Electro-hydrogen coupling system planning method based on adaptive time granularity

Abstract

The invention discloses an electricity-hydrogen coupling system planning method based on adaptive time granularity, and the method comprises the steps: carrying out the decomposition of an annual net load time sequence based on an STL algorithm, and extracting an intra-day component and a seasonal component; performing longitudinal granularity division on the intra-day components by using Ward clustering to generate typical days, and then performing transverse granulation on the typical days by using Ward clustering to obtain time granularity combinations of different typical days; performing transverse granularity division on the seasonal components by using Ward clustering to obtain a time granularity combination of the seasonal components; and establishing a time granularity structure containing an intra-day component and a seasonal component. Determining a specific time granularity structure based on an improved PSO algorithm, wherein the specific time granularity structure comprises an optimal particle number in each typical day and season component; establishing an electricity-hydrogen coupling system collaborative planning model based on the specific time granularity structure; and solving the electric-hydrogen coupling system collaborative planning model, realizing supply and demand matching on a plurality of time scales, and realizing electric-hydrogen coupling system planning of self-adaptive time granularity.

Description

Technical Field

[0001] The present invention relates to the technical field of planning of renewable energy and hydrogen energy coupling systems, and in particular to a method for planning an electric-hydrogen coupling system based on adaptive time granularity. Background Art

[0002] With global concern about the energy crisis and the significant emission of greenhouse gases, the proportion of installed renewable energy generation capacity has expanded rapidly in recent years. This increasing share of renewable energy and the rapid consumption of power systems, coupled with the need for flexible resource adjustment, have heightened the challenges of power system balance and security across multiple timescales. Hydrogen, with its wide range of flexible adjustments and scalable, long-term storage capabilities, is being adopted as a long-term energy storage technology to improve source-load matching and support the safe and stable operation of power systems with a high proportion of renewable energy.

[0003] However, the coupling of multiple time scales greatly increases the complexity of the planning problem. In particular, hourly operation throughout the year leads to a large number of decision variables in the optimization problem, which may lead to a sharp increase in the amount of calculation, making the planning problem difficult to solve. Reducing the complexity of the model mainly relies on two technical means: time series simplification and device model simplification. The former refers to how to define the time range and simulation time step in the optimization process, and the latter refers to the level of refinement of the device model. Reducing complexity by modifying the refinement of the device model is relatively easy in actual application, but it also has the disadvantages of limited scope of application and lack of dynamic response, and it is impossible to accurately evaluate the performance of the device in the system. However, there are few methods to reduce complexity by using different time series modeling, especially using different time granularity combinations for different energy storage scenarios and time spans. Summary of the Invention

[0004] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide an electric-hydrogen coupled system (EHCS) planning method based on adaptive time granularity. By adaptively optimizing the time granularity, the time cost and complexity of the planning model are significantly reduced while meeting the planning accuracy requirements, and the challenges faced by the current energy system in supply and demand matching and renewable energy integration are effectively addressed.

[0005] The purpose of the present invention is achieved through the following technical solutions:

[0006] A planning method for an electric-hydrogen coupling system based on adaptive time granularity, comprising:

[0007] Based on the Standard Template Library algorithm STL, the annual net load time series is decomposed to extract the intraday component and seasonal component;

[0008] Perform vertical granularity division on the intraday components to generate typical days, then perform horizontal granularity division on the typical days to obtain time granularity combinations of different typical days; perform horizontal granularity division on the seasonal components to obtain time granularity combinations of seasonal components; and establish a time granularity structure that includes intraday components and seasonal components.

[0009] Determine the specific temporal granularity structure, including the optimal number of particles in each typical day and seasonal component;

[0010] Establish a collaborative planning model for the electric-hydrogen coupling system based on a specific time granularity structure;

[0011] Solve the collaborative planning model of the electric-hydrogen coupling system to obtain the capacity values and scheduling plans of various equipment in the electric-hydrogen coupling system.

