DE102011082840A1 - Method for controlling loads of power network, involves molding loads and determining objective function for predetermined prediction time and considering loads as constraints during optimization of modeling of loads - Google Patents

Abstract

The method involves molding loads (102-106) and determining an objective function for predetermined prediction time. The loads are considered as constraints during optimization of the modeling of loads. Each load is modeled by temporal evolution of an energy store. The objective function is minimized in the context of optimization, and deviation of total load is penalized by a target load. An individual load is switched based on determined optimization, where objective function has square error that is minimized by optimization process. Independent claims are also included for the following: (1) a device for controlling consumers of a power network (2) a system comprising a device for controlling consumers of a power network.

Description

The invention relates to a method and a device for controlling consumers of a power grid. Also, a corresponding system comprising at least one such device is proposed.

In a future smart grid, the intention is to use the feed-in of regenerative energies by adapting the demand to the supply. Further motives for a load management are a high flexibility of the load to avoid load peaks as well as a reduced fluctuation of the load curves and thus a lower demand for control energy.

• (1) eine direkte Steuerung,
• (2) eine Steuerung über Preisanreizsignale sowie
• (3) eine marktbasierte Interaktionen, z.B. Auktionen.

For load management, there are essentially the following interactions between management and consumers:

• (1) a direct control,
• (2) control over price incentive signals as well
• (3) market-based interactions, eg auctions.

Known approaches are aimed at controlling household appliances, e.g. Refrigerators in private apartments or houses. For such consumers, the store is charged as soon as it is "empty" (in the case of the refrigerator: cooling on reaching an upper threshold temperature) and discharged as soon as it is "full" (in the case of the refrigerator: reaching a lower threshold temperature).

With the load management can be achieved that the energy storage is charged or discharged before it is "empty" or "full". This makes it possible to shift the time of electrical consumption.

[ Stadler 2007 ] describes a study in which a large number of consumers are influenced by a direct control or price signal. The large number of consumers statistically reduces the relative fluctuations in overall consumption.

[ Chef 2009 ] describes a price-based load management method in which the loads report a price to the management system at each time interval, which decreases the shorter the time it takes the memory to reach one of the thresholds. The management system then determines a price threshold so that a certain total consumption is achieved when consumers then switch according to the price signal.

The object of the invention is to improve the above-mentioned approaches and in particular to provide an efficient way of controlling consumers.

This object is achieved according to the features of the independent claims. Preferred embodiments are in particular the dependent claims.

• – bei dem die Verbraucher modelliert werden,
• – bei dem eine Zielfunktion für eine vorgegebene Prognosedauer bestimmt wird,
• – bei dem die Zielfunktion optimiert wird, wobei bei der Optimierung die Modellierung der Verbraucher als Nebenbedingungen berücksichtigt werden.

To solve the problem, a method for controlling consumers of a power grid is proposed,

• - where consumers are modeled,
• In which an objective function is determined for a predefined prognosis duration,
• - In which the objective function is optimized, with the optimization of the modeling of consumers are considered as constraints.

It is advantageous that a target load profile can be achieved without violating predetermined secondary conditions.

By means of this optimization method, predetermined load curves can be readjusted or adjusted well even with a few consumers, and fluctuations in the load curve can be significantly reduced.

It should be noted that the consumer can also be referred to as an energy store or comprises at least one energy store.

The consumer is e.g. switched on, if a control variable reaches or exceeds an upper threshold, and it is e.g. switched off if the control variable reaches or falls below a lower threshold value. The thresholds may be provided with hysteresis to prevent unwanted switching between on and off states.

It is advantageous that the control of the consumer takes place in a forward-looking manner and thus a significant improvement in the load management can be achieved.

A development is that the consumers are modeled by modeling a temporal evolution of at least one energy storage of each consumer.

