load to the full extent from crest segment to the transfer of paddy section, and this is on all four with the target of carrying out DSM Work.But the transfer of load from peak to paddy is not unconfined, needs the constraint of acceptor's mains supply power, also to consider the load electrical characteristics of distribution garden resident and the requirement for comfort level simultaneously.
Target: purchases strategies is minimum, namely
Wherein,
it is the output power of main electrical network t period; R
tfor the electricity price of t period; Δ t is the time span of minimum period, namely 15 minutes.
In the load operation planning process that in a few days demand response is formulated before then referring to the research object execution day, when after the peak clipping demand (i.e. aforesaid Demand Side Response demand) obtaining grid company issue, suboptimization is again carried out to the operation action of load each in system or device, under the condition counting peak clipping compensation income, electricity consumption total cost is in a few days made to reach minimum.
Target: purchases strategies is minimum, namely
Wherein,
it is the output power of main electrical network t period; R
tfor the electricity price of t period; Δ t is the time span of minimum period, namely 15 minutes; Te is the start periods of demand response time window; Ts is the processing completion time used for them of demand response time window;
for the output power of main electrical network t period obtained by planning optimization a few days ago; W is that the demand response that grid company is issued compensates electricity price.
Computing method involved by technical scheme of the present invention comprise Grey Models, Artificial Neural Network, genetic algorithm and genetic algorithm are carried out to the penalty function algorithm of convergence correction.
Grey Models is a kind of load forecasting method of being used widely, and has the advantage that modeling information needed is less, modeling accuracy is higher.
Compared to Grey Models, Artificial Neural Network overcomes the deficiency of classic method in process nonlinear problem, the impact of correlative factor on load can be taken into full account, be more suitable for the prediction curve calculating the larger power-adjustable cooling and heating load of nonlinear degree.
Comprehensive use Grey Models and the embodiment of Artificial Neural Network as follows:
First, carry out frequency domain decomposition to historical load sequence and obtain taking week as the component in cycle, deducting from original loads sequence with week is the component in cycle, obtains without cycle sequence;
Secondly, for without cycle sequence, to each moment point, set up GM (1,1) model, and calculate history matching value and forecasted future value, add with week to be the component in cycle after having calculated, obtain gray model match value and the predicted value of original series;
Again, selected reference day, then to each moment point of a day, set up neural network, containing 5 input nodes, wherein 1 is the difference of gray model match value and Base day load, other 4 are respectively maximum temperature, minimum temperature, medial temperature and the date type variable quantity relative to the Base day, output node 1, is the change of actual load relative datum daily load, trains network;
Finally, by the temperature prediction value of day to be predicted, the Grey Model value of load, deduct input neural network after Base day corresponding amount, the predicted value of output load variable quantity, adds that Base day load obtains the final predicted value of load.
Genetic algorithm is a kind of direct search method not relying on particular problem.Before execution algorithm, first provide a group " chromosome ", that is to say initial population.Then, among " environment " that this population be placed in problem, and by the principle of the survival of the fittest, therefrom select " chromosome " that relatively more conforms and copy.A new generation " chromosome " group is produced again by intersection, mutation process.Constant evolution goes down like this, finally converges to the individuality of " conforming most ".
When using genetic algorithm for solving, first determine decision variable span.
Wherein, removable load operation start periods span is determined according to staff's work habit, regards as basic data; The determination of storage-type load operation start periods span needs the effect playing storage-type device to greatest extent, will at trough section storing electrical energy, interior energy and cold to greatest extent, at crest segment small size electric energy, interior energy and the cold saved to greatest extent, store full as far as possible before paddy section terminates, before crest segment terminates, export least residue capacity to.Calculate work in peak, the paddy mode of the span starting the period of storage-type load as follows:
A () is according to the strategy of " storing full before paddy section terminates; export least residue capacity to before crest segment terminates " as far as possible as far as possible, hop count when calculating shared by the running status of 3 kinds of storage-type loads in peak, paddy, hop count when can calculate the non-operating state of 3 kinds of storage-type loads in peak, paddy accordingly.
B () obtains according to hop count during the non-operating state of 3 kinds of storage-type loads in peak, paddy the span that it runs start periods in peak, paddy.In one embodiment, originated in for the 25th period, end at the crest of 88 periods, when it is total, hop count is 88-25+1=64; It is 50 that one storage-type load is full of from energy the time hop count being discharged into zero needs, then this storage-type load hop count when the non-operating state of crest is 64-50=14, and then obtain its crest segment and run start periods span and be: [25,38] (38=25+14-1).
By using above-mentioned method, the operation interval that its stored energy can be provided, release energy for storage-type load, and then draw its prediction curve.
Further, based on the model of genetic algorithm be the following fitness function model of genetic algorithm:
Wherein, fit (z) is fitness function, and z is decision variable, and C (z) is the purchases strategies under decision variable, C
0for optimizing front purchases strategies.
When needing to be revised the convergence of above-mentioned fitness function, penalty function can be used.Correction impact based on the penalty function of the model of genetic algorithm is following penalty function model:
Wherein, f (z)=weight × Σ price
i× loadbeyond
i, wherein, z is decision variable, and weight is weight coefficient, price
ifor i period electricity price, loadbeyond
ifor i period out-of-limit value.
