DE102011114344A1 - Power control device and method for load balancing a network - Google Patents

Abstract

The invention relates to a method for load balancing a network with a plurality of electrical consuming (3), in which the consumers (3) each Nutzintervalle be assigned with a dependent on the respective consumer interval duration and the course of the power supplied to the consumers within the Nutzintervalle (11a to 11g; 32, 37, 48) are each controlled so that a) the consumer (3) in each useful interval (11a to 11g; 32, 37, 48) over the useful interval (11a to 11g; 32, 37, 48) averaged power corresponds to a consumer specified average power (12a to 12g; 31), and b) within each useful interval (11a to 11g; 32, 37, 48) the power (13; 35) is controlled in such a way that that at any time a total load of the consumers corresponds to a total power available to the network within specified tolerances. Furthermore, the invention relates to a power control device (1) for controlling consumers (3) according to the method.

Description

The invention relates to a method for load balancing of a network in which a load of the consumers connected to the network is adapted to a total power available to the network, and to a power control device for controlling a consumer connected to a supply network.

The increasing use of regenerative energy sources, in particular wind and solar energy, requires a growing share of primary and secondary control power as well as power plant reserves in regional and supraregional energy supply networks. These services are particularly costly but indispensable for maintaining security of supply in the event of heavily fluctuating generation. The primary, secondary and minute reserve power is primarily provided in conventional power plant blocks in the form of variable power bands and through the use of pumped storage power plants, gas turbine power plants or standby power plant blocks.

In the prior art, it is known to turn consumers on or off depending on control signals, communication channels, and / or local frequency or voltage measurements. However, this is acceptable only for a small number of power customers, since it is usually expected that the power customer can decide for himself at any time when which device is switched on or off. Only in devices such as refrigerators in which a short-term delay of turning on a customer is not detected, offers such a scheme. Furthermore, for some electrical appliances, it is only relevant that a function be performed overnight, regardless of when the function is performed at night.

As a result of such control methods, only a small proportion of the consumer power can be involved in total load balancing. Furthermore, the availability of controllable or interruptible consumer power varies both seasonally and time of day. Overall, there is a small temporal displacement potential for controllable and interruptible load shares. Such a regulation is thus applicable at most for the frequency relief after interference events.

Also known from the prior art is an indirect load balancing by variable tariffs, such that electricity is offered at reduced power consumption times. However, this results in similar disadvantages as mentioned above, wherein in addition the so-called payback effect can be observed, d. h., If after several hours of high network utilization, the network utilization decreases and a discounted price of electricity is offered, peak loads are to be observed, resulting from the simultaneous connection of power-price-dependent controlled consumers. Furthermore, the billing-related expenditure in relation to the low control power provided is clearly too high. Thus, a tariff-dependent control usually pays off for the customer, since, for example, higher costs are incurred for the complex meters, which exceed the savings for the cheaper electricity price.

The object of the invention is to enable a power control device and a method for load balancing of a network, which provides a high control power with little effort and significantly reduces a need for control. Furthermore, network stability and security of supply are to be increased. In addition, as far as possible no restrictions on the periods of time in which electrical consumers can be used should arise for the electricity customer.

These objects are achieved by a power control device according to claim 1 and a method according to claim 5. Advantageous developments and refinements emerge from the features of the subclaims.

The power control device according to the invention for controlling an electrical consumer connected to a supply network comprises a control unit and a detection unit. The detection unit is configured to detect an external control signal, and / or arranged to determine an electrical parameter of the supply network, which depends on whether a load connected to the supply network is greater or less than a reference value corresponding to a total power of the network , The control unit is designed to supply the load in such a way that, on the one hand, continuous operation of the load is possible and, on the other hand, the load is involved in load balancing of the network.

For this purpose, the control unit supplies the consumer at useful intervals of a predetermined length, in each case on average, an average power predetermined by the consumer. Within the useful intervals, the time profile of the power is controlled so that in a first phase of each useful interval the time profile of the power is controlled by a value of the electrical parameter and / or the external control signal, such that the load with a power, the deviates, is operable or is operated in the course of time from the specified average power. Power is preferably increased from the average power when the reference value is greater than the load connected to the utility grid, and decreases when the reference value is less than the load connected to the utility grid. A performance that is deviated from average performance over time is understood to mean performance that is actively deriving from average performance. A power deviating from an average power solely due to automatically resulting physical effects (ie, a frequency and / or voltage dependent load of the load) is not to be understood as power deviating from an average power in the sense of the invention. An example of an automatically resulting changed power is a reduced power consumption of a (non-controlled) transformer at a reduced line frequency.

By the control in Nutzintervallen, wherein in the Nutzintervallen the consumer in each case on average a predetermined average power is supplied, it is ensured that there are no adverse effects on the function of the consumer. It is thus exploited that sufficient for most consumers a fuzzy power setting, such that a certain energy is supplied per time interval to meet. In order to avoid an impairment of the function of the consumer, the length of the useful intervals is preferably chosen so that an inertia of the consumer is sufficiently large to compensate for the fluctuations of the actual power within each useful interval, i. H. the length of the useful intervals is selected depending on a supply quality demanded by the consumer.