[0012] Furthermore, the standard template library algorithm STL can extract the long-term fluctuation characteristics and short-term variation characteristics of the annual net load; the long-term fluctuation characteristics are used to match hydrogen energy, and the short-term variation characteristics are used to match battery energy storage.

[0013] The formula for decomposing the annual net load time series based on the Standard Template Library algorithm STL is as follows:

[0014]

[0015] P Im (t) represents the annual net load time series, represents the intraday component of hourly t throughout the year, represents the seasonal component of hourly t throughout the year, Represents the spurious component at each hour t throughout the year.

[0016] Furthermore, the granularity division of intraday components, typical day components and quarterly components is carried out using Ward's system clustering method based on the sum of squared deviations.

[0017] Furthermore, the specific time granularity structure is determined based on the improved particle swarm optimization algorithm PSO, including:

[0018] When using Ward clustering to divide the intraday components into vertical granularity, first divide the intraday components into Converted into a 365×24 matrix, each row is an initial particle x i , i=1,2...,365; each initial particle is regarded as an original cluster and merged recursively to minimize the variance within the cluster until the convergence condition is met;

[0019]

[0020] During the granulation process, the centers of clusters I and J are determined by formula (2): and Formula (3) calculates the dissimilarity D(I,J) between each pair of adjacent clusters I and J; where N I and N J are the number of particles in cluster I and cluster J respectively. The two nearest neighbor clusters are merged according to the dissimilarity matrix; the cluster center is taken is the particle cluster S after longitudinal granulation k The typical particle is the typical day k; particle cluster S k The number of particles in is the frequency of the typical day. After the daily components are divided into vertical granularity to generate typical days, a suitable granularity combination needs to be found horizontally for each typical day. Then, the same method as the vertical granulation is used to recursively granulate the 24 hours within the typical day.

[0021] Seasonal component Ward clustering is used to perform horizontal granulation to form S particles, each of which contains several hours. The value of each particle in the seasonal component after clustering is defined as P season (s), s=1,2,…,S,P season (s) represents the seasonal component of the net load in the particle, and the value of each hour in the same particle is the same; the number of particles obtained after horizontal granulation on a typical day k is defined as D k , k=1,2,...K; each particle also contains several hours, and the value of each particle in the intraday component after clustering is defined as P day (k,n)k,n=1,2,3...K, represents the value of the intraday component of the net load in the nth particle on the kth typical day;

[0022] In the same particle of the intraday component, the equipment output status remains unchanged in short cycles within 5 hours; in the same particle of the quarterly component, the equipment output status remains unchanged in long cycles of more than 5 hours.

[0023] Furthermore, when solving the coordinated planning model of the electric-hydrogen coupling system, the input includes a time granularity combination of quarterly components and a time granularity combination of typical days, which are used for seasonal peak regulation and intraday source-load mismatch adjustment.

[0024] Furthermore, in the collaborative planning model of the electric-hydrogen coupling system, the annual operation and maintenance cost of the electric-hydrogen coupling system EHCS is used as the objective function, and the constraints include long-cycle equipment constraints, short-cycle equipment constraints and electric-hydrogen balance constraints.

[0025] The present invention also provides an electric-hydrogen coupling system planning device based on adaptive time granularity, comprising:

[0026] Net load decomposition module, used to decompose the annual net load time series and extract the intraday component and seasonal component;

[0027] The time granulation module is used to perform vertical granularity division of intraday components to generate typical days; and to perform horizontal granularity division of typical days to obtain time granularity combinations of different typical days; and to perform horizontal granularity division of seasonal components to obtain time granularity combinations of seasonal components; in order to establish a time granularity structure that includes intraday components and seasonal components;

[0028] Optimization solution module, used to determine the specific time granularity structure;

[0029] The electric-hydrogen coupling system collaborative planning module is used to establish an electric-hydrogen coupling system collaborative planning model based on a specific time granularity structure;

[0030] The calculation module is used to solve the collaborative planning model of the electric-hydrogen coupling system.