Except penalty function, following three kinds of methods can also be used to improve the convergence of genetic algorithm: to control evolution population size---strengthen population size, along with the increase scale of evolutionary generation is gradually reduced to constant scale at the evolution initial stage; To the improvement of mutation operator---improve the adaptability of mutation operator, increase mutation probability, along with the increase of evolutionary generation is reduced to constant gradually in evolution initial stage appropriateness; The chaos optimization of excellent individual---the some individuals in addition chaotic disturbance high to fitness, impels genetic algorithm to jump out local optimum with this.
Above-mentioned use based on gray model Forecasting Methodology, use based on gray model and neural network ensemble Forecasting Methodology, use based on the model of genetic algorithm and affiliated use based on the model of genetic algorithm and the correction that the power added under the comfort level sacrifice allowed the same day with power-adjustable cooling and heating load reduces the determination fitness function that amplitude is penalty function affect and calculate, for calculating for final goal so that purchases strategies is minimum.
This method has carried out simulation calculation by test macro to model and algorithm.Result of calculation shows, after optimizing, next day, out-of-limit situation that total load curve occurs at partial period was eliminated, and " peak load shifting " is respond well simultaneously, and total electric cost obviously reduces.
Above-mentioned all collections, with calculate obtains curve number of effective points certificate be spaced apart 15 minutes.
The curve that all calculating obtains, all by data output interface, exports with the form of data file.
The data of above-mentioned all collections, for gathering the data of each type load of each user in intelligent power distribution garden by user's efficiency control of intelligent terminal.
According to another embodiment of the invention, a kind of energy management system of distribution garden is provided.
As shown in Figure 2, the energy management system of the distribution garden provided according to the embodiment of the present invention comprises:
Load classification module 21, all power loads in distribution garden, according to the characteristic of power load each in distribution garden, are divided into storage-type load, removable load, interruptible load, power-adjustable cooling and heating load and uncontrollable load by load classification module 21;
Prediction module 22, prediction module 22 is at prediction day proxima luce (prox. luc), obtain temperature prediction curve next day, uncontrollable demand history data and power-adjustable cooling and heating load historical data respectively, and according to next day temperature prediction curve, uncontrollable demand history power data and power-adjustable cooling and heating load historical power data, calculate and obtain uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day;
Plan optimization module 23 a few days ago, plan optimization module 23 is at prediction day proxima luce (prox. luc) a few days ago, obtain photovoltaic generating system time daily output prediction curve respectively, interruptible load work program next day data, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, and according to photovoltaic generating system time daily output prediction curve, interruptible load work program next day data, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, uncontrollable load prediction curve next day and power-adjustable cooling and heating load prediction curve next day, calculate and obtain removable load operation curve next day and storage-type load operation curve next day,
In a few days demand response module 24, in a few days demand response module 24 is predicting same day day, the power obtained respectively under the comfort level sacrifice that the Demand Side Response information issued on Utilities Electric Co. same day and power-adjustable cooling and heating load allowed the same day reduces amplitude, and according to photovoltaic generating system when daily output prediction curve, interruptible load work program on same day data, uncontrollable load prediction curve on the same day, removable load electrical characteristic data, storage-type load operation supplemental characteristic, maximum electric power restriction, electricity price information, power under the Demand Side Response information that Utilities Electric Co. same day is issued and the comfort level sacrifice that power-adjustable cooling and heating load allowed the same day reduces amplitude, calculate and obtain the start working time period value of removable load under Demand Side Response, the beginning power stage time period value of storage-type load under Demand Side Response, and interruptible load load decreasing value under Demand Side Response, and result of calculation is exported with curve form,
The horizontal assessment module 25 of planning optimization a few days ago, the horizontal assessment module 25 of planning optimization is after prediction day a few days ago, obtain photovoltaic generating system actual power curve on the same day, interruptible load actual electricity consumption curve on the same day, removable load actual electricity consumption curve on the same day, storage-type load actual operation curve on the same day, uncontrollable load actual electricity consumption curve on the same day, power-adjustable cooling and heating load actual electricity consumption curve on the same day, and perform all the actual electricity consumption curve produced according to optimisation strategy, actual power curve, actual operation curve with calculate without the comparing property of electricity consumption situation under optimisation strategy,
The in a few days horizontal assessment module 26 of demand response, in a few days the horizontal assessment module 26 of demand response is after prediction day, and the electricity consumption situation under the actual electricity consumption curve of the interruptible load produced according to Demand Side Response strategy execution according to all, removable load, storage-type load and power-adjustable cooling and heating load and actual operation curve and non-reference requirement respond compares comparative calculating.
In sum, by means of the technical solution of the present invention, with the typical case's intelligence adapted electricity garden in practical application for research object, by carrying out analyzing to part throttle characteristics all kinds of in garden and concluding, a kind of energy management method based on gray model, neural network and combinations genetic algorithms is proposed, the method can the optimal operation model of each type load in garden under reasonable construction dsm condition, minimum for target with garden total electric cost next day, strict implement time-of-use tariffs, and guarantee to meet mains supply power constraint; And by genetic algorithm, problem is solved, thus realize peak load shifting, strengthen demand response analysis and Control and efficiency coordination optimization, reach the object reducing electric cost, for distribution garden provides the total solution based on comprehensive diversity load efficiency optimization and intelligent demand response.
Those of ordinary skill in the field are to be understood that: the foregoing is only specific embodiments of the invention; be not limited to the present invention; within the spirit and principles in the present invention all, any amendment made, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.