Patent Application
20140365419
The invention relates to a method for determining a power outage probability of an electrical power grid, a method for an adaptation of a power generation capacity and a method for determining an energy storage unit size. It also relates to a data processing unit.
Electrical power grid systems usually combine power generated by fuel-based power plants (i.e. coal, oil, gas) with variable energy sources from wind, solar and micro-hydel. Fuel-based power plants must be scaled up and down to match the rise and fall of energy production from the variable energy sources and varying demands.
Two basic methods are used to deal with changing demands. First, the fuel-based power plants can be run below their maximum output and then quickly increase the amount of generated power when needed (spinning reserve). Second, additional power plants (e.g. based on combustion gas turbines) can be brought up (in a matter of minutes) to provide a larger generating capacity. Both above methods are costly. Spinning reserve plants running below maximum output usually produce at less than their best efficiency and the expensive generating equipment at the additional power plants remain unused much of the time.
Apart from cost issues, existing power generation systems adjust the power based on the knowledge of deterministic demands provided in advance by daily or monthly usage profiles. They are not designed to handle real-time stochastic energy sources and demands that arise from grid-connected intermittent energy sources such as solar cells and wind turbines.
Grid energy storage refers to methods used to store electricity at large-scale in the electrical power grid. Electrical energy is stored during times when production from power plants exceeds consumption and the energy of the storage is used when consumption exceeds production. Energy storage has two potential benefits. First, it can increase efficiency and lower the cost of energy production. Energy storage can reduce the peak of generated power and power plants need not be drastically scaled up and down to meet momentary consumption. This has the advantage that fuel-based power plants can be operated more efficiently and easily at lower power production levels. Second, it can facilitate the use of variable energy sources and demands. Using storage, an operator of a power grid can adapt energy production to energy consumption, both of which can vary randomly over time.
Despite the high potential and increased usage of energy storage in electric power grids, it is not yet clear how to design and operate such a system in an efficient manner. A potential design approach would be to size the energy storage before system operation in order to ensure that the demand is always met.
A known method of sizing a battery consists of determining the specific demand requirements and selecting a battery size capable of supplying that load for the specified time. ANSI/IEEE 485 is the industry reference for this type of cell sizing. ANSI/IEEE 1115, IEEE recommended practice for sizing nickel-cadmium batteries for stationary applications, provides equivalent sizing information for nickel-cadmium batteries. Both methods assume a deterministic demand duty cycle and size the battery based on the highest section of the duty cycle. This yields a conservative design when the peak load of the worst duty cycle is much higher than the average and cannot be applied to the case of stochastic energy sources and/or demands.
Some research papers refer to battery sizing techniques for stochastic energy sources.
A first article from P. Arun, R. Banerjee and S. Bandyopadhyay, “Sizing Curve for Design of Isolated Power Systems”, Advances in Energy Research (AER-2006), proposes a design space approach for photovoltaic systems with energy storage, where the design space consists of all feasible power generation and storage sizes in order for the system to operate properly.
A second article from A. Roy, S. Kedare, S. Bandyopadhay, “Design of Wind Power Generation systems for Industrial Application Incorporating Resource Uncertainty.”, Chemical Engineering Transactions, Volume 18 (2009), extends the teachings of the first article by proposing a design space methodology for sizing and optimizing a wind-power battery system by incorporating the uncertainty of the wind resource in the design stage using a reliability target.
The methods described in these articles have the following limitations. First, they both assume a priori knowledge of demands. Second, the second article assumes that wind generation power follows a Weibull distribution and the method works only for that distribution.
It is the object of the invention to provide improved technologies for dimensioning and controlling an electrical power grid.
This object is solved by a method according to the independent claim 1, a method according to the independent claim 8, a method according to the independent claim 9 and a data processing unit according to the independent claim 10. Advantageous embodiments of the invention are the subject of dependent claims.
According to one aspect, the invention provides a method for determining a power outage probability of an electrical power grid, in particular a smart grid, wherein a power generation facility and an energy storage unit are used to distribute power to at least one load unit, for a time period, the method comprising the following steps carried by a processor of a data processing unit:
According to another aspect, the invention provides a method for an adaptation of a power generation capacity of an electrical power grid, in particular a smart grid, the method comprising the following steps:
According to another aspect, the invention provides a method for determining an energy storage unit size for an electrical power grid, in particular a smart grid, comprising a power generation facility and a load unit, the method comprising the steps:
According to another aspect, the invention provides a data processing unit for determining a power outage probability of an electrical power grid, in particular a smart grid, wherein a power generation facility and an energy storage unit are used to distribute power to at least one load unit, for a time period, the data processing unit comprising a processor for:
The electrical power grid comprises a power generation facility with a power generation capacity, an energy storage unit, e.g. a battery, and a load unit with a demand of electrical energy, for example a household. The demand of the load unit must be matched by the power generation capacity. When the demand of the load unit is less than the power generation capacity, the remaining energy is stored in the energy storage unit. When the load unit demand exceeds the power generation capacity, the energy storage unit provides the additional capacity to serve the demand of the load unit if the energy storage unit is not empty. A power outage occurs when the load unit demand exceeds the power generation capacity and there is no electrical energy in the energy storage unit.