[0031] Furthermore, the standard template library algorithm STL is used in the net load decomposition module to decompose the annual net load time series; the granularity division of intraday components, typical day components and quarterly components is all performed using the systematic clustering method Ward based on the sum of squared deviations.

[0032] Furthermore, the specific time granularity structure is determined based on the improved particle swarm optimization algorithm PSO, including:

[0033] When using Ward clustering to divide the intraday components into vertical granularity, first divide the intraday components into Converted into a 365×24 matrix, each row is an initial particle x i , i=1,2...,365; each initial particle is regarded as an original cluster and merged recursively to minimize the variance within the cluster until the convergence condition is met;

[0034]

[0035] During the granulation process, the centers of clusters I and J are determined by formula (2): and Formula (3) calculates the dissimilarity D(I,J) between each pair of adjacent clusters I and J; where N I and N J are the number of particles in cluster I and cluster J respectively, and the two nearest neighboring clusters are merged according to the dissimilarity matrix; the cluster center x S is the particle cluster S after longitudinal granulation k The typical particle is the typical day k; particle cluster S kThe number of particles in is the frequency of the typical day. After the daily components are divided into vertical granularity to generate typical days, a suitable granularity combination needs to be found horizontally for each typical day. Then, the same method as the vertical granulation is used to recursively granulate the 24 hours within the typical day.

[0036] Seasonal component Ward clustering is used to perform horizontal granulation to form S particles, each of which contains several hours. The value of each particle in the seasonal component after clustering is defined as P season (s), s=1,2,…,S,P season (s) represents the seasonal component of the net load in the particle, and the value of each hour in the same particle is the same; the number of particles obtained after horizontal granulation on a typical day k is defined as D k , k=1,2,...K; each particle also contains several hours, and the value of each particle in the intraday component after clustering is defined as P day (k,n)k,n=1,2,3...K, represents the value of the intraday component of the net load in the nth particle on the kth typical day;

[0037] In the same particle of the intraday component, the equipment output status remains unchanged in short cycles within 5 hours; in the same particle of the quarterly component, the equipment output status remains unchanged in long cycles of more than 5 hours.

[0038] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the method for planning an electric-hydrogen coupling system based on adaptive time granularity are implemented.

[0039] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the electric-hydrogen coupling system planning method based on adaptive time granularity.

[0040] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0041] 1. Accurately extract net load characteristics: This paper decomposes the annual net load time series into intraday components and seasonal components through the STL decomposition method. It not only extracts the short-term intraday variation characteristics, but also retains the long-term seasonal fluctuation characteristics, providing a data basis for subsequent time granulation and optimization.

[0042] 2. Efficient Time Granulation Technology: By employing Ward clustering to perform both horizontal and vertical granulation of time series, we achieve a combined construction of time granularity for typical daily and seasonal components, forming a multi-perspective, multi-layered time granulation structure. This approach fully accounts for the differences in net load characteristics at different time scales, providing a more precise granularity for optimization solutions.

[0043] 3. Significantly Reduced Computational Complexity: Based on an improved PSO algorithm, the optimal time granularity structure is selected, significantly reducing computational costs while ensuring planning accuracy. The dual optimization design of horizontal and vertical granularity makes the collaborative planning model more flexible in handling multi-time scale characteristics, making it suitable for complex energy system scenarios.

[0044] 4. Integrated Energy Storage System Optimization: This invention combines the long-term energy storage characteristics of hydrogen with the rapid charge and discharge characteristics of batteries to establish a coordinated planning model for a coupled electricity-hydrogen system. This achieves energy supply and demand matching across multiple timescales, effectively addressing seasonal energy imbalances and intraday peak load regulation in power systems with a high proportion of renewable energy, and significantly improving the operational flexibility and stability of the energy system.

[0045] 5. Improving planning efficiency and accuracy: This method, through the introduction of adaptive time granularity, strikes a balance between planning efficiency and model accuracy, balancing the overall accuracy of annual planning with the accuracy of local details. Furthermore, this method is highly versatile and applicable to the planning needs of a variety of complex energy systems.