Since, in a first phase of the useful interval, the consumer can thus be supplied with a current power which differs from average power and which can be controlled depending on the electrical parameter or the external control signal and thus on the requirements of the supply network with regard to load balancing contribute to the load balancing of the supply network.

As a result, the proportion of primary and secondary control power or, in some cases, minutes or hours of reserve power required for a certain level of security of supply is significantly reduced, without jeopardizing network stability.

It should be noted that the reference value of the total power of the network, i. H. the total power available to the network. The network can be a nationwide power supply network or a smaller distribution network. Likewise, the network may be an in-house network or the like, in which an avoidance of high peak power is sought. This objective (reduction of peak power) is the primary control objective for in-service networks (ie also more important than load balancing in the higher-level grid connected to the in-house network), since electricity utility electricity prices often come from electricity costs and costs for a peak power to be provided by the utility when needed. A reduction of peak power thus leads to lower electricity costs. In this case, the reference value may correspond to a desired value of the consumption, i. H. this setpoint of consumption is understood as the power available to the network within the meaning of claim 1.

In a preferred embodiment, the control unit has a logic unit and a power limiter, wherein the power limiter is designed to supply the consumer with a power dependent on a signal output by the logic unit.

Further preferably, the control unit may include a power controller, which may be part of the logic unit in particular. In this preferred embodiment, the power limiter is designed to switch an output of the control unit such that in the first phase of the useful interval an actual power supplied to the load depends on a difference of the reference value and the load connected to the supply network, while the load is in a second Phase maximum or minimum power is supplied. The ability to feed the maximum or minimum power in a second phase compensates for the difference between the current power and the average power provided by the load, which is integrated in the first phase. One Such power regulator can of course also be provided in embodiments in which the detection unit is designed to detect an external control signal. In these embodiments, the actual power supplied to the load in the first phase depends on the external control signal.

beschrieben werden. Hierbei ist W_Intervall die während des Intervalls insgesamt verrichtete Arbeit, P(t, x) die Leistung abhängig von der Zeit sowie einer Regelgröße x (beispielsweise einer Abweichung der Netzfrequenz von einer Sollfrequenz), t₁ die Zeit zum Ende der Regelphase (als Nullpunkt der Zeit wurde der Intervallbegin gewählt), t₂ die Zeit zum Ende der Ausgleichsphase (und damit die Zeit zum Ende des Nutzintervalls), P_max/min die in der Ausgleichsphase zugeführte, minimale oder maximale Leistung und P_⌀ die vom Verbraucher geforderte durchschnittliche Leistung im Intervall.In another embodiment, the control unit has both a power limiter and a power regulator as described above, wherein the power regulator is arranged to control a three-point switch, wherein the three-point switch is connected between the power limiter and the consumer. The three-point switch can be used to connect the load to the power limiter in the first phase (control phase) and to a connection for the maximum or minimum power in the second phase (compensation phase). The work done in the entire interval can therefore be determined by the equation

to be discribed. In this case, W _{interval is} the work performed overall during the interval, P (t, x) the power depending on the time and a controlled variable x (for example, a deviation of the mains frequency from a nominal frequency), t ₁ the time at the end of the control phase (as zero point the time interval start was chosen), t ₂ the time to end of the balancing phase (and thus the time to the end of the useful interval), P _{max / min} the minimum or maximum power supplied in the balancing phase and P _⌀ the average power demanded by the consumer in the interval.

It should be noted that a three-position controller (i.e., a three-point switch driven by a power controller) is only one (particularly advantageous) example of a controller useful in the context of the invention. Instead, for example, a dynamic controller or another controller can be used to switch from the control phase to the compensation phase. In addition, the controller may be configured to control the maximum and / or minimum power, i. H. the controller can change the maximum and / or minimum power time-dependent. Maximum and / or minimum power may be increased and / or decreased within one interval. Likewise, the maximum and / or minimum power may be constant within each interval, with maximum and / or minimum power increased and / or reduced in at least one interval from the previous interval.

In a method according to the invention for load balancing of a network with a large number of electronic consumers, the consumers are each assigned useful intervals with an interval duration which is dependent on the respective consumer. Within the useful intervals, the power is controlled in each case so that the consumer averaged over the useful interval one of the consumer specified average power corresponding power is supplied and on the other hand within each useful interval, the total load of the consumers of a total available power within predetermined tolerances.

By this it is meant that the total load of the total power available in the network corresponds to the extent that the locally valid standard regarding the mains frequency and the nominal voltage of the network is fulfilled. For example, the total load of the available power in the network can correspond so far that the mains frequency deviates from a nominal frequency by no more than 200 millihertz, and the voltage in the network deviates by less than 10% from a nominal voltage. So, for example, those in the European Standard EN 50160 be fulfilled.