The evolution of the energy storage over time may include a consumer load component and a consumer discharge component. Both components can be considered as terms of temporal evolution.

wobei

n

den Verbraucher,

E rel / n

einen normierten Speicherinhalt pro Verbrau cher n,

E amb / n

eine Außentemperatur, die dafür sorgt, dass der Verbraucher n entladen wird (Wärmeverlust),

u_n

eine Stellgröße,

τ_n

eine Zeitkonstante für die Entladung,

k_n

eine Kopplungsstärke,

υ_n

einen Zufallsprozess

bezeichnen. An exemplary development is that the temporal development of a consumer is modeled as follows:

in which

n

the consumer,

e rel / n

a normalized memory content per consumer,

e amb / n

an outside temperature which ensures that the consumer n is discharged (heat loss),

u _n

a manipulated variable,

τ _n

a time constant for the discharge,

k _n

a coupling strength,

υ _n

a random process

describe.

It is also a further development that the temporal development of the consumer is modeled (eg estimated) on the basis of data available online, in particular by the parameters τ _n , k _n and E amb / n based on readings for E rel / n and u _n be determined.

Preferably, a Kalman filter can be used for this purpose.

In particular, it is a development that the objective function is minimized as part of the optimization, which punishes a deviation of a total load of a target load.

It is also a development that is determined by the optimization, when a single consumer is switched.

Switching involves either switching on the load or switching off. Optionally, different intermediate stages of switching on and off can be realized.

mit einer Randbedingung 0 < E rel / n,t < 1 bestimmt wird, wann der Verbraucher n umgeschaltet wird, wobei

P tar / t

einen vorgegebenen Summenlastverbrauch zu ei ner Zeit t,

P rated / n

einen Energieverbrauch eines Verbrauchers n, wenn dieser lädt (z.B. u = 1),

T

ein Zeitintervall einer Prognose

bezeichnen. Furthermore, it is a development that based on a manipulated variable

with a boundary condition 0 <E rel / n, t <1 it is determined when the consumer n is switched, where

P tar / t

a given total load consumption at a time t,

P rated / n

an energy consumption of a consumer n when it is charging (eg u = 1),

T

a time interval of a forecast

describe.

In an additional development, the optimization is performed by determining at a time at one end of a time interval for each consumer how the target function would change if the consumer were switched at that time.

A next development is that the objective function includes a quadratic error and the optimization minimizes the quadratic error.

angegeben wird. An embodiment is that as a target function

is specified.

An alternative embodiment is that the one consumer is switched, which achieves the greatest improvement of the objective function.

A next embodiment is that the optimization is performed until it has converged or a predetermined number of iterations has been reached.

• – die Verbraucher modellierbar sind,
• – eine Zielfunktion für eine vorgegebene Prognosedauer bestimmbar ist,
• – die Zielfunktion optimierbar ist, wobei bei der Optimierung die Modellierung der Verbraucher als Nebenbedingungen berücksichtigt werden.

The above object is also achieved by a device for controlling consumers of a power network, comprising a processing unit which is set up such that

• - the consumers are modelable,
• An objective function can be determined for a predefined duration of the prognosis,
• - The objective function can be optimized, with the optimization of the modeling of consumers are considered as constraints.

The processing unit mentioned here may in particular be embodied as a processor unit and / or an at least partially hard-wired or logical circuit arrangement, which is set up, for example, in such a way that the method can be carried out as described herein. Said processing unit may include any type of processor or computer or computer necessary peripherals (memory, input / output interfaces, I / O devices, etc.).

The above explanations regarding the method apply to the device accordingly. The device may be implemented in one component or distributed in several components.

A development is that the device is part of a load management or load management system.

Another development is that the device comprises a configuration unit, by means of which parameters for the optimization can be set.

Also, the above object is achieved by means of a system comprising at least one of the devices described here.

In particular, the system may be configured such that the device receives status information from the consumers and performs control of the consumers based on the status information.

The solution presented here further comprises a computer program product which can be loaded directly into a memory of a digital computer, comprising program code parts which are suitable for performing steps of the method described here.

Furthermore, the above problem is solved by means of a computer-readable storage medium, e.g. any memory comprising computer-executable instructions (e.g., in the form of program code) adapted for the computer to perform steps of the method described herein.

The above-described characteristics, features, and advantages of this invention, as well as the manner in which they will be achieved, will become clearer and more clearly understood in connection with the following schematic description of exemplary embodiments which will be described in detail in conjunction with the drawings. In this case, the same or equivalent elements may be provided with the same reference numerals for clarity.

It shows:

1 an exemplary arrangement with a central unit (balancing or load management unit) for the coordination of multiple consumers.