[0046] 6. Ease of Application and Scalability: The proposed method and device, based on a modular design, can be integrated into existing electric-hydrogen coupling system planning platforms, facilitating practical application. Furthermore, by optimizing the temporal granularity structure, it is highly scalable and can flexibly adapt to the needs of energy scenarios of varying scales and characteristics. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a general framework diagram of the EHCS planning method based on adaptive time granularity in a specific implementation method;

[0048] Figure 2 Schematic diagram of a transverse-vertical granulation method in a specific embodiment;

[0049] Figure 3 is an example of a typical daily time granularity of the method of the present invention in the embodiments;

[0050] Figure 4 is the error evaluation result of each typical day of the method of the present invention in the embodiment;

[0051] Figure 5a and Figure 5bThey are respectively comparative diagrams of intra-day component processing results of different methods in the embodiments of the present invention;

[0052] Figure 6 This is an example of the time granularity of the seasonal component in the embodiment of the present invention. DETAILED DESCRIPTION

[0053] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0054] The present invention provides a method for planning an electric-hydrogen coupling system based on adaptive time granularity. Figure 1 ,include:

[0055] Time decomposition: To achieve reliable seasonal energy storage planning, the annual net load time series is decomposed based on the Standard Template Library (STL) algorithm to extract intraday and seasonal components, effectively preserving long-term fluctuation characteristics.

[0056] Adaptive optimization of temporal granularity: Using seasonal and intraday components as input, Ward clustering, a systematic clustering method based on the sum of squared deviations, is used to achieve vertical and horizontal granularity. Ward clustering is used to perform vertical granularity division on the intraday components to generate typical days, reducing scenario complexity while preserving intraday fluctuation information. Ward clustering is then used to perform horizontal granularity division on the typical days, resulting in temporal granularity combinations for different typical days. Because seasonal component data changes relatively slowly, Ward clustering is used to directly perform horizontal granularity division on the seasonal components to obtain temporal granularity combinations for the seasonal components. A temporal granularity structure is then established that includes both intraday and seasonal components.

[0057] During the horizontal granulation process, the number of particles within the time granularity structure is optimized based on the improved particle swarm optimization algorithm PSO (Wu Xiaogang, Liu Zongqi, Tian Liting, et al. Distribution network energy storage site selection and sizing based on improved multi-objective particle swarm optimization algorithm [J]. Power System Technology, 2014, 38(12):3405-3411. DOI:10.13335 / j.1000-3673.pst.2014.12.021.) to determine the specific time granularity structure. The number of particles corresponds to the granularity division, and the two are in a one-to-one correspondence. The specific time granularity structure includes the optimal number of particles in each typical day and seasonal component. This embodiment takes the power fluctuation and comprehensive cost of the electrolytic cell as the target and obtains the optimal time granularity structure from multiple angles and levels. The time granularity size is made to conform to the characteristics of each energy storage medium, greatly reducing the computational complexity without significantly reducing the planning accuracy.

[0058] A coordinated planning model for the electric-hydrogen coupling system is established based on a specific time granularity structure. This model combines the long-term storage characteristics of hydrogen energy with the rapid charge and discharge characteristics of batteries to address the seasonal energy imbalance and intraday peak load regulation issues of the hydrogen coupling system.

[0059] Solve the collaborative planning model of the electric-hydrogen coupling system, achieve supply and demand matching on several time scales, and realize electric-hydrogen coupling system planning with adaptive time granularity.

[0060] Due to its long-term storage properties, hydrogen energy is primarily used for seasonal peak load regulation. Therefore, the input data for the solution is a combination of quarterly component time granularity. This means that the operation of electrolyzers, hydrogen storage tanks, and fuel cells is analyzed using the quarterly component particles as the time step. Battery energy storage and hydrogen gas turbines, on the other hand, can be used to adjust intraday source-load mismatches due to their flexibility. Therefore, the input data is a typical day and the corresponding time granularity combination.