The method is based on a probabilistic framework for the computation of the power outage probability. It derives from the notion of an effective bandwidth used in teletraffic theory applying large deviations analysis to data buffers fed by stochastic sources in telecommunication systems. The method is based on the observation that the energy storage unit size can be modelled as a “reverse” data buffer, where the data source is mapped to a load unit demand and the buffer transmission capacity serving the source is mapped to the power generation capacity satisfying the demand.
An effective load unit demand is determined from the load unit demand for each time interval of a time period. A grid parameter depends on the power generation capacity, the energy storage unit size and the effective load unit demand. The grid parameter is optimized for its maximum value for all time intervals of the time period. Due to the optimization of the grid parameter, a particular distribution of the power generation capacity and/or the load unit demand does not have to be assumed. The power outage probability is then computed from the grid parameter.
In some situations it might be necessary to react to a varying load unit demand by adapting the power generation capacity. If the power outage probability for a given set of the power generation capacity, the energy storage unit size and the load unit demand is smaller than a target reliability threshold, the power generation capacity has to be adjusted, for example increased, to avoid a power outage. Preferably, the load unit demand is provided as a real-time demand measured by a meter. This allows an operator of the power grid to adjust the power generation capacity on short notice and assure that the demand is met. By repeated adjustment of the power generation capacity and comparison of the corresponding power outage probability, a new power generation capacity can be determined that assures that the power outage probability is equal or larger than the target reliability threshold. In a preferred embodiment, adjusting the power generation capacity comprises decreasing the capacity if the corresponding power outage probability is larger than the target reliability threshold. Hereby, the minimum power outage capacity that relates to the power outage probability that is equal to the target reliability threshold can be determined.
Further, in a planning stage for a new electrical power grid or an upgrade of an existing power grid, the size of the energy storage unit can be determined based an the expected load unit demand and the power generation capacity. The load unit demand is provided from a predetermined power usage profile and it is assumed that this profile is valid in the future. By computing the power outage probability for some values of the energy storage unit size and comparing the probability with a target reliability threshold, the energy storage unit size sufficient for the power grid at hand can be determined.
Each of the above methods can be executed by the data processing unit that is connected to the database. Preferably, the processor of the data processing unit executes each step of the above methods, respectively.
In a preferred embodiment, the load unit demand comprises several load unit demand distributions for each time interval, determining the effective load unit demand comprises determining an effective load unit demand distribution for each time interval from each load unit demand distribution, respectively, wherein the load unit demand distributions for each time interval are read-out from the database, and a multiplex parameter is determined that aggregates the several load unit demand distributions, wherein the grid parameter comprises the multiplex parameter and is further optimized for its minimum value with respect to the multiplex parameter by the processor of the data processing unit. The multiplex parameter relates to an aggregate demand of several load units in the power grid.
According to a further embodiment, the load unit demand distributions are stochastic distributions. In contrast to the prior art, no specific form of the load unit demand distributions is expected. Instead, arbitrary stochastic distributions are used to determine the power outage probability. Hereby, a huge flexibility of demand distributions is provided. For example, the load unit demand distributions may refer to the real demands of households, office buildings, public buildings and / or industrial facilities.
In still another preferred embodiment, the load unit demand distributions are provided as predetermined power usage profiles. The predetermined power usage profiles can be measured over a certain time period, for example. Using known profiles, the power outage probability can be determined for similar circumstances in the future. For example, the demand distributions of several households of a district are measured over winter. Assuming that the demand of each household will be the same in winter, the power outage probability can be determined for the next winter for the case that additional households are built in the district.
In a further embodiment, the power usage profiles are related to a daily, weekly, monthly or yearly power usage.
In still a further embodiment of the invention, the load unit demand distributions are provided as measurement values that are measured in real-time and provided to the database by a power meter. Real-time measurement of the demand distributions allows an identification of potential problems in providing electrical energy. If the power outage probability becomes too large, indicating that a reliable energy supply can not be provided, a grid operator can react by increasing the power generation capacity, for example. Preferably, each load unit demand distribution is provided by a power meter that is associated to the respective load unit.
In a preferred embodiment, the measurement values are measured at each load unit or at an energy storage unit.
In another preferred embodiment, the power generation capacity comprises a stochastic power generation distribution. Hereby, the power generation capacity of intermittent energy sources, e.g. solar cells or wind turbines, can be taken into account.
According to a further embodiment, the power generation capacity comprises several individual power generation capacities. An energy supply by several power plants connected to the electrical power grid can be considered.
In still a preferred embodiment, the power generation capacity refers to at least one power generation plant of the following group: nuclear power plant, coal power plant, oil power plant, gas power plant, solar power plant, hydro power plant and wind power plant.
According to another further embodiment, the energy storage unit size comprises several individual energy storage unit sizes. Depending on the size of the electrical power grid and / or on specific regional circumstances, the energy storage unit can be provided as one (large) unit or as several (smaller) units.
Following, the invention will be described in further detail, by way of examples, and not by way of limitation, in the figures of the accompanying drawings, in which like reference numerals refer to similar elements and in which:
An electrical power grid 2, in particular a smart grid, is considered where a power generation facility PGF 4 distributes power to several load units LUs 6, such as homes or industrial facilities, using a battery which serves as an energy storage unit ESU 8. The LUs 6 create electricity demands that must be matched by the PGF 4. When the aggregate demand of the LUs 6 is less than the power generation capacity, the remaining energy is stored in the ESU 8. When the aggregate demand exceeds the capacity, the ESU 8, if non-empty, provides the additional capacity to serve the excess demand of the LUs 6. A power outage, at all LUs 6, occurs when the demands exceed the generation capacity and the ESU 8 has no energy left.
A data processing unit 10, comprising a processor 12 (
The processor 12 implements the steps described in the following, with reference to the flowchart of
First, the ESU 8 considers LU 6 demand distributions during a time period T, divided at step 20 in T/t smaller time intervals of duration t. During each time interval [(i−1)t, it], the demand lj([i−1)t, it]), i=1, . . . ,T/t of each LU j is determined at step 22. The demand distributions of the LUs 6 are either provided in advance (e.g. as daily power usage profiles) or they are measured in real-time using smart grid power meter technologies, for example a meter. If the LU 6 demands are measured they can either be measured and provided by the LUs 6 to the ESU 8 or measured at the ESU 8.
Then, for the time period T, the power outage probability P(outage) is computed, at step 24, as a function of the N demands lj, the ESU 8 size B and the power generation capacity C based on the following formula:
P(outage)=exp(−IN). (1)
The exponent −l in equation (1) is computed by solving the following optimization problem:
Similar to the notion of an effective bandwidth in communication networks, the quantity aj(s,t) in equation (3) can be viewed as an “effective demand” of each LU j. The effective demand takes values between a peak demand and an average demand of LU j. The effective demand can be computed using the following equation:
The optimization problem defined by equation (2) is solved by solving two separate optimization problems.
First, for a fixed t the value s* of s that minimizes J(s,t) is determined:
Second, the value t* of t that maximizes J*(t) is determined:
Both equations (5) and (6) can be solved using numerical techniques which search in the space of the parameters s and t, for example a brute force enumeration. The parameters s and t have the following physical interpretations. The parameter t* represents the most likely time duration until the ESU 8 will become empty and a power outage occurs. The parameter s corresponds to the way the demands of the N LUs 6 are multiplexed and create the aggregate demand that depletes the energy of the energy storage unit 8.
With the method to determine the power outage probability, two fundamental design and control issues of (smart) electrical power grids can be addressed: a real-time adaptation of a power generation capacity and a sizing of an energy storage unit.
The above steps are repeated for different values of B, C and time periods T to yield several values of P(outage). The output of this step is a system design space, namely a set of graphs which quantify the relationship between these quantities and aid in controlling the power generation in real time or sizing the ESU 8.
Examples of applications of the above method are as follows.
While there has been illustrated and described what are presently considered to be the preferred embodiments of the present invention, it will be understood by those skilled in the art that various other modifications may be made, and equivalents may be substituted, without departing from the true scope of the present invention. Additionally, many modifications may be made to adapt a particular situation to the teachings of the present invention without departing from the central inventive concept described herein. Furthermore, an embodiment of the present invention may not include all of the features described above. Therefore, it is intended that the present invention is not limited to the particular embodiments disclosed, but that the invention includes all embodiments falling within the scope of the appended claims.
Expressions such as “comprise”, “include”, “incorporate”, “contain”, “is” and “have” are to be construed in a non-exclusive manner when interpreting the description and its associated claims, namely construed to allow for other items or components which are not explicitly defined also to be present. Reference to the singular is also to be construed as a reference to the plural and vice versa.
A person skilled in the art will readily appreciate that various parameters disclosed in the description may be modified and that various embodiments disclosed and/or claimed may be combined without departing from the scope of the invention.
Thus, the invention can be applied at a substation level with a large energy storage serving homes equipped with smart meters. It can also be applied at the home level with a smaller energy storage serving a plurality of home electrical appliances.
| Number | Date | Country | Kind |
|---|---|---|---|
| 11306813.4 | Dec 2011 | EP | regional |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/EP2012/076603 | 12/21/2012 | WO | 00 | 6/27/2014 |