In the following, in particular the interaction (1), "direct control" will be discussed in more detail. Preferably, this direct control is designed anticipatory, so that energy storage can be prepared for extreme requirements.

There are different consumers (devices) that need electrical energy. These electrical consumers are e.g. switched on, if a control variable reaches or exceeds an upper threshold and they are e.g. switched off if the control variable reaches or falls below a lower threshold value. The thresholds may be provided with hysteresis to prevent unwanted switching between on and off states.

One goal now is to control the devices so that a target load profile is achieved without violating given constraints.

The present load management problem can be formulated as a discrete optimization problem. In this case, a target function is to be minimized, which penalizes the deviation of the total load curve from a target load curve.

A consumer's time evolution can be modeled as follows:

n

einen Energiespeicher (Verbraucher),

E rel / n

einen normierten (relativen) Speicherinhalt pro Energiespeicher n, wobei der normierte Speicherinhalt in einem Intervall zwischen 0 und 1 notiert werden kann,

E amb / n

eine Außentemperatur, die dafür sorgt, dass der Energiespeicher n entladen wird (Wärmeverlust),

u_n

eine Stellgröße, z.B. "0" für "Entladen" und "1" für "Laden" des Energiespeichers n,

τ_n

eine Zeitkonstante für die Entladung des Ener giespeichers n,

k_n

eine Kopplungsstärke (je größer die Kopplung, desto schneller wird der Energiespeicher n geladen),

υ_n

einen Zufallsprozess, der die Entladerate mo difiziert (in der Praxis kann hierfür eine Prognose eingesetzt werden).

Where:

n

an energy store (consumer),

e rel / n

a normalized (relative) memory content per energy store n, wherein the normalized memory content can be noted in an interval between 0 and 1,

e amb / n

an outside temperature which ensures that the energy store n is discharged (heat loss),

u _n

a manipulated variable, eg "0" for "discharging" and "1" for "charging" the energy store n,

τ _n

a time constant for the discharge of the energy storage n,

k _n

a coupling strength (the larger the coupling, the faster the energy storage n is charged),

υ _n

a random process that modifies the discharge rate (in practice, a forecast can be used for this).

mit der Randbedingung 0 < E rel / n,t < 1. An optimization should now show when the consumer n must be switched, ie it should be the manipulated variable u * / n, t be true

with the boundary condition 0 <E rel / n, t <1.

According to the boundary condition, the memory switches n if it is full.

u * / n,t

die zeitabhängige Stellgröße für den Energie speicher n,

P tar / t

einen vorgegebenen Summenlastverbrauch zu ei ner Zeit t,

P rated / n

einen Energieverbrauch des Energiespeichers n, wenn dieser lädt (z.B. u = 1),

T

ein Zeitintervall der Prognose.

They denote:

u * / n, t

the time-dependent manipulated variable for the energy storage n,

P tar / t

a given total load consumption at a time t,

P rated / n

an energy consumption of the energy store n when it is charging (eg u = 1),

T

a time interval of the prognosis.

The optimization is achieved by determining for each consumer, at a time t beginning at the end of the entire considered time interval, how a target function would change if the consumer were switched over at time t.

durch Umschalten erreicht, wird auch umgeschaltet; falls keine Verbesserung erzielbar ist, wird nicht umgeschaltet. The consumer, the biggest improvement of the objective function, for example, the following quadratic error

achieved by switching is also switched; if no improvement can be achieved, it will not switch.

This process is repeated until the process converges or a preset number of iterations is reached.

verwendet werden. Alternatively, an absolute error can also be the target function

be used.

For the configuration of this method, the model parameters may be either known (e.g., predetermined) or estimated from (e.g., previous) measurement data.

The inventive step consists in a model predictive approach comprising an objective function for a finite forecast horizon and models for the consumers describing their temporal changes. Based on this, predictive load management can be achieved.

Another inventive step is the choice of the optimization method. This scales linearly with a number N of memories and quadratically with a number T of the time steps considered, and thus much better than a general discrete optimization method that has to evaluate at least part of the 2 ^{T · N} possible solutions.

By means of this optimization method, predefined load curves can be readjusted or adjusted well even with a few memories, and fluctuations in the load curve can be significantly reduced.

