Method and control unit for controlling a power network

Abstract

The invention proposes a method for controlling a power network (2) having a plurality of network nodes (3) by means of a control unit (1), the method being characterized at least by the following steps: - (S1a, S1b, S1ab) providing an a-priori distribution (6) for state variables of the power network (2); - (S2) providing a measurement distribution (7) of measured variables in a given state of the power network (2); - (S3) forming an a-posteriori distribution (42) for the state variables by means of the a-priori distribution (6) and the measurement distribution (7); - (S4) determining a probability (4) for a critical state for at least one of the network nodes (3) by means of the marginal distribution (43) determined for the network nodes from the a-posteriori distribution (42); and - (S5) using the control unit (1) to perform measures that are beneficial to the network and / or system when the determined probability (4) is greater than or equal to a set threshold value. The invention also relates to a control unit (1) for carrying out the method.

Description

[0001] Description

[0002] Method and control unit for controlling a power grid

[0003] The invention relates to a method according to the preamble of patent claim 1 and a control unit according to the preamble of patent claim 15.

[0004] Due to the progressive installation of additional consumers, for example with regard to electromobility, and volatile decentralized generation units, for example using photovoltaic systems or wind turbines, critical operating conditions can increasingly occur, particularly in medium and low voltage networks.

[0005] In addition to complex and costly grid expansion, operating approaches are planned that specifically control consumers in critical grid situations to avoid overloads. Such critical grid situations or conditions can be characterized by a violation of the permissible voltage range, the triggering of a fuse, or by exceeding maximum power. However, these operating approaches require knowledge of the state of the power grid, for example, how close an avoidable voltage range violation or a current limit is at one or more grid nodes.

[0006] Currently, little to no measurement infrastructure is installed for medium- and low-voltage grids. This allows only imprecise estimates of the condition to be determined, while for a grid operator, the only relevant factor is whether a limit violation is imminent or likely.

[0007] According to the state of the art, a network state estimation is performed to solve this problem. This involves attempting to account for the accuracy of the available measured values ​​using weights in a quadratic objective function to be minimized. The resulting estimated network state is deterministic. This can lead to a high degree of uncertainty regarding the true network state.

[0008] Furthermore, probabilistic methods for state estimation are known. However, these only consider the expected values ​​or maximum values ​​as estimates, so that a deterministic result is obtained.

[0009] The present invention is based on the object of providing an improved control of a power grid based on a probabilistic approach.

[0010] The object is achieved by a method having the features of independent patent claim 1 and by a device having the features of independent patent claim 15. Advantageous embodiments and further developments of the invention are specified in the dependent patent claims.

[0011] The method according to the invention for controlling a power grid with several network nodes is carried out by means of a control unit and is characterized by the following steps:

[0012] - Providing an a priori distribution for state variables of the power grid;

[0013] - Providing a measurement distribution of measured variables for a given state of the power grid;

[0014] - Forming a posterior distribution for the state variables using the prior distribution and the measurement distribution;

[0015] - Determining a probability for a critical state for at least one of the network nodes using the marginal distribution determined from the posterior distribution for the network node; and

[0016] - Implementation of grid-friendly and / or system-friendly measures by the control unit if the determined probability is greater than or equal to a specified threshold. The order of the steps of the method does not imply a chronological order of the steps. In particular, these can - as far as possible - be carried out in parallel. In other words, one or more steps of the method are preferably carried out in parallel.

[0017] The method according to the invention and / or one or more functions, features, and / or steps of the method according to the invention and / or one of its embodiments can be computer-aided. In particular, the control unit comprises one or more computing units designed and configured to determine the probability, for example, by means of commands.

[0018] The power grid is an electrical distribution network, in particular a low-voltage network and / or a medium-voltage network. The power grid has several network nodes. Furthermore, the power grid typically has several lines that extend from one of the network nodes to another of the network nodes. The topology of the power grid can comprise strands and / or rings.

[0019] The power grid also has one or more state variables, with one or more state variables typically being assigned to each network node. In particular, each network node has one or more complex-valued voltages, i.e., voltage magnitudes and voltage angles (angles), as state variables. The power grid can be configured as a multi-phase system.

[0020] State variables of the power grid can be complex-valued voltages, complex-valued currents, phase asymmetries, higher harmonics, for example, by modeling the systems / devices connected to the grid, and / or transformer rated power. Complex-valued voltages are particularly preferred as state variables. In principle, however, any technical power grid variable can be used that characterizes non-critical and critical operation of the power grid through an associated limit value.

[0021] System-supporting measures include, in particular, control interventions or adjustments related to frequency maintenance, voltage maintenance, safe operation, and / or supply restoration of the power grid. These are typically carried out by a grid operator of the power grid, so the control unit can be comprised of a grid control device of the grid operator.

[0022] Grid-friendly measures can include the control of individual or multiple electrical systems, particularly generators, consumers, and / or storage facilities, which, through their controlled operation, contribute to reducing grid bottlenecks, the need for grid expansion, and / or optimized grid operation. In this case, control by the control unit is typically carried out indirectly via local control units of the systems.

[0023] The control unit according to the invention is thus designed and / or configured to carry out system-supporting and / or grid-supporting measures. This can occur directly or indirectly. For example, the control unit determines control signals based on the determined probability, which it transmits to system-supporting systems, for example transformers, and / or grid-supporting systems, for example within energy systems. The systems are then controlled according to the transmitted control signals. This results in indirect control of the power grid by the control unit.

[0024] The distributions or probability distributions can be provided discretely and / or as a probability density. In principle, the distributions can be integrated over their associated variables. In the discrete case, integration is understood as a corresponding summation. The distributions are typically multidimensional due to the number of network nodes and / or lines. In particular, they are designed as multidimensional Gaussian distributions or normal distributions.

[0025] In a first step of the method according to the invention, an a priori distribution is provided for the state variables. For example, the distribution of the voltage magnitudes and the angles for each of the network nodes is thereby known. The a priori distribution can be determined and provided based on measurements of the state variables and / or derived variables, such as active power and / or reactive power, over a specified period of time, for example one year. The a priori distribution p(x) thus quantifies the probability p with which a network state or state x exists.

[0026] According to a second step of the present invention, a measurement distribution of measured variables is provided for a given state of the power grid. In other words, the measurement distribution is given by p(z|x), i.e., by the probability distribution of the measurements z for a given state x of the power grid. The measurement distribution can be determined, for example, from measured values ​​of the aforementioned measured variables, in particular power, voltage, current, and the like.

[0027] In a third step of the method according to the invention, a posterior distribution p(x|z) is formed for the state variables x for given measurements z using the prior distribution p(z|x) and the measurement distribution p(x). For this purpose, the Bayesian approach (Bayes' theorem) is particularly preferably used, according to which p(x|z) ≤ p(z|x) · p(x). Thus, after the third step, the statistical distribution of the state variables for given measurements or measured values ​​is determined. The distribution p(x|z) is advantageous because it contains the information about the probability that a network state exists for given measurements. However, this posterior distribution, i.e. after knowledge of the measured values, is not immediately known. However, according to the invention it can be determined from the distribution of the state variables and the measurement distribution, which can be determined a priori, by using Bayes' theorem.

[0028] According to a fourth step of the method according to the invention, a probability for a critical state (critical probability) for at least one of the network nodes is determined using the marginal distribution determined from the a posteriori distribution for the network node. Advantageously, in contrast to known methods which, for example, only use expected values, the entire distribution of the states is used to determine the probability for a critical state of the power grid or the network node. In order to obtain a distribution for each network node, the a posteriori distribution is marginalized in each case, i.e. the other network nodes are integrated or the results are summed over the other network nodes. This allows the marginal distribution or edge distribution to be determined for each network node.The probability of a critical state of the network node is then determined from the determined marginal distribution for at least one of the network nodes. The determination is based on the entire distribution. For example, the probability is determined using an integral over a subset of the marginal distribution. This advantageously includes the entire distribution. Furthermore, a probabilistic statement regarding the presence of a critical network state can be determined.

[0029] In a fifth step of the method according to the invention, the power grid is controlled by implementing grid-friendly and / or system-friendly measures or interventions. According to the invention, these are carried out when the determined probability reaches or exceeds a specified threshold. In other words, measures, for example for grid stabilization, are only implemented when the probability of a critical grid state exceeds a certain value. This advantageously creates a probabilistic and thus non-deterministic approach to grid operation management. For example, the determined probability indicates the probability of a violation of the voltage range of the power grid occurring at one of the grid nodes.Furthermore, several thresholds can be defined, whereby the respective measures depend on the achievement of the respective threshold.

[0030] In other words, according to the present invention, in contrast to using only the expected value as a deterministic estimate, the entire posterior distribution is used to determine the probability of a critical grid state. This improves the control of the power grid.

[0031] According to the present invention, probabilistic approaches, in particular Bayesian methods, are used. This makes it possible to pass stochastic input variables (values ​​subject to uncertainty) to a state estimator. The output variable of the present invention is also a stochastic distribution (a posterior distribution), which is evaluated accordingly. This provides the network operator with information about which outgoing lines and / or network nodes are likely to be affected by a critical state. The control unit then reacts accordingly with control and / or regulation interventions.

[0032] The control unit according to the invention for controlling a power grid with multiple network nodes comprises a computing unit and is characterized in that the computing unit is configured to carry out a method according to the present invention and / or one of its embodiments. In particular, a network control device of a network operator of the power grid comprises the control unit according to the invention.

[0033] Similar, equivalent and equally effective advantages and / or configurations of the control unit according to the invention result from the method according to the invention.

[0034] According to an advantageous embodiment of the invention, the state variables are formed by voltage magnitudes and angles at the network nodes.

[0035] This advantageously utilizes the state variables preferred for electrical networks. The power grid can be multiphase, meaning each phase of each network node has a voltage magnitude and an angle. The angles of the multiple phases can be interdependent.

[0036] In an advantageous development of the invention, the a priori distribution of the state variables is determined and provided by measured values ​​of active power and / or reactive power and their associated measurement deviations and / or their temporal variability, in particular within one year.

[0037] In other words, a distribution of active power and / or reactive power is determined for each network node through measurements. The measurements or measured values ​​can be based on real-time data, historical data, annual electricity consumption and / or forecasts. The a priori distribution of the state variables, in particular the complex-valued voltages (voltage magnitude and angle), can then be determined from the distribution of active power and reactive power. In particular, the temporal variability of the measured values ​​over a specified period of time, in particular one year, is used as a measure of the width of the a priori distribution. The distribution of active power and / or reactive power can also be referred to as the background distribution.Particularly preferred is the a priori distribution of the state variables, in particular the complex-valued voltages, determined from the measured values ​​of the active power and / or reactive power using a load flow calculation. Thus, the a priori distribution of the state variables, in particular the complex-valued voltages, is determined from the background distribution of the active power and / or reactive power using a load flow calculation.

[0038] According to an advantageous embodiment of the invention, the a priori distribution is determined approximately by using a multidimensional Gaussian distribution, wherein the associated expected values ​​and variances are determined from the measured values ​​and the associated measurement deviations.

[0039] This advantageously achieves sufficient accuracy. In other words, the a priori distribution

[0040] Φ priori by a multidimensional normal distribution approximated, that is, it is Here, μ priori the expected value vector (vector of expected values) and ∑ priori the covariance matrix of the measured values ​​.

[0041] For example, where x denotes the vector of state variables .

[0042] For example, x = (Re(V T ), In(V T )) T , where V denotes the vector of complex-valued voltages. The a priori

[0043] The distribution of the complex-valued voltage can be determined from the background distribution of the complex-valued powers (active power and reactive power) using an affine transformation that includes the result of a linear load flow calculation. This transformation is typically implemented in the form of a matrix.

[0044] In an advantageous development of the invention, the measurement distribution is determined and provided by measured values ​​of the state variables, by measured values ​​of active power and reactive power and / or by measured values ​​of currents and the respective associated measurement deviations.

[0045] In this case, it is again particularly advantageous if the measurement distribution is determined approximately by using a multidimensional Gaussian distribution, whereby the associated expected values ​​and the variances are determined from the measured values ​​and the associated measurement deviations.

[0046] In other words, the measurement distribution can be approximated by h(μprior), where h denotes a measurement function that maps the expected values ​​to the measured values. For example, where underlined

[0047] Quantities characterize complex quantities. Furthermore, the apparent powers for N + l network nodes and L lines, that is i ∈ {0 ...N}, fi ∈ {0 ... L} and ti ∈ {0...L} . The first six variables of the measurement function mentioned can be determined using RTU sensors (Remote Terminal Unit) and the remaining variables using PMUs (Phasor Measurement Unit) and the associated measurement errors. The network impedances are each marked with Y, whereby the network topology is taken into account. Furthermore, external sensors connected and / or linked to an RTU can transmit measured values ​​to a control system of the power grid.

[0048] Thus, the measurement distribution can be approximately determined from the active power P and the reactive power Q, i.e. the apparent power S, using the measurement function.

[0049] According to an advantageous embodiment of the invention, the posterior distribution is determined by the product of the prior distribution and the measurement distribution. This advantageously uses a Bayesian approach to determine the posterior distribution of the state variables. If the prior distribution and the measurement distribution are designed as a multidimensional Gaussian distribution, then the posterior distribution advantageously also forms a multidimensional Gaussian distribution, i.e. • The expected value vector μposteriori and the covariance matrix ∑posteriori are uniquely defined by measure|true, priori as well as ∑measure|true and ∑ priori determinable and certain. In other words, the product of two Gaussian distributions results in a Gaussian distribution.

[0050] In an advantageous development of the invention, the probability of the critical state is determined by an integral over the marginal distribution, wherein the integral extends up to a threshold value defined for the state variable.

[0051] The marginal distribution of a network node is determined by forming a trace (integration or sum) over the other network nodes. The marginal distribution is therefore a one-dimensional distribution. If the posterior distribution is a multidimensional Gaussian distribution, then the marginal distribution forms a one-dimensional Gaussian distribution. The probability of the critical state is now advantageously calculated for one of the network nodes according to p critical = determined if a lower threshold V lower limit ( limit value ) for the state variable X of the network node is specified , or according to if an upper threshold X upper limit(limit value) for the state variable X of the network node. The probabilities can be specified in percent, whereby the mentioned integrals include a factor of 100%. For several network nodes, the probabilities for a critical state can be combined into a vector p critical The lower limit value X is therefore preferred as the threshold value for the state variable. lower limit or upper limit X upper limit a range of values ​​permissible for the state variable is used.

[0052] The state variable is particularly preferably the magnitude of the voltage, i.e. X = |V| . The permissible value range for the magnitude of the voltage, or the voltage, extends, for example, in Germany from 207 volts (lower limit) to 253 volts (upper limit). Thus, the permissible range is characterized by ± 10 percent of the nominal value of 230 volts. Other countries, such as the USA, also have a symmetrical percentage band around their respective nominal value as the permissible range.

[0053] Preferably, the permissible value range is formed by ±X percent of a nominal value of the state variable, where X has a value between 0 and 10, or if not specified as a percent, between 0 and 0.1.

[0054] This advantageously creates a symmetrical permissible percentage band for the state variable.

[0055] In particular, according to an advantageous embodiment of the invention, the permissible value range is formed by ± 10 percent, ± 6 percent, ± 5 percent or ± 4 percent of a nominal value of the state variable.

[0056] In particular, different network operators may have different permissible value ranges. Typically, network operators in Germany use the values ​​± 10 percent, ± 6 percent, or ± 4 percent, so that the limit value used to determine the critical probability is advantageously aligned with the permissible value ranges used.

[0057] In an advantageous development of the invention, an active power feed, a reactive power feed, a display of a status message, a switching operation, a control of one or more controllable local network transformers and / or a reduction in power are / is carried out by the control unit as system-serving measures.

[0058] This allows resources typically available within the power grid to be used advantageously for control and thus for mitigating or preventing the critical condition.

[0059] It is particularly preferred if the status message comprises at least three levels, each of the levels being defined by a value range for the determined probability.

[0060] In other words, a traffic light system comprising three levels, identified, for example, by the colors green (normal), yellow (warning), and red (alarm), is advantageous. Based on the levels or warning levels, more specific measures for controlling the power grid can thus advantageously be provided. This is advantageous because, in principle, interventions should be as small and as small as possible. For example, at the second level (yellow / warning), different grid-friendly and / or system-friendly measures can be planned and implemented than at the third level (red / alarm).

[0061] The three levels are particularly preferably defined by the value ranges from 0 percent to 5 percent, from greater than 5 percent to 15 percent and by greater than 15 percent.

[0062] In other words, the three levels are defined by 0 < p critical < 5% , 5% < p critical< 15% and Pcritical > 15%. This advantageously creates a particularly advantageous three-stage plan for avoiding critical network conditions. Further advantages, features, and details of the invention emerge from the exemplary embodiments described below and from the drawings. The following schematically show:

[0063] Figure 1 is a flow diagram of an embodiment of the method according to the invention;

[0064] Figure 2 shows an example of a priori distribution, measurement distribution and a posteriori distribution;

[0065] Figure 3 shows an example marginal distribution for a network node; and

[0066] Figure 4 shows a traffic light system of critical probabilities for a power grid.

[0067] Elements of the same type, value or effect may be provided with the same reference symbols in one or more of the figures.

[0068] Figure 1 shows an exemplary flow diagram of an embodiment of the present invention.

[0069] For this purpose, Figure 1 shows a control unit 1 according to an embodiment of the present invention, which is designed and configured to control a power grid based on a determined probability of a critical state of the power grid.

[0070] The power grid comprises several network nodes. Typically, a critical state of the power grid is characterized by a limit value for the operation of the power grid being violated, i.e., being below or exceeded, at at least one of the network nodes. For example, there is a violation of the permissible voltage range, which has a lower and upper limit value. The present method is designed to determine a probability for a critical state of one or more network nodes and, based on the determined probability, to intervene to stabilize the grid so that the probable critical state is mitigated or, in the best case, does not occur.

[0071] A first step S 1 of the method shown comprises three sub-steps S la, S lab, S lb .

[0072] In the first sub-step S1a, a background distribution is determined. The background distribution corresponds to the statistical distribution of the active power and reactive power, i.e. the apparent power, at each network node in the power grid. The background distribution can be determined from recorded real-time measurements for the power, from historical data (SM data), from annual energy consumption and / or from forecasts, for example for the feed-in power of one or more photovoltaic systems. The measurements can include active power, reactive power, voltage magnitudes, voltage angles (angles), current magnitudes and / or current angles. The aforementioned data from which the background distribution is determined is identified by the zeroth step S0 and forms what is known as input data. In this case, a background distribution is determined from each of the network nodes if possible.For example, the active power is measured at several network nodes. This typically varies at each network node. Through multiple measurements, the distribution of active power at each of the network nodes used can be determined.

[0073] From the background distribution or background distributions determined in step S1a, the a priori distribution of the complex-valued voltages is determined by an affine transformation of a load flow calculation, which typically maps the complex power to the complex voltages (see reference symbol Slab). The a priori distribution is thus determined in sub-step S1b. For this purpose, it can further be assumed that the a priori distribution of the state variables, preferably the complex voltages, is approximated by a multi-dimensional Gaussian distribution or normal distribution, whereby the associated expectation vector is formed by the mean values ​​of the state variables, preferably the complex voltages, and the associated covariance matrix of the state variables, preferably the complex voltages, is formed by the measurement deviations of the state variables.

[0074] In a second step S2 of the method, a measurement distribution of measured variables for a given state of the power grid is provided. The measurement distribution is determined using measured values ​​of the recorded measured variables. The measured variables are, in particular, the complex powers (active power and reactive power), the complex voltages and the complex currents. The respective measured values ​​and their associated required variances can be determined using RTU sensors (Remote Terminal Unit) and PMUs (Phasor Measurement Unit). In particular, the complex powers and the voltage values ​​can be determined using RTU sensors. The voltage angles and complex currents can be determined, in particular, using PMUs. Furthermore, the network topology of the power grid can be used to determine the measurement distribution.The input data mentioned for determining the measurement distribution are again marked as input with the reference symbol S 0 .

[0075] The measurement distribution can be approximately determined using a multidimensional Gaussian distribution or normal distribution. This is defined by an expected value vector and a covariance matrix. The expected value vector of the measured variables, preferably the complex power values, is determined by the mean values ​​of the measured variables. The covariance matrix is ​​determined by the measurement deviation of the measured variables. This allows the measurement distribution to be determined based on the aforementioned measurements.

[0076] In a third step S3 of the method, a posterior distribution is determined using the prior distribution and the measurement distribution. In the present case, this is done using Bayes' theorem. If the prior distribution of the state variables x, in particular the complex voltages, is denoted by p(x) and the measurement distribution of the measured variables z, in particular the complex powers, for a given state x is denoted by p(z|x), then the posterior distribution p(x|z) can be determined by p(x|z) = N · p(z|x) · p(x), where N is a proportionality factor. The proportionality factor N is independent of x and can be determined by ensuring that the integral or the sum of the probabilities p(z|x) has the value 1.

[0077] The determined a posteriori distribution is advantageous because it determines the distribution of the state variables, particularly the complex voltages, for given measurements of the measured variables, especially the complex power. This corresponds to typical operation of the power grid, where the state cannot be measured or is insufficiently measured. The measured variables, however, especially the complex power, can be measured sufficiently well.

[0078] In a fourth step S4 of the method, a probability for a critical state for at least one of the network nodes is determined by means of the marginal distribution determined from the a posterior distribution for the network node.

[0079] The marginal distribution is determined by integration over the other network nodes. This can be carried out for each of the network nodes, so that a marginal distribution is available for each of the network nodes. The respective marginal distribution is then integrated up to a lower or upper limit for the state variable at the network node in question. In particular, the aforementioned limit values ​​restrict a permissible voltage range, for example from 207 to 253 volts. For each of the limit values, in particular for 207 volts and for 253 volts, a probability of its violation can then be determined from the marginal distribution for each of the network nodes.

[0080] In a fifth step S5 of the method, the control unit 1 implements grid-friendly and / or system-friendly measures if the determined probability is greater than or equal to a specified threshold. For example, the first measures are implemented if the voltage range is violated with a determined probability of greater than 5 percent. Second measures can preferably be implemented if the voltage range is violated with a determined probability of greater than 15 percent. The same or different thresholds can be specified for the network nodes.

[0081] Alternatively or additionally, the most probable state of the power grid can also be determined in the fifth step S5. This is done by maximizing the posterior distribution (maximum posterior estimator; abbreviated to MAP). In other words, the state that maximizes the posterior distribution is determined. This corresponds to the most probable state.

[0082] Figure 2 shows an example of an a priori distribution 6 , a measurement distribution 7 and an a posteriori distribution 42 .

[0083] The value of a state variable of a network node, for example, the voltage magnitude at the network node, is plotted on the abscissa 101 of the diagram shown in Figure 2. The value of the respective distribution is plotted on the ordinate 102.

[0084] Here, distributions 6, 7, and 42 were each approximated by a Gaussian distribution. Figure 2 shows that the posterior distribution 42 is sharper than the prior distribution 6 and the measurement distribution 7. This is because it includes the information from the prior distribution 6 and the measurement distribution 7. This advantageously allows for an improved determination of whether a critical state is likely. Based on this, improved control of the power grid can be achieved.

[0085] According to Bayes' theorem, p(x|z) = p(z|x) · p(x) / p(z) . Here, p(x|z) is the posterior distribution 42 , p(z|x) is the measurement distribution 7 , and p(x) is the prior distribution 6 of the state variables. The further prior distribution of the measurements p(z) is independent of the true state x and can be assumed to be constant. Thus, p(x|z) ∞cp(z|x) · p(x) .

[0086] Figure 3 shows the marginal distribution 43 of a network node of the power grid.

[0087] The value of a state variable of a network node, for example, the voltage magnitude at the network node, is plotted on the abscissa 101 of the diagram shown in Figure 3. The value of the respective marginal distribution 43 is plotted on the ordinate 102.

[0088] The marginal distribution 43 is formed from the determined posterior distribution by summing or integrating the other network nodes. In other words, the other network nodes are integrated. This allows a one-dimensional distribution, the marginal distribution 43, to be determined from the multidimensional posterior distribution for each network node. The marginal distribution 43 describes the distribution of the state variables of a network node when the realizations of the other network nodes are not taken into account. For example, if V1 and V2 are the voltage magnitudes of a first and second network node and p(V1,V2l z) is the joint multidimensional posterior distribution, then the marginal distribution for the first network node can be determined by . The mar- The initial distribution of the second network node results accordingly. Figure 3 also shows a lower limit 5 for the value of the state variable, in particular the voltage magnitude. Furthermore, an upper limit, as in the permissible voltage range, can be provided alternatively or additionally.

[0089] By integrating the marginal distribution 43 up to the lower limit 5, the area 4 under the marginal distribution 43 up to the limit 5 is determined. This determines a probability 4 for violating the lower limit 5. This corresponds to the determined probability for a critical network state.

[0090] In other words, the voltage or the

[0091] Voltage amount for the lower limit and for the upper limit . Is a If one of the determined probabilities for a critical state of the power grid, in this case for the violation of the voltage range from 203 volts to 253 volts, is greater than or equal to a specified threshold value, then the control unit initiates system-friendly and / or network-friendly measures which, at best, prevent the critical state or at least mitigate it.

[0092] Figure 4 shows a three-stage warning system based on the determined probability of a critical state of a network node 3 or a power grid 2.

[0093] The illustrated three-stage warning system comprises the stages "Normal", "Warning", and "Alarm" 301, 302, 303 as operating stages or warning stages of the power grid 2. These can also be represented using a color system, comparable to a traffic light. The color green is assigned to the first stage 301 "Normal", the color yellow to the second stage 302 "Warning", and the color red to the third stage 303 "Alarm". The various colors mentioned are represented in the figure by different hatchings. Basically, the stages 301, 302, 303 or warning stages are identified by value ranges of the determined probability for a critical state (critical probability). The first stage 301 thus has a critical probability of 0 percent to X percent, preferably from 0% to 5%.The second stage 302 has a critical probability of X percent to Y percent, preferably from 5% (excluded) to 15%. The third stage 303 has a critical probability above Z percent, preferably above 15% (excluded). Here, X <Y<Z .

[0094] Furthermore, Figure 4 shows a power grid 2 with a plurality of network nodes 3. For each of the network nodes 3 of the power grid 2, a critical probability for a state variable of the power grid 2 at the respective network node 3 can be determined according to the present invention and / or one of its embodiments. In particular, the state variable is the voltage or the voltage magnitude of the complex-valued voltage. However, one of the network nodes 3 can have a plurality of state variables, and for each of the state variables, a critical probability can be determined from associated limit values ​​according to the present invention and / or one of its embodiments.

[0095] In Figure 4, the network nodes 3 of the power grid 2 are colored green, yellow, and red, or shaded accordingly, according to their determined critical probability. This makes it clear when and at which point in the power grid 2 a problem is likely to occur and at which point in the power grid 2 grid-friendly and / or system-friendly measures are required.

[0096] Although the invention has been illustrated and described in detail by the preferred embodiments, the invention is not limited by the disclosed examples and other variations can be derived therefrom by those skilled in the art without departing from the scope of the invention.

[0097] Reference symbol list

[0098] 1 control unit

[0099] 2 Power grid

[0100] 3 network nodes

[0101] 4 Probability of critical condition

[0102] 5 Threshold value for state variable

[0103] 6 A priori distribution

[0104] 7 Measurement distribution

[0105] 42 A-posteriori distribution

[0106] 43 Marginal distribution

[0107] SE entrance

[0108] 101 Abs zisse

[0109] 102 Ordinates

[0110] 301 first stage

[0111] 302 second stage

[0112] 303 third stage

[0113] Sla first step - Determination of background distribution

[0114] Slab first step - load flow calculation

[0115] Slb first step - Determination of a priori distribution

[0116] 52 second step

[0117] 53 third step

[0118] 54 fourth step

[0119] 55 fifth step

Claims

Patent claims 1. Method for controlling a power grid (2) with several network nodes (3) by means of a control unit (1), characterized by the following steps: - (Sla, Sib, Slab) Providing an a priori distribution (6) for state variables of the power grid (2) ; - (S2) providing a measurement distribution (7) of measured variables for a given state of the power grid (2); - (S3) forming a posterior distribution (42) for the state variables by means of the prior distribution (6) and the measurement distribution (7); - (S4) determining a probability (4) for a critical state for at least one of the network nodes (3) by means of the marginal distribution (43) determined from the a posterior distribution (42) for the network node; and - (S5) Implementation of grid-friendly and / or system-friendly measures by the control unit (1) if the determined probability (4) is greater than or equal to a specified threshold value.

2. Method according to claim 1, characterized in that the state variables are formed by voltage amounts and angles at the network nodes (3).

3. Method according to claim 1 or 2, characterized in that the a priori distribution (6) is determined and provided by measured values ​​of active power and / or reactive power and their associated measurement deviations and / or their temporal variability.

4. Method according to claim 3, characterized in that the determination of the a priori distribution (6) from the measured values ​​of the active power and / or reactive power is carried out by means of a load flow calculation (slab).

5. Method according to claim 3 or 4, characterized in that the a priori distribution (6) is approximated by a using a multidimensional Gaussian distribution, whereby the corresponding expected values ​​and variances are determined from the measured values ​​and the corresponding measurement deviations.

6. Method according to one of the preceding claims, characterized in that the measurement distribution (7) is determined and provided by measured values ​​of the state variables, by measured values ​​of active power and reactive power and / or by measured values ​​of currents and the respectively associated measurement deviations.

7. The method according to claim 6, characterized in that the measurement distribution (7) is determined approximately by using a multidimensional Gaussian distribution, wherein the associated expected values ​​and the variances are determined from the measured values ​​and the associated measurement deviations.

8. Method according to one of the preceding claims, characterized in that the a posteriori distribution (42) is determined by the product of the a priori distribution (6) and the measurement distribution (7).

9. Method according to one of the preceding claims, characterized in that the probability (4) for the critical state is determined by an integral over the marginal distribution (43), wherein the integral extends up to a threshold value (5) defined for the state variable.

10. The method according to claim 9, characterized in that the lower or upper limit value of a permissible value range for the state variable is used as the threshold value (5) for the state variable.

11. Method according to claim 10, characterized in that the permissible value range is defined by ±10 percent, ±6 percent, ±5 percent or ±4 percent of a nominal value of the state variable.

12. Method according to one of the preceding claims, characterized in that as system-serving measures an active power feed, a reactive power feed, a display of a status message, a switching operation, a control of one or more controllable local network transformers and / or a reduction are / is carried out by the control unit (1).

13. The method according to claim 12, characterized in that the status message comprises at least three levels (301, 302, 303), each of the levels (301, 302, 303) being defined by a range of values ​​for the determined probability (4).

14. Method according to claim 13, characterized in that the three stages (301, 302, 303) are defined by the value ranges from 0 percent to 5 percent, from greater than 5 percent to 15 percent and by greater than 15 percent.

15. Control unit (1) for controlling a power grid (2) with a plurality of network nodes (3), comprising a computing unit, characterized in that the computing unit is configured to carry out a method according to one of the preceding claims.

Images