In one embodiment of the method, the useful intervals of the consumers are each subdivided into a first phase serving as a control phase and a second phase following the first phase and serving as a compensation phase. In the first phase of the utility intervals, the current power of each consumer is adjusted to compensate for a difference between the current total available power in the utility grid and the total average power provided by the consumers.

In the second phase, the current power of the consumer is regulated so that the power averaged over the useful interval and supplied to the respective consumer corresponds to the average power given by the consumer. That is, if the consumer was supplied too high a power (ie, higher than the required average power) in the first phase, then in the second phase (balancing phase) the consumer is given a reduced power (preferably the minimum power) while the load is being consumed if it has been fed too low an output in the first phase, an increased power (preferably the maximum power) is supplied in the second phase. The regulation of the current performance thus takes place in the second phase, preferably completely independent of the total average power of the consumer and the power available in the network. Instead, in the second phase it is achieved that the consumer is supplied with the required average power averaged over the entire useful interval. The division of the useful intervals thus achieves that in the first phase the consumer can participate in a load balancing of the network, while ensuring that the consumer continues to receive the average power necessary for the consumers to be fully functional.

In an embodiment of the method which is also particularly suitable for sensitive electronic consumers, not only an average power but also a maximum and / or a minimum actual power is predetermined for the consumers. In this embodiment, the consumer is controlled such that the current power is always greater than or equal to the minimum and less than or equal to the maximum current power. This prevents the short-term, too high performance from damaging the electronic consumer. Also, lowering the current power too much could lead to a consumer outage, and this is prevented by the set minimum power.

In an optimal embodiment with respect to the total available control power, in the balancing phase, the actual power may correspond to the maximum or minimum power. As a result, a particularly long control phase is possible, being changed from the control phase in the compensation phase as averaged over the respective useful interval one of the consumer specified average power corresponding performance can be realized only by constantly supplying the maximum or minimum current power for the remaining time interval can. It is therefore preferably switched at the latest possible time of the control phase in the compensation phase.

Alternatively, however, it is also possible that when currently only a small control power is needed, consumers with a relatively short useful interval are prematurely switched from the control phase in the compensation phase.

In a further embodiment, the consumers are each assigned a delay time, which indicates how quickly the current power of the respective consumer is adapted to an available power for load balancing in the network. In such embodiments, a difference between the currently available power and the total average power of the load is first offset by a change in the current power of the load with the least delay time.

As soon as the consumers with the least delay time switch from the control phase to the compensation phase, the load compensation is taken over by electrical consumers with a higher delay time. These then on the one hand equalize the difference between the total average power provided by the consumers and the total power currently available in the network and, on the other hand, an additional load resulting from the switching of the consumers with a shorter delay time from the control phase to the balancing phase. The function of fast-reacting consumers involved in load balancing is thus gradually taken over by slower consumers.

The load balance thus works preferably according to a snowball principle. Consumers who can take part in the load balancing only briefly (small length of the useful intervals) but with high dynamics (short delay time), the power differences to be compensated "continue" to slowly reacting load groups which can be activated over longer useful intervals. Thus, with a persistent power difference (ie difference between a power available in the grid and the total average power required by the consumers) in the overall system, the temporal displacement potential at the controllable load groups (consumers) is gradually increased and, in the meantime, the controllability of the faster-reacting load groups or consumer restored. This principle is referred to as fractal load control according to the invention. It should be noted that the above-described "handover" without a data exchange between power control devices of different consumers is feasible. The "passing on" is already possible because power control devices of slowly responding consumers within the delay time observe the electrical parameters of the supply network, the course of this electrical parameter implies how the faster responding consumers have responded to the course of the electrical parameter. This results in the extent to which the slow-reacting consumers have to take part in the load balancing at the end of their delay time.

It has been found that a particularly good control can be achieved if the lengths of the useful intervals and the delay times are chosen so that consumers with longer useful intervals also have longer delay times. Particularly preferably, the delay time is chosen to be proportional to the length of the useful interval.

Depending on the embodiment, an electrical parameter of the network, preferably the mains frequency or an overvoltage or undervoltage, can be detected in the method, it being determined by means of the parameter whether the total load of the consumers of a total power available to the network within predetermined tolerances (see above) corresponds. The control of the currently supplied power is then dependent on a deviation of the parameter from a setpoint of the parameter. If, for example, a total power available in the network is greater than a total load of the consumers, a grid frequency increases. In order to regulate the network frequency as accurately as possible to the desired value, as described above, the consumer is thus counteracted by temporarily increasing the current power with low delay times.

Alternatively or in addition to a regulation via a measured electrical parameter, a control can be provided via an external control signal.

This is expedient, for example, if electricity is purchased under the condition that the electricity price is composed of an overall power-dependent component and a maximum-power-dependent component. In this case, power consumption for a power-consuming operation decreases when it consumes uniform power.

In such embodiments, therefore, it is not primarily a load balancing in the supplying power network that is aimed at in the interest of the operation but a uniform power consumption of the operation. This can not be determined on the basis of a network parameter such as the network frequency, since this depends on various other non-operating consumers. In order to achieve the most uniform possible power consumption of the operation and thus low electricity costs in such situations, an external control signal - which can indicate, for example, a total current consumed by the operation - is determined by the detection unit.

In a particularly preferred embodiment, the electrical parameters of the network are determined separately for the individual consumers and the current power of the consumers is controlled separately. Accordingly, a separate power control device can be provided for each consumer.

By such a configured method, a difference between generated and demanded electrical power can be compensated decentralized, by load groups or consumers reacting at different rates. For this purpose, each load involved in the load balancing is preferably equipped with an individual power control device. Thus, a self-sufficient, decentralized load balancing is possible, with no additional data lines between the power control devices are necessary.

It should be noted that with a single consumer in the context of this application, not necessarily a single electrical device must be meant; rather, device groups up to entire households or blocks of houses can be jointly controlled as a single consumer.

In a further advantageous embodiment of the method, at least one power generator or storage supplying energy to the network is also included in the control, so that in the case of a long-term difference between the currently available power and the total average power of the consumers, an energy delivery of the Adjusted to the total average consumer power.

The fractal load balancing can thus be integrated into the regulation of the energy supply system such that long-lasting differences in the power balance (ie deviations of the power available in the grid from the power required by the consumers), which exhaust the capacity of the fractal load balancer, by anticipatory activation of reserve power in the power plants, while providing most of the primary and secondary control power, and in part the minute reserve power provided by the fractal load balancing.

It is particularly preferred for consumers to cycle through the useful intervals in such a way that control and compensation phases alternate for each of the consumers.

It should be pointed out that the current interval duration, the current average power specified by the consumer, the minimum current power and the maximum current power and / or the current delay time do not necessarily have to be constant, but are also varied over time or depending on operating parameters of the consumer can. By "current" is meant the value of the respective size in a current interval. The "current interval duration" is thus the duration of the current interval. In particular, successive intervals may have a different interval duration.

If, for example, a fluorescent tube is controlled, an interval duration for the connection process may be relatively short, whereby a relatively precise regulation of the actual output is necessary. Accordingly, there is little difference between the average power given by the consumer and the maximum and minimum power. So it is a high quality of care required. Once the fluorescent lamp has been switched on, such high quality of care is no longer necessary. For the continuous lighting operation, the fluorescent tube is then driven with a larger useful interval duration and also a distance between the minimum and maximum power and the average power is larger (relative to the average power). Due to the fact that the interval duration, the average power specified by the consumer, the minimum current power, the maximum current power and / or the delay time are not constant over time, but are time-dependent and / or controlled as a function of the operating parameters of the consumer, one available one Control power can be maximized because the values mentioned can always be adapted to the current operating conditions of the consumer.

Embodiments of the invention will be explained in more detail with reference to FIGS. Show it:

1 A first embodiment of a power control device on the consumer,

2 an exemplary performance curve for a power control device according to claim 1,

3 a sketch to illustrate the interaction of a variety of power control devices,

4 a progression of the current power for several consumers with different delay time,

5 a more detailed variant of a power control device,

6a -E diagrams to illustrate the regulation in a variety of consumers and

7 Representation to illustrate a multi-level control.

In 1 is an arrangement 1 consisting of a possible embodiment of a power control device 2 as well as a consumer 3 shown. The illustrated power control device 2 for a single consumer 3 (this may be a single piece of electrical equipment, but also an entire block or factory) includes a power-limiting electronic component 4 , which sets the power consumption for the consumer within a given band, a sensor 5 (This may, for example, be a mains frequency sensor) and alternatively or in addition to the sensor 5 an external communication channel 6 for receiving a control signal and a logic or control unit 7 , Furthermore, the arrangement has a network connection 8th on that with the electronic component 4 connected is.

Depending on the embodiment, the logic or control unit 7 adapted to receive a signal from the sensor 5 and / or an external control signal via the external communication channel 6 to recieve. Whether a signal from a sensor or an external control signal is used depends mainly on the application. Is by the power control device 2 the aim of avoiding a peak load in an operation, that is depends the control target (minimization of peak power) not from the overall network but a local size such as power consumption during operation, preferably an external communication signal (this can indicate, for example, the current power consumption of the operation) is used , If, however, to optimize the stability of the entire network regulated, preferably the signal from the sensor 5 used.

Depending on the signal or signals, the logic or control unit controls 7 the power limiter 4 so that the consumer 3 one through the logic or control unit 7 specified power is supplied.

This is based on 2 explained in more detail. In 2 is on the x-axis 9 the time and on the y-axis 10 the power applied. The x-axis 9 is in several in this case the same length of use intervals 11a to 11g divided. In addition, an upper and a lower dashed line P _max or P _min are shown over all the illustrated intervals, the upper dashed line indicating the maximum load deliverable to the load and the lower dashed line indicating the minimum power that can be supplied to the load. It should be noted that neither the maximum nor the minimum deliverable power must be constant in time; In order to ensure clarity of the figure, however, the maximum and the minimum power were shown as constant in this case.

Between the upper dashed line P _max and the lower dashed line P _min is separately for each interval a further broken line, 12a to 12g representing an average power demanded by the consumer in the respective interval. Furthermore, a solid line can be seen 13 which indicates the current performance at the respective time (ie a time course of the performance). In every interval 11a to 11g corresponds to an integral of the time course 13 the performance over the respective interval, a product of the interval duration and the average power required by the consumer 12a to 12g , That is, averaged over the interval, the consumer is supplied with the power required by him.

Each of the intervals 11a to 11g is in a first phase 14 and a second phase 15 divided. At each time ta to tg is switched from the first phase to the second phase. Within the first phases 14 (also called rule phases) is the current performance 13 regardless of the for the respective interval 11a to 11g required average power 12a to 12g controlled. Instead, to control the current power (except for the constraint that the current power must always be between the maximum and minimum power), it only takes into account which actual power is optimal in terms of load balancing.

For this purpose, z. B. the signal of the sensor 5 be taken into account, in which case if the sensor 5 a frequency above a target frequency measures, the consumer one over the average power 12a lying current performance 13 is fed and then when through the sensor 5 too low a frequency is detected, the consumer 3 one below an average power 12a lying current performance 13 is supplied. Of course, a corresponding regulation is also possible if an electrical parameter other than the frequency is used or if an external control signal is used for regulation.

At the time 16a the average power supplied up to that point in time (ie the integral of the power) differs 13 from the start of the interval to the time 16a ) so much of an average power demanded by the consumer 12a that the required average power 12a can only be achieved by the fact that the maximum power P _{max is} supplied to the consumer for the remainder of the interval. At intervals in which the consumer within the first phase 14 too high a power has been supplied (for example, the interval 11c ), a corresponding averaged over the interval performance is only possible by constant supply of the minimum power P _min at the appropriate time.

At the respective time 5a to 5g will then be in the second phase 15 the respective payload interval 11a to 11g switched within this second phase, the consumer either the minimum or the maximum power is supplied, this is done independently of a desirable for load balancing the network performance. Within the second phase 15 (also referred to as compensation phases) is the current performance of the consumer 3 ie independent of a signal from the sensor 5 and / or one over the external communication channel 6 supplied signal regulated.

It should be noted that in 2 a simplified history of the current performance is shown; in other embodiments, the consumer is assigned a delay time Θv ⁱ , wherein in each interval the control phase 14 only after expiry of the delay time begins. For control are thus by the logic or control unit 7 the delay time, the duration of the interval interval, the length of the already past time portion within the averaging interval, a mean interval power during this time portion, an average power to be achieved in the interval, a maximum power and a minimum power are taken into account as control parameters. Depending on the embodiment, the delay time may be the beginning of each Form intervals, for example, the consumers get their average power supplied in the delay time. However, it can also be provided that the delay time for the intervals is formed in each case by a last section of the compensation phase of the respective preceding interval. In this case, the consumers can each contribute to the load balance from the beginning of each interval, wherein only in a first interval (ie, when connecting the consumer to the supply network) lying within this first interval, the control phase preceded delay time is necessary.

The interaction of a large number of controlled consumers is based on the 3 and 4 explained in more detail.

In 3 is a lot of arrangements after 1 . 1.1 to 1.x , shown, which are each connected in parallel to each other. The arrangements 1.1 to 1.x In each case a power is supplied with a current network frequency. There is a connection for this 17 intended. Furthermore, a nominal value of the mains frequency is supplied to the arrangements. This is indicated by the arrow 18 indicated. About one each by the arrangements 1.1 to 1.x The power taken from the grid affects the arrangements 1.1 to 1.x load balancing in the network. This is indicated by the arrows 19.1 to 19.x shown.

The arrow 20 indicates that power is supplied to the grid by power generators. By the supplied power as well as by the arrangements 1.1 to 1.x Consumed power is an electrical parameter of the network, such as the grid frequency, influenced. This is done by the box 21 indicated. The correspondingly influenced power grid in turn supplies the arrangements 1.1 to 1.x as by the arrow 17 is indicated. The arrangements 1.1 to 1.x are each as under the 1 and 2 explained operated controlled.

The individual arrangements 1.1 to 1.x each include a consumer with a different delay time. As a result, a time-offset load compensation by the arrangements 1.1 to 1.x possible. This will be explained in more detail by 4 explained.

In 4 Several graphs are shown above each other, where the graphs are each an x-axis 16 on which the time is plotted and where the x-axes 16 the individual graphs are scaled identically. The graphs 22 . 24 and 26 represent here the time course of a control power demand plotted on the y-axis, while the graphs 23 . 25 and 27 show the time course of the power (y-axis) of a first, second and third consumer.

The graph 22 represents the standard requirement that arises from the disconnection or connection of other consumers and / or from fluctuations in electricity production. As can be seen, consists in a first section 28 an oversupply of network-available power during a second period of time 29 too little power is available in the network. The first consumer whose history of the current performance in the graph 23 has a useful interval length which corresponds to about half of the illustrated time range. The end of the first payload interval is in the graph 23 through a vertical line 30 clarified.