Tests with given load curves allow to investigate certain characteristics and to compare them with simulated versions of this method. An example of a characteristic behavior is the reduction of the memory contents before an increase of the target load profile.

1 shows an exemplary arrangement with a central unit 101 (Balancing or load management unit) to coordinate multiple consumers 102 to 106 , At the consumers 102 to 106 it may be, for example, a battery, a refrigerator, a heater or a pump.

Each of the consumers 102 to 106 returns a current status to an optimizer 108 the central unit 101 , The optimizer unit 108 controls the consumers 102 to 106 ,

The optimizer unit 108 uses a forecasting unit 107 to provide efficient control of consumers 102 to 106 to reach. The optimizer unit 108 and the forecasting unit 107 be from a configuration unit 109 set. The configuration unit 109 This includes, for example, models, parameters and optimization criteria, eg via a user interface 110 can be defined, modified and / or visualized. Thus, it is possible for a user via the user interface 110 Inputs made by the configuration unit 109 for the forecasting unit 107 as well as the optimizer unit 108 be implemented.

While the invention has been further illustrated and described in detail by the at least one embodiment shown, the invention is not so limited and other variations can be derived therefrom by those skilled in the art without departing from the scope of the invention.

Literature:

• [Chef 2009] Koch et al., "Active Coordination of Thermal Household Appliances for Load Management Purposes", IFAC Symposium on Power Plants and Power Systems Control, 5-8 July 2009, Tampere, Finland, 2009 ,
• [Stadler 2007] Stadler et al., "The Adaptive Fridge - Comparing Different Control Systems for Enhanced Load Shifting of Electricity Demand", 21st Conference on Informatics for Environmental Protection - Enviroinfo Warsaw 2007, Shaker Publishing, 2007, 199-206 ,

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited non-patent literature

• Stadler 2007 [0006]
• Chef 2009 [0007]

Claims (16)

Method for controlling consumers of a power grid, - where consumers are modeled, In which an objective function is determined for a predefined prognosis duration, - In which the objective function is optimized, with the optimization of the modeling of consumers are considered as constraints. The method of claim 1, wherein the consumers are modeled by modeling a temporal evolution of at least one energy storage of each consumer. The method of claim 2, wherein the time evolution of a consumer is modeled as follows:

where n is the consumer, E rel / n a normalized memory content per consumer n, E amb / n an external temperature that causes the load n to be discharged (heat loss), u _n a manipulated variable, τ _n a time constant for the discharge, k _n a coupling strength, υ _n a random process. Method according to Claim 3, in which the time evolution of the consumer is modeled on the basis of data available online, in particular by the parameters τ _n , k _n and E amb / n based on readings for E rel / n and u _n be determined. Method according to one of the preceding claims, in which the objective function is minimized as part of the optimization, which punishes a deviation of a total load from a target load. Method according to one of the preceding claims, wherein the optimization determines when a single load is switched. Method according to Claim 6, in which reference is made to a manipulated variable

with a boundary condition 0 <E rel / n, t <1 it is determined when the consumer n is switched, where P tar / t a given total load consumption at a time t, P rated / n an energy consumption of a consumer n when it loads (eg, u = 1), T denotes a time interval of a forecast. A method according to any one of the preceding claims, wherein the optimization is performed by determining, at a time at one end of a time interval for each consumer, how the target function would change if the consumer were switched at that time. The method of any one of the preceding claims, wherein the objective function comprises a quadratic error and the optimization minimizes the quadratic error. Method according to Claim 9, in which the target function

is specified. Method according to one of claims 9 or 10, in which the one consumer is switched, which achieves the greatest improvement of the objective function. Method according to one of the preceding claims, in which the optimization is carried out until it has converged or a predetermined number of iterations has been reached. Device for controlling consumers of a power network, comprising a processing unit which is set up such that - the consumers are modelable, An objective function can be determined for a predefined duration of the prognosis, - The objective function can be optimized, with the optimization of the modeling of consumers are considered as constraints. The device of claim 13, wherein the device is part of a load management. Device according to one of claims 13 or 14, comprising a configuration unit by means of which parameters for the optimization can be set. A system comprising at least one device according to any one of claims 13 to 15, wherein the device obtains status information from the consumers and performs control of the consumers based on the status information.

Images