[0061] Specifically, the horizontal and vertical granulation of time series are as follows:

[0062] This embodiment uses the STL algorithm to decompose the annual net load time series. The trend and seasonal changes of the time series data are estimated by smoothing the time series data, and the residual term is obtained by subtracting it from the original data.

[0063] P Im (t) = P PV (t)-P load (t) (1)

[0064]

[0065] Formula (1) describes the source of energy imbalance, P PV (t),P load (t),P Im (t) represents the photovoltaic power generation power, load power and net load power hourly throughout the year. Formula (2) represents the annual net load time series P extracted based on the STL algorithm Im (t) The hourly intraday component obtained throughout the year Seasonal Components and spurious components

[0066]

[0067] This embodiment uses Ward clustering to granulate the intra-day components horizontally and vertically. When granulating vertically, first granulate the intra-day components Converted into a 365×24 matrix, each row is an initial particle x i(i=1,2...,365). Each initial particle is treated as an original cluster and merged recursively to minimize the variance within the cluster until the convergence condition is met. The purpose of longitudinal granulation is to obtain a representative typical day. During the granulation process, the centers of clusters I and J are determined by formula (3) and Formula (4) calculates the dissimilarity D(I,J) between each pair of adjacent clusters I and J. I and N J are the number of particles in clusters I and J respectively, and the two nearest neighbor clusters are merged according to the dissimilarity matrix. Take the cluster center is the particle cluster S after longitudinal granulation k The typical particle is the typical day k; particle cluster S k The number of particles in the typical day is the frequency of that typical day. After longitudinal granulation into different typical days, a suitable particle size combination must be found in the transverse direction for each typical day. The present invention uses the same method as the longitudinal granulation to recursively granulate the 24 hours within a typical day in the transverse direction.

[0068] Horizontal and vertical granulation diagram Figure 2 As shown, the intraday volume With seasonal components After clustering, a multi-level and multi-granular structure is formed: seasonal component The horizontal granulation forms S particles, each particle contains several hours, and the value of each particle in the seasonal component after clustering is defined as P season (s)s=1,2,…,S, represents the seasonal component of the net load in the particle. The value of each hour in the same particle is the same; the intraday component After vertical granulation, K typical days are formed, and then each typical day is granulated horizontally. The number of particles obtained after horizontal granulation of typical day k is defined as D k ,k=1,2,...K. Each particle may also contain several hours. Similarly, the value of each particle in the intraday component after clustering is defined as P day (k,n)k,n=1,2,3...K, represents the value of the intraday component of the net load in the nth particle on the kth typical day.

[0069] In the same particle of the intraday component, the output status of short-cycle equipment remains unchanged; similarly, in the same particle of the quarterly component, the output status of long-cycle equipment also remains unchanged.

[0070] Specifically, the time granularity adaptive optimization based on the improved PSO algorithm is as follows:

[0071] Different particle numbers will result in different granularity combinations and affect the planning results and operating efficiency of the electric-hydrogen coupling system. This embodiment uses an improved PSO algorithm to optimize the number of particles. The present invention sets two goals, which are expressed as follows:

[0072] f=min{f1,f2}(5)

[0073] In formula (5), f1 is the annual comprehensive cost of the electric-hydrogen fusion system, and f2 is the power fluctuation rate of the electrolyzer in the system.

[0074] f1=C inv +C mc +C oc (6)

[0075]

[0076] In formula (6), C inv is the annual investment cost of the system, C mc is the annual maintenance cost of the system, C oc is the annual operating cost of the system. In formula (7), P EL (s) is the power consumption of the electrolytic cell per hour in the s-th particle of the quarterly component, is the average value of electrolytic cell power throughout the year, N s is the number of hours contained in the sth particle.

[0077] The selection of the global optimal solution of the population is as follows:

[0078] When solving multiple objectives, the non-inferior solution set must be updated after each iteration. In this embodiment, there are two objective functions f1 and f2. In order to maintain the scale of the Pareto solution set and the uniformity of the solution distribution, the particle x i The dense distance is defined as:

[0079]

[0080] In formula (8), x j 、x k is the distance x i The two nearest particles; f1(x j ) and f2(x j ) refer to particles x j The values of the two objective functions; f1 max and is the maximum value of the two objective functions among all particles.

