Two-dimensional clustering unit combination method with preprocessing

Abstract

The invention relates to the technical field of power system dispatching operation, in particular to a two-dimensional clustering unit commitment problem modeling method with preprocessing, which comprises the following steps of: establishing a unit clustering model and dividing unit groups according to technical parameters of units; according to the unit group operation parameters, a unit clustering unit commitment preprocessing model of a relaxation integer variable is constructed; establishing a time clustering model according to a scheduling result of the preprocessing model, and extracting a typical time period; and in combination with unit clustering and time clustering algorithms, a unit commitment model is embedded, a two-dimensional clustering unit commitment model is constructed, and a long-time-scale unit commitment decision of the power system is obtained. Compared with the prior art, the method has the advantages that the problem of large-scale system unit commitment with the year as the time scale is solved, and the final result retains the constraint on non-continuous variables such as unit start and stop on the aspect of high time resolution, so that the flexibility of the system can be described more strictly and accurately.

Description

Technical Field

[0001] The present invention relates to the technical field of power system dispatching and operation, and in particular to a two-dimensional clustering unit combination method and equipment with preprocessing. Background Art

[0002] With the large-scale access of renewable energy, the seasonal fluctuations of new energy represented by wind energy are obvious, and the operational flexibility of the power system on a long-term scale has become a key issue. It is particularly important to consider the unit combination constraints in the long-term power system operation optimization scenario. At the same time, it is also very important to reduce the problem scale in the calculation process to achieve a balance between effectiveness and feasibility.

[0003] For many years, there have been abundant research results in the field of unit combination and clustering methods at home and abroad. For example, Reference 1 "Distributed Robust Unit Combination with Energy Storage Based on Wind Power Forecast Error Clustering" (Shi Yunhui, Wang Luyu, Chen Wei. Automation of Electric Power Systems, 2019, 43(22): 3-12, 121) clusters wind power forecast errors based on the Dirichlet process Gaussian mixture model, and establishes an uncertainty set that considers the correlation of wind power forecast errors between wind farms. Then, a distributed Robust unit combination model considering energy storage is proposed, and an objective function considering the cycle aging cost of the energy storage system is established. The column constraint generation algorithm is used to solve it, which improves the robustness of the energy storage unit combination model. Reference 2 "Evaluation method for power system flexibility based on clustering of typical operation scenarios" (You Guangzeng, Tang Xiangying, Hu Yan, Tai Nengling, Zhu Xinchun, Li Lingfang. Journal of Shanghai Jiaotong University, 2021, 55(7): 802-813.) proposes an evaluation method for power system flexibility to cope with the uncertainty brought about by the increase of renewable energy sources (such as wind power, photovoltaics and hydropower). In the article, the authors use the improved K-means algorithm to cluster the operation scenarios of new energy and loads to obtain typical operation scenarios of the power system. These typical scenarios evaluate the flexibility of the power system from three perspectives: supply and demand balance, power flow distribution and transmission capacity. The article defines flexibility evaluation indicators, including the flexibility index of supply and demand up and down adjustments within the region and the grid flexibility index, as well as the flexibility index of inter-regional transmission channels. By calculating these indicators, the flexibility of the power system under different typical scenarios can be evaluated, and the comprehensive evaluation index can be obtained by weighting according to the probability of the scenario.

[0004] According to various research results, there are many patent applications involving large-scale unit combination modeling. Patent CN108491983B.1 "A large-scale unit combination optimization method for a multi-stack fuel cell power generation system" proposes a large-scale unit combination optimization method for a multi-stack fuel cell power generation system. Through the K-means clustering method and the Monte Carlo simulation method, the fuel cell stacks are clustered according to the scenario and a probability distribution model is established. Chinese patent application CN202410593079.7 "A typical scenario extraction method and system for a power system" proposes a typical scenario extraction method for a power system. According to the historical operation data of the power system and the grid data of the power system, a full-time production simulation considering the combination of AC power flow units is performed; the simulation results of the full-time production simulation are subjected to data extraction related to the scenario extraction index; the extracted data is clustered to complete the typical scenario extraction of the power system. Chinese patent application CN202110184290.X "Medium-term optimal dispatching method for multi-source power systems based on conditional risk constraints" proposes a risk-averse two-stage medium-term random clustering unit combination model. In the unit clustering process, hydropower units are divided according to river basins, and thermal power units are divided according to optimal power flow; the uncertainty scenario expectations of the second stage are evaluated as a combination of conditional risk value (CVaR) and the expected cost of all uncertainty scenarios.

[0005] However, existing clustering unit combination methods all use a single clustering method to extract typical days, which underestimates unit flexibility and has the inherent defect that it is difficult to realize large-scale system unit combination calculations over a long time scale. Summary of the invention

[0006] The purpose of the present invention is to provide a two-dimensional clustering unit combination method with preprocessing in order to solve the technical problem that the existing large-scale unit combination method is difficult to realize large-scale system unit combination calculation under long time scale.

[0007] The purpose of the present invention can be achieved by the following technical solutions:

[0008] A unit commitment method with two-dimensional clustering with preprocessing, comprising the following steps:

[0009] According to the technical parameters of the units, a unit clustering model is established to divide the unit groups;

[0010] According to the operating parameters of the unit group, a unit clustering unit combination preprocessing model with relaxed integer variables is constructed, wherein the preprocessing model relaxes the number of starting units and the number of units in operation on the grid into continuous variables, and obtains the dispatching results;

[0011] According to the scheduling results of the preprocessing model, a time clustering model is established to extract typical time periods;

[0012] Combining the unit groups divided by the unit clustering model with the typical time periods extracted by the time clustering model, the unit combination model is embedded to construct a two-dimensional clustering unit combination model to obtain the unit combination decision of the power system on a long time scale.

[0013] As an optimal technical solution, the unit clustering method is specifically as follows: k cluster centers are determined by a machine learning method; then, based on the similarity of the unit technical parameters, a cluster center close to it is selected for each unit; each cluster center and the units close to it form k unit groups.

[0014] As a preferred technical solution, the unit technical parameters used for the unit clustering include: maximum output, minimum output, maximum climbing power, minimum online / offline time, output cost, startup cost, and one or more of the total number of units in the corresponding class of the representative unit.

[0015] As a preferred technical solution, the objective function of the unit clustering unit commitment preprocessing model with relaxed integer variables is to minimize the system operation cost:

[0016]

[0017] Among them, T is the total number of time periods; C is the total number of unit groups obtained by unit clustering; They represent the startup cost, fixed generation cost and variable generation cost of the unit group c respectively; with U c，t They represent the number of starting units and the number of units in operation at time t of unit group c, and are relaxed to continuous variables; P c，t Represents the power generation of unit group c at time t.

[0018] As a preferred technical solution, the constraint conditions of the unit clustering unit commitment preprocessing model of the relaxed integer variables include: one or more of unit combination constraints and power balance constraints.

[0019] As a preferred technical solution, the scheduling results solved by the unit clustering unit commitment preprocessing model of the relaxed integer variables include one or more of the unit group power, unit group start and stop commands, and unit group online status.

[0020] As a preferred technical solution, the time clustering model extracts typical time periods as follows:

[0021] The scheduling results of the unit clustering unit combination preprocessing model of the relaxed integer variables are spliced ​​into daily operation parameter vectors; according to the similarity between the daily operation parameter vectors, the daily operation parameter vectors with similar distances in the sense of the two norm are clustered into one category; and the day closest to the mean of the daily operation parameter vectors of each day in the category is taken as the representative day of the category.

[0022] As an optimal technical solution, the daily operating parameters constituted by the scheduling results of the preprocessing model include: one or more of the unit group power, unit group online status, unit group start-up command, and unit group shutdown command solved by the unit clustering unit combination preprocessing model.

[0023] As a preferred technical solution, the objective function of the two-dimensional cluster unit combination model is to minimize the system operation cost:

[0024]

[0025] Among them, P is the total number of typical time periods obtained by time clustering, C is the total number of unit categories, They represent the startup cost, fixed power generation cost and variable power generation cost of the unit group c respectively, U c，p are integer variables representing the number of startup units and the number of units in operation in the network of unit group c in a typical period p, respectively. c，p represents the power generation of the generator group c in a typical period p, It represents the weight coefficient of the representative day, which is the number of hours in the time period set represented by the typical time period p.

[0026] As a preferred technical solution, the constraint conditions of the two-dimensional clustering unit combination model include: one or more of the unit clustering unit combination intra-day constraints, unit clustering unit combination inter-day constraints, and power balance constraints.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] The present invention proposes a unit combination method with preprocessed units and time two-dimensional clustering; firstly, the units are clustered, and a preprocessing model of unit clustering unit combination with relaxed integer variables is constructed to obtain the scheduling result of the preprocessing model; based on the scheduling result of the preprocessing model, the unit combination is time clustered according to daily operation characteristics such as unit output; finally, the time clustering algorithm and the unit clustering algorithm are combined, the unit combination model is embedded, and a two-dimensional clustering unit combination model is constructed, which realizes the solution of the unit combination problem of a large-scale system with a year as the time scale. The final optimization result retains the constraints on non-continuous variables such as unit start and stop at a high time resolution, so that the flexibility of the system can be described more strictly and accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 The present invention is a flowchart of the steps of a unit combination method with two-dimensional clustering with preprocessing. DETAILED DESCRIPTION

[0030] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0031] Example 1

[0032] As a first implementation mode of the present invention, the purpose of this embodiment is to provide a modeling method for a two-dimensional clustering unit commitment problem with preprocessing, which is used to obtain long-time scale unit commitment decisions for large-scale power systems.

[0033] As process Figure 1 As shown, a modeling method for a two-dimensional clustering unit combination problem with preprocessing mainly includes steps S1 to S4.

[0034] S1, establish a unit clustering model and divide the unit groups according to the technical parameters of the unit;

[0035] S2, constructing a unit clustering unit commitment preprocessing model with relaxed integer variables according to the unit group operation parameters;

[0036] S3, based on the scheduling results of the preprocessing model, a time clustering model is established to extract typical time periods;

[0037] S4, combining unit clustering and time clustering algorithms, embedding the unit commitment model, constructing a two-dimensional clustering unit combination model, and obtaining the unit commitment decision of the power system on a long time scale.

[0038] The following describes in detail steps S1 to S4 of the modeling method for the two-dimensional clustering unit combination problem with preprocessing in this embodiment.

[0039] Step S1, establishing a unit clustering model and dividing the unit groups according to the technical parameters of the unit.

[0040] Specifically, in this implementation example, the unit clustering method is: determine k cluster centers through machine learning methods; then select a cluster center close to each unit according to the similarity of the unit technical parameters; each cluster center and the units close to it form k unit groups. As shown in Table 1, the unit technical parameters used for clustering are: maximum output, minimum output, maximum ramp power, minimum online / offline time, output cost, startup cost, and one or more of the total number of units in the corresponding class of the representative unit.

[0041] Table 1 Technical parameters of the units used for clustering in the model

[0042]

[0043] Machine learning to process unit classification requires first establishing a vector composed of technical parameters for each unit. and p g Represent the maximum and minimum output of the unit respectively, and the unit technical parameter vector processed by machine learning is as follows:

[0044]

[0045] Then the parameter vector x g , calculate its relationship with each cluster center c i The distance d(x g ,c i )=||x g -c i ||2, the one with the smallest d value corresponds to the cluster center with the closest distance. Step S2, constructing a unit clustering unit combination preprocessing model with relaxed integer variables according to the unit group operation parameters.

[0046] Specifically, in this implementation example, the objective function of the unit clustering unit commitment preprocessing model with relaxed integer variables is to minimize the system operation cost:

[0047]

[0048] Among them, T is the total number of time periods, C is the total number of units, They represent the startup cost, fixed power generation cost and variable power generation cost of class C units respectively. with U c，t They represent the number of starting units and the number of units in operation at time t of unit group c, respectively. In this relaxation model, they are relaxed to continuous variables. c，t Represents the power generation of unit group c at time t.

[0049] In this implementation example, the constraint conditions of the unit clustering unit commitment model of the relaxed integer variables include: one or more of unit combination constraints and power balance constraints.

[0050] In this embodiment, the unit combination constraints include:

[0051] Unit output constraints:

[0052] Unit climbing constraints:

[0053] Logical constraints on unit switch action and online status:

[0054] Minimum online and offline time constraints for units:

[0055] in, P c They represent the upper and lower limits of the power generation capacity of the C-type units. They represent the climbing and descending capabilities of the C-class units, TU c , TD c They respectively represent the minimum online and offline time of the representative unit of Class C unit.

[0056] Specifically, in this embodiment, the relaxation model scheduling result includes: one or more of the unit group power, unit group start and stop commands, and unit group online status solved by the unit clustering unit combination model of the relaxed integer variables.

[0057] Step S3, based on the scheduling results of the preprocessing model, a time clustering model is established to extract typical time periods.

[0058] Specifically, in this embodiment, the time clustering algorithm extracts typical time periods as follows: the scheduling results of the unit clustering unit combination preprocessing model of the relaxed integer variables are spliced ​​into daily operating parameter vectors; according to the similarity between the daily operating parameter vectors, similar days are clustered into one category, and the similarity is measured by the distance in the sense of the two norm, and the day closest to the mean of the daily operating parameter vectors of each day in the category is taken as the representative day of the category.

[0059] Specifically, in this embodiment, the daily operating parameters constituted by the scheduling results of the preprocessing model include: one or more of the unit group power, unit group online status, unit group start-up command, and unit group shutdown command solved by the unit clustering unit combination preprocessing model.

[0060] First, in a method similar to step S1, these operating parameters are combined into a vector representing the daily operating characteristics. dk , U dk , They represent the parameters related to the power of the unit group, the online status of the unit group, the startup command of the unit group, and the shutdown command of the unit group on the dkth day. The daily operating parameters on the dkth day are:

[0061]

[0062] Afterwards, according to the similarity between the daily operation parameter vectors, the daily operation parameter vectors with similar distance in the sense of the two norm are grouped into one category, such as the calculation between dk and dj is:

[0063] D dk,dj =||v dk -v djk ||

[0064] Then, within each category, the day closest to the mean of the daily operating parameter vector is taken as the representative day to achieve typical time period extraction.

[0065] Step S4, combining unit clustering and time clustering algorithms, embedding the unit combination model, and constructing a two-dimensional clustering unit combination model.

[0066] Specifically, in this embodiment, the objective function of the two-dimensional clustering unit combination model is to minimize the system operation cost:

[0067]

[0068] Among them, P is the total number of typical time periods obtained by time clustering, C is the total number of unit categories, They represent the startup cost, fixed power generation cost and variable power generation cost of the unit group c respectively, U c，p are integer variables representing the number of startup units and the number of units in operation in the network of unit group c in a typical period p, respectively. c，p represents the power generation of the generator group c in a typical period p, It represents the weight coefficient of the representative day, which is the number of hours in the time period set represented by the typical time period p.

[0069] Specifically, in this embodiment, the constraint conditions of the two-dimensional clustering unit combination model include: one or more of the unit clustering unit combination intra-day constraints, unit clustering unit combination inter-day constraints, and power balance constraints.

[0070] In this example, the intra-day constraints of the unit clustering unit combination include:

[0071] Unit output constraints:

[0072] Unit climbing constraints:

[0073] Logical constraints on unit switch action and online status:

[0074] Minimum online and offline time constraints for unit types:

[0075] Unit output constraints:

[0076] Unit climbing constraints:

[0077] Among them, p represents hours, p g，p is the output of unit g in representative hour p, is the upper limit of the output of unit g, RP t represents the first t hours of the day.

[0078] In this example, the system grid constraints include:

[0079] Line flow constraints:

[0080] Line capacity constraints:

[0081] Among them, F l,p is the power flow of line l at time p, B l,p is the negative reciprocal of the line l reactance, θ l+,p ,θ l-,p is the voltage phase angle at both ends of the line, and F l are the upper and lower limits of the transmission power of line 1 respectively.

[0082] In this example, the power balance constraint is:

[0083]

[0084] In summary, the present invention realizes the solution of the large-scale system unit combination problem on a yearly time scale, and its final optimization result retains the constraints on non-continuous variables such as unit start and stop at a high time resolution, thereby being able to describe the flexibility of the system more strictly and accurately.

[0085] Example 2

[0086] As another embodiment of the present invention, the electronic device of the present invention includes a central processing unit (CPU), which can perform various appropriate actions and processes according to computer program instructions stored in a read-only memory (ROM) or computer program instructions loaded from a storage unit into a random access memory (RAM). In the RAM, various programs and data required for device operation can also be stored. The CPU, ROM and RAM are connected to each other via a bus. An input / output (I / O) interface is also connected to the bus.

[0087] Multiple components in the device are connected to the I / O interface, including: input units, such as keyboards, mice, etc.; output units, such as various types of displays, speakers, etc.; storage units, such as disks, optical disks, etc.; and communication units, such as network cards, modems, wireless communication transceivers, etc. The communication unit allows the device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunication networks.

[0088] The processing unit performs the various methods and processes described above, such as methods S101 to S104. For example, in some embodiments, methods S101 to S104 may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit. In some embodiments, part or all of the computer program may be loaded and / or installed on the device via a ROM and / or a communication unit. When the computer program is loaded into the RAM and executed by the CPU, one or more steps of methods S101 to S104 described above may be performed. Alternatively, in other embodiments, the CPU may be configured to execute methods S101 to S104 in any other appropriate manner (e.g., by means of firmware).

[0089] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0090] The program code for implementing the method of the present invention can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code can be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0091] In the context of the present invention, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0092] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0093] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A two-dimensional clustering unit commitment problem modeling method with preprocessing, characterized in that: The following steps are involved: According to the technical parameters of the units, a unit clustering model is established to divide the unit groups; According to the operating parameters of the unit group, a unit clustering unit combination preprocessing model with relaxed integer variables is constructed, wherein the preprocessing model relaxes the number of starting units and the number of units in operation on the grid into continuous variables, and obtains the dispatching results; According to the scheduling results of the preprocessing model, a time clustering model is established to extract typical time periods; Combining the unit groups divided by the unit clustering model with the typical time periods extracted by the time clustering model, the unit combination model is embedded to construct a two-dimensional clustering unit combination model to obtain the unit combination decision of the power system on a long time scale.

2. The two-dimensional clustering unit combination problem modeling method with preprocessing according to claim 1 is characterized in that: The unit clustering method is specifically as follows: k cluster centers are determined by a machine learning method; then, a cluster center close to each unit is selected according to the similarity of the unit technical parameters; each cluster center and the units close to it form k unit groups.

3. The two-dimensional clustering unit combination problem modeling method with preprocessing according to claim 2 is characterized in that: The unit technical parameters used for the unit clustering include: maximum output, minimum output, maximum ramp power, minimum online / offline time, output cost, startup cost, and one or more of the total number of units in the corresponding class of the representative unit.

4. The two-dimensional clustering unit combination problem modeling method with preprocessing according to claim 1 is characterized in that: The objective function of the unit clustering unit commitment preprocessing model with relaxed integer variables is to minimize the system operation cost: Among them, T is the total number of time periods; C is the total number of unit groups obtained by unit clustering; They represent the startup cost, fixed generation cost and variable generation cost of the unit group c respectively; with U c，t They represent the number of starting units and the number of units in operation at time t of unit group c, and are relaxed to continuous variables; P c，t Represents the power generation of unit group c at time t.

5. The two-dimensional clustering unit combination problem modeling method with preprocessing according to claim 4 is characterized in that: The constraint conditions of the unit clustering unit commitment preprocessing model of the relaxed integer variables include: one or more of unit combination constraints and power balance constraints.

6. The two-dimensional clustering unit combination problem modeling method with preprocessing according to claim 4 is characterized in that: The scheduling result solved by the unit clustering unit commitment preprocessing model of the relaxed integer variables includes one or more of the unit group power, the unit group start and stop command, and the unit group online status.

7. The two-dimensional clustering unit combination problem modeling method with preprocessing according to claim 1 is characterized in that: The typical time periods extracted by the time clustering model are specifically: The scheduling results of the unit clustering unit combination preprocessing model of the relaxed integer variables are spliced ​​into daily operation parameter vectors; according to the similarity between the daily operation parameter vectors, the daily operation parameter vectors with similar distances in the sense of the two norm are clustered into one category; and the day closest to the mean of the daily operation parameter vectors of each day in the category is taken as the representative day of the category.

8. The two-dimensional clustering unit combination problem modeling method with preprocessing according to claim 7 is characterized in that: The daily operation parameters constituted by the scheduling results of the preprocessing model include: one or more of the unit group power, unit group online status, unit group start command, and unit group shutdown command solved by the unit clustering unit combination preprocessing model.

9. The two-dimensional clustering unit combination problem modeling method with preprocessing according to claim 8 is characterized in that: The objective function of the two-dimensional cluster unit combination model is to minimize the system operation cost: Among them, P is the total number of typical time periods obtained by time clustering, C is the total number of unit categories, They represent the startup cost, fixed power generation cost and variable power generation cost of the unit group c respectively, U c，p are integer variables representing the number of startup units and the number of units in operation in the network of unit group c in a typical period p, respectively. c，p represents the power generation of the generator group c in a typical period p, τ p It represents the weight coefficient of the representative day, which is the number of hours in the time period set represented by the typical time period p.

10. The two-dimensional clustering unit combination problem modeling method with preprocessing according to claim 9, characterized in that: The constraint conditions of the two-dimensional clustering unit commitment model include: one or more of the unit clustering unit commitment intra-day constraints, unit clustering unit commitment inter-day constraints, and power balance constraints.

Images