For the sake of simplicity, the average power required by the consumer has been shown to be constant in this embodiment. This required average power is given by the horizontal line 31 indicated. Furthermore, dashed lines show a maximum and a minimum power P _max and P _min . In the same way, the required average power and the maximum and minimum power are in the graph 25 and 27 indicated.

The first consumer whose history of the current performance in the graph 23 is shown, gets to the beginning of the first interval 32 a current power, the average power demanded by the consumer 31 equivalent. With expiration of the delay time 33 (The end of the delay time is indicated by the vertical dashed line 34 clarifies), the first consumer begins to participate in load balancing. The delay time is used here to decide whether a control intervention is required or not.

Since too much power is initially available in the network, the consumer initially has too high a current performance 35 fed. To a through the vertical dashed line 47 indicated time can be one of the required average power 31 corresponding over the interval 32 average power can only be realized by the fact that the first consumer for the remainder of the interval 32 the minimum voltage is supplied. Accordingly, the consumer is in the compensation phase 36 connected. With the beginning of the second interval 37 the consumer can (after another delay time) participate again in the regulation.

The regulation requirement taking into account the control of the first consumer is in the graph 24 shown. As can be seen, there is still a certain need for regulation within the delay time of the first consumer. However, since the interval time of consumers who require a very high quality of care can be in the range of a few milliseconds, this period is negligibly short.

From the moment when the first consumer participates in the regulation, there is initially no need for further regulation until the first consumer is switched to the equalization phase. At this point in time, the demand for rules suddenly increases because, on the one hand, there is more power available in the network than is needed and, on the other hand, the first consumer is there for the remainder of its first interval 32 less than the average power 31 must consume.

The second consumer whose history of the current power in the graph 25 is shown has a higher interval duration. The first interval 48 of the second consumer ends at a time 38 , After a delay time 39 For example, the second consumer may participate in a scheme and thus offers the balancing power needed, taking account of the first consumer. At a time 49 The second consumer switches to its compensation phase, so that there is again need for regulation.

The regulation requirement taking into account the second consumer is in the graph 26 shown. The corresponding required control power is provided by a third consumer whose interval duration extends beyond the illustrated time range.

Like the graphene 22 to 27 can be seen, a rule requirement is first met by fast consumers, and then, if the fast consumers require compensation for the initial regulation, the regulation and the corresponding compensation is taken over by slower consumers. The load balancing thus takes place according to a snowball principle, whereby gradually more and more consumers are included in the load balancing. The control method can thus be called a fractal load balancing.

It should be noted that frequency deviations do not necessarily have to be completely compensated, ie it is not necessary to provide the complete control of the difference between the power provided to the network and the total load of the load required by the load control according to the invention. Instead, it is also possible to accept a certain frequency deviation, whereby, depending on the frequency deviation, it is controlled how strongly the individual consumers participate in the regulation. In this case, the regulation of the loads increases the value for the derivation of the grid power according to the grid frequency, ie while without control interventions the derivative dP _grid / dω _{grid is} approximately equal to 0.024 P _soll / Hz, where P _{grid is} the power currently supplied to the grid, ω _{Net is} the current grid frequency and P _is the power needed in the grid (that is, a power change of about 2.4% leads to a frequency change of 1 Hz), this value could be increased so much that even significant power fluctuations still an acceptable frequency deviation Consequence would have.

For a more detailed description of the power control devices 2 is in 5 a more detailed variant of a power control device shown. On the left in 5 can be seen that the power control device, as indicated by the arrow 39 indicated, a current grid frequency and, as indicated by the arrow 40 indicated, a target frequency is supplied. The mains frequency and the nominal frequency become an observer 41 fed.

The Observer 41 (This may be, for example, a control circuit) decides whether the controlled by this power control device consumer after a delay time involved in the scheme or not. For this purpose, the observer carries out a signal-theoretical evaluation. In this case, customary signal evaluation methods such as moving averaging, evaluation of characteristic signal components, signal transformations or other methods known from the prior art can be used. In the signal-theoretical evaluation, the observer takes into account 41 not only a deviation of the current network frequency of the target frequency, but also the limited control ability of the other consumers involved or resulting from switching the consumer with a shorter delay time in compensation phases additional deviation. The behavior of consumers with shorter delay time can be the observer 41 estimate the course of the network frequency within the delay time.

A result of the evaluation is sent from the observer to another unit 42 passed a subsequent load balancing scheme 43 depending on the results of the evaluation by the observer 41 active or inactive switches. Will the load balancing circuit 43 When switched active, it controls an output power exclusively depending on load balancing requirements. Will the load balancing circuit 43 On the other hand, if not activated, the consumer can be supplied with the average power demanded by the consumer.

The load balancing circuit 43 Below is a three-point switch 44 arranged. In the three-point switch is a connection for the maximum power, for the minimum power as well as for the output of the load balancing circuit 43 intended. On the output side, the three-point switch has a connection for the consumer 57 on. The three-point switch 44 is also equipped with a power regulator 45 connected.

The power controller 45 monitors to what extent the consumer so far 57 In the current control interval, power supplied deviates from a required average power. At a time when an average power averaged over the interval can only be realized according to the requirements of the consumer if the maximum or minimum power is continuously supplied to the consumer (cf. 2 . 16a to 16g ), the power controller switches the three-point switch 44 such that one output of the three-way switch 44 connected to the port for maximum or minimum power. The output of the three-point switch 44 is fed to the consumer.

The behavior of a system with a large number (here 2000 pieces) of consumers is based on the 6a to 6e explained in more detail.

In 6a the simulated time course of the sum of the compensation reactions of the consumers is shown. On the y-axis 50 is a deviation of the current power P (t) (in arbitrary units) from a power demanded by the consumers on average and on the x-axis 46 Time in seconds. The course of the deviation of the network frequency (not shown) has a substantially identical form as in FIG 6a illustrated course. At the beginning, there is almost no mains frequency deviation, so that only a negligibly small compensation reaction (control power) of the consumers can be detected. Between a time of 30 seconds and about 80 seconds there is an overfrequency in the network, while between a time of about 80 seconds and about 130 seconds is an underfrequency.

In 6b becomes a regulatory contribution depending on a delay time of the consumer at the time 15 Seconds after simulation start (see 6a ). On the x-axis 51 Here, the number of the consumer is plotted, with a low number of the consumer corresponds to a short delay time. On the y-axis 52 is the control power in any units (scaled differently than in 6a ). As can be seen, almost no regulation is carried out by the consumers, since the grid frequency essentially corresponds to the nominal frequency.

In 6c is a corresponding diagram for the time 58 Seconds after the simulation starts. As can be seen, a control power is provided mainly by high delay time components. This is because an overfrequency has already been present for some time, so that the control power has already shifted from the faster consumers to the consumers with a long delay time.

In the 6d and 6e is a corresponding diagram for the times 84 Seconds and 120 seconds. At the time 84 Seconds (see 6a ), has just before the zero crossing of the control power, ie the change from overfrequency to underfrequency, occurred. Since only recently has an underfrequency, are still used to a large extent consumers with a short delay time for load balancing.

As in 6e It can be seen, this has until the time 120 Seconds so far postponed that much of the control power is provided by the consumer with high delay time.

In 7 is shown that a control as described above can also be carried out in multiple stages. A high voltage network 53 (this is the highest network level, in the figure above) are multiple high voltage networks 54 assigned. These in turn supply medium-voltage networks 55 , These are with substations 56 for low-voltage networks, for example for individual blocks, with low-voltage networks being the lowest level. Any of the high voltage networks can 54 here as a consumer of the extra high voltage network 53 be considered. The medium voltage networks 55 in turn, consumers of high-voltage networks 54 and the substations 56 Consumers of medium voltage networks 56 , It should be noted that although only 3 high-voltage grids are shown here, however, that is usually much more high-voltage grids 54 to the extra high voltage grid 53 can be connected. This applies equally to medium-voltage grids 55 and the substations 56 ,

Now there is a rule requirement due to a decreasing power frequency, can be lowered by inventively controlled substations briefly the voltage of the lower network levels, with after some time to compensate an increased voltage is provided by the substations available.

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

• Standard EN 50160 [0019]

Claims (15)

Power control device ( 2 ) for controlling an electrical consumer connected to a supply network ( 3 ), comprising - a control unit ( 4 . 7 ; 41 - 45 ) and - a registration unit ( 5 . 6 ), the registration unit ( 5 . 6 ) is adapted to detect an external control signal and / or to determine an electrical parameter of the supply network, which depends on whether a load connected to the supply network is greater or less than a reference value corresponding to the total power of the network, and wherein the control unit is adapted , the consumer at useful intervals ( 11a - 11g ; 32 . 37 . 48 ) of a predetermined length each on average one by the consumer ( 3 ) given average power ( 12a - 12g ; 31 ), whereby a temporal course of the performance ( 13 ; 35 ) in a first phase ( 14 ) of each useful interval ( 11a - 11g ; 32 . 37 . 48 ) is controlled depending on a value of the electrical parameter and / or the external control signal such that the consumer ( 3 ) with a time course of the given average power ( 12a to 12g ; 31 ) different performance ( 13 ; 35 ) is operable. Power control device ( 2 ) according to claim 1, characterized in that the control unit ( 4 . 7 ; 41 - 45 ) a logic unit ( 7 . 41 . 42 . 45 ) and a power limiter ( 4 . 43 ), the power limiter ( 4 . 43 ) is designed to assist the consumer ( 3 ) one by one through the logic unit ( 7 . 41 . 42 . 45 ) output signal dependent power ( 13 ; 35 ). Power control device ( 2 ) according to claim 1 or 2, characterized in that the control unit ( 4 . 7 ; 41 - 45 ) comprises a power controller ( 45 ), which is adapted to an output of the control unit ( 4 . 7 ; 41 - 45 ) so that in the first phase ( 14 ) of the useful interval ( 11a to 11g ; 32 . 37 . 48 ) a course of the consumer ( 3 ) supplied power ( 13 ; 35 ) depends on a difference between the reference value and the load connected to the supply network, while the consumer ( 3 ) in a second phase ( 15 ) a maximum or minimum power (Pmax or Pmin) can be fed. Power control device ( 2 ) with the features of claims 2 and 3, characterized in that the power regulator ( 45 ) for controlling a three-point switch ( 44 ), wherein the three-point switch ( 44 ) between the power limiters ( 4 . 43 ) and the consumer ( 3 ) is switched. Method for load balancing a network with a plurality of electrical consumers ( 3 ), in which consumers ( 3 ) are each assigned to user intervals with an interval duration predetermined by the respective consumer, and the course of the power supplied to the consumers within the useful intervals ( 11a - 11g ; 32 . 37 . 48 ) is controlled in such a way that a) the respective consumer ( 3 ) in each use interval ( 11a - 11g ; 32 . 37 . 48 ) over the useful interval ( 11a - 11g ; 32 . 37 . 48 ) averaged power delivered by a consumer given average power ( 12a - 12g ; 31 ) and b) within each useful interval ( 11a - 11g ; 32 . 37 . 48 ) a time course of the performance ( 13 ; 35 ) is at least temporarily regulated to a value deviating from an average power specified by the consumer. Method according to Claim 5, characterized in that useful intervals ( 11a - 11g ; 32 . 37 . 48 ) the consumer ( 3 ) in each case in a first phase serving as a control phase ( 14 ) and a second phase following the first phase, which will serve as a balancing phase ( 15 ; 36 ), whereby in the first phase ( 14 ) of the useful intervals ( 11a 11g ; 32 . 37 . 48 ) the time course of the service ( 13 . 35 ) of the respective consumer ( 3 ) is adjusted to compensate for the difference between the current total supply available in the supply network and the total average consumer power provided by consumers, and in the second phase ( 15 ; 36 ) the performance ( 13 ; 35 ) of the respective consumer ( 3 ) is regulated so that the over the useful interval ( 11a - 11g ; 32 . 37 . 48 ), the respective consumer ( 3 ) supplied power ( 13 ; 36 ) the average power provided by the consumer ( 12a - 12g ; 31 ) corresponds. Method according to one of claims 5 or 6, characterized in that for the consumer ( 3 ) in addition to the average power ( 12a - 12g ; 31 ) a maximum and / or a minimum power (Pmax or Pmin) is predetermined, whereby the consumers ( 3 ) are controlled in such a way that the power ( 13 ; 35 ) is always greater than or equal to the minimum and equal to or less than the maximum power (Pmax or Pmin) over its time course. Method with the features of claims 6 and 7, characterized in that in the balancing phase ( 15 ; 36 ) the performance ( 13 ; 35 ) corresponds to the maximum or minimum power (Pmax or Pmin), wherein within the useful intervals ( 11a to 11g ; 32 . 37 . 48 ) of the control phase ( 14 ) into the balancing phase ( 15 ; 31 ), as soon as, over the respective useful interval ( 11a - 11g ; 32 . 37 . 48 ), one of the consumers ( 3 ) given average power ( 12a - 12g ; 31 ) corresponding performance only by constantly supplying the maximum or minimum power (Pmax or Pmin) for the remaining useful interval ( 11a 11g ; 32 . 37 . 48 ) can be realized. Method according to one of claims 5 to 8, characterized in that the consumers ( 3 ) each one delay time ( 33 ), where a difference between the current available power and the total average consumer power ( 3 ) first about a change in the time course of the performance ( 13 ; 35 ) the consumer ( 3 ) with the least delay time ( 33 ) is compensated. Method according to one of Claims 5 to 9, characterized in that an electrical parameter of the network, preferably the network frequency, is detected, it being determined by means of the parameter whether the total load of the consumers ( 3 ) corresponds to a total power available to the network within predetermined tolerances, wherein the control of the time course of the power supplied to the consumer ( 13 ; 35 ) is dependent on a deviation of the parameter from a setpoint of the parameter. Method according to claim 10, characterized in that separately for the individual consumers ( 3 ) determines the electrical parameter of the network and the time course of the power ( 13 ; 35 ) is controlled. Method according to one of Claims 5 to 11, characterized in that at least one power generator supplying power to the network is regulated in such a way that, given a long-term difference between the available power and the total average power of the consumers ( 3 ) an energy output of the energy producer to the total average consumer power ( 3 ) is adjusted. Method according to one of claims 5 to 12, characterized in that the useful intervals ( 11a - 11g ; 32 . 37 . 48 ) are cycled. Method according to one of claims 5 to 13, characterized in that the interval duration, the consumer specified average power ( 13 ; 35 ), the minimum power (Pmin), the maximum power (Pmax) and / or the delay time ( 31 ) for at least one of the consumers ( 3 ) is varied over time depending on operating parameters of the consumer and / or time-dependent and / or depending on other reference variables. Method according to one of Claims 5 to 14, characterized in that network connections of a lower network level are controlled as consumers of a higher network level.

Images