[0081] After solving the dense distance of each Pareto solution, sort them from large to small according to the dense distance, and the one with the largest dense distance is regarded as the global optimal solution.

[0082] Specifically, the collaborative planning model of the electric-hydrogen coupling system is established based on a specific time granularity structure as follows:

[0083] Objective function

[0084] In this embodiment, the collaborative planning model for the electric-hydrogen coupling system is solved with the annual operation and maintenance cost of the electric-hydrogen coupling system EHCS as the objective function to obtain the optimal configuration plan to achieve lower carbon emissions and higher renewable energy utilization.

[0085] F=min{C mc+ C oc}(9)

[0086] Constraints

[0087] 1) Long-cycle equipment constraints

[0088]

[0089] In formula (10), EL is the electro-hydrogen conversion coefficient of the electrolyzer, P EL (s) represents the power of the electrolytic cell in s particles, V EL (s) represents the hydrogen production of the electrolyzer in s particles. and They represent the upper and lower limits of the electrolytic cell power, and Respectively represent the upper and lower limits of the electrolyzer ramp rate. Similarly, formula (11) represents the corresponding operating constraints of the fuel cell. In formula (12), V ST (s) represents the amount of hydrogen in the sth particle of the hydrogen storage device in the quarterly component; and are the amount of hydrogen stored and released by the sth particle in the hydrogen storage device. ST (s) is a 0-1 variable that controls the hydrogen storage and access state in the sth particle of the hydrogen storage device. and are the maximum amount of hydrogen stored and withdrawn per hour; N s is the number of hours contained in the sth particle. is the capacity of the hydrogen storage equipment in the station.

[0090] 2) Short-cycle equipment constraints

[0091]

[0092] In formula (13), Y bat (k,n) is a 0-1 variable representing the access state of the nth particle of the battery on a typical day k. and are the charging and discharging powers of the kth particle of the battery on a typical day. The maximum power of battery charging and discharging; N k,n is the number of hours contained in the nth particle of a typical day k, E bat (k,n) and E bat (k, n-1) are the stored energy of the battery at the beginning of the nth particle and the beginning of the n-1th particle on a typical day k; is the design capacity of the battery. Similarly, equation (14) represents the corresponding operating constraints of the gas turbine.

[0093] 3) Electricity-hydrogen balance constraints

[0094]

[0095] In formula (15), the first equality constraint ensures the intraday power balance of EHCS; the second equality constraint ensures the seasonal power balance; and the third equality constraint ensures the hourly hydrogen balance in each particle throughout the year. LOSS (s) is the amount of hydrogen lost in the sth particle.

[0096] Preferably, the dataset used in this example is the hourly photovoltaic power generation and load demand in a certain area in northern China from January 1, 2019 to December 31, 2019. Based on the original data, the unbalanced power time series is obtained and decomposed into three components using the STL algorithm: seasonal component, intraday component, and stray component.

[0097] (1) Comparison model settings

[0098] In order to verify the effectiveness of the adaptive time granularity planning method proposed in this invention, the following three methods are set up for comparative analysis of planning accuracy:

[0099] Method 1: Based on the hourly time series solution, simulate the operation of each hour (8760 hours) throughout the year without approximation. This method is selected as the benchmark case.

[0100] Method 2: Use uniform time granularity for planning. The selected time granularity is 6 hours for seasonal energy storage equipment and 2 hours for daily energy storage equipment.

[0101] Method 3: The adaptive time granularity planning method proposed in this invention.

[0102] In order to achieve the purpose of controlling variables, method 4 is introduced in the time series error comparison part to achieve a comparison with method 2:

[0103] Method 4: Granulate the selected typical day into 12 time granularities according to the horizontal granulation method of the present invention.

[0104] (2) Model evaluation

[0105] For the granulation results of various methods, the root mean square error (RMSE) is used as the error evaluation indicator:

[0106]

[0107] Among them, the smaller the RMSE index, the better the prediction effect.

[0108] (3) Comparative experimental results

[0109] Figure 3 The time granularity of typical days 1, 7, 13, and 19 is shown in the method proposed in this invention. The time granularity of these four typical days is [14, 15, 17, 17]. The error results of each typical day are as follows Figure 4 As shown, it can be seen that the relative error of the method based on the present invention is the smallest.

[0110] Table 1 shows the number of variables and computation time for the proposed planning method and the comparative method. As can be seen, Method 1, selected as the baseline, has 97,355 variables and a total computation time of 3,255 seconds. Method 3, based on the present invention, reduces the number of variables to 40,553, a 58.35% reduction compared to the baseline, while also reducing the total computation time by 67.31%. This effectively demonstrates the proposed method's superiority in reducing computational costs.

[0111] Table 1 Performance comparison of different planning methods

[0112]

[0113] Figure 5a and Figure 5b The intraday component time series and error distributions of different methods are shown. Figure 5a It shows the results of intraday component time series after being processed by different methods. The picture contains the intraday component curves of methods one to four. Figure 5a You can intuitively see the processing effects of different methods. Figure 5b It reflects the error distribution of the intraday component time series after being processed by different methods. Figure 5b The error of different methods can be seen intuitively. The curve used in method 1 is the original unbalanced power. Figure 5a As can be seen from the figure, in the first 6 hours and the last 6 hours of the day, the three methods produce similar processing results because the unbalanced power changes slowly. However, when the power imbalance changes greatly during the noon period, the processing differences between the three cases are quite large. Figure 5b It can be seen that method 2 has the largest error during this period, and method 3 based on the present invention has the smallest overall error.

[0114] like Figure 6 As shown in the figure, for long-term energy storage devices, since they are not suitable for frequent power changes, the volatility of the unbalanced power is greatly reduced after decomposition by the STL algorithm, and the average granularity length of the seasonal component is greater than 5 hours, which is in contrast to the granularity division of a typical day and is more in line with the long-term and seasonal characteristics of hydrogen storage.

[0115] In summary, the case results show that:

[0116] 1) The proposed method effectively leverages the characteristics of long-term and short-term energy storage systems to determine the optimal time-granularity combination. By optimizing time granularity, the EHCS is more adaptable to changes in system power imbalances, enabling it to better allocate and utilize renewable energy.

[0117] 2) In terms of computing performance, compared with the traditional planning method based on 8760 hours, the average error of the proposed method is only 4.89%, while the computing time is reduced by 67.31%; compared with the fixed time granularity method, the average error of the proposed method is reduced by 9.79%.

[0118] 3) As the demand for EHCS continues to increase, the role of the present invention will become more obvious and help relevant stakeholders make energy planning decisions.

[0119] Preferably, the embodiments of the present application further provide a specific implementation of an electronic device capable of implementing all steps of the method for planning an electric-hydrogen coupling system based on adaptive time granularity in the above embodiment. The electronic device specifically includes the following contents:

[0120] Processor, memory, communications interface, and bus;

[0121] Among them, the processor, memory, and communication interface communicate with each other through the bus; the communication interface is used to realize information transmission between related devices such as server-side devices, metering devices, and user-side devices.

[0122] The processor is used to call the computer program in the memory, and when the processor executes the computer program, it implements all the steps of the electric-hydrogen coupling system planning method based on adaptive time granularity in the above embodiment.

[0123] An embodiment of the present application also provides a computer-readable storage medium capable of implementing all steps of the method for planning an electric-hydrogen coupling system based on adaptive time granularity in the above-mentioned embodiment. The computer-readable storage medium stores a computer program, which, when executed by a processor, implements all steps of the method for planning an electric-hydrogen coupling system based on adaptive time granularity in the above-mentioned embodiment.

[0124] The various embodiments in this specification are described in a progressive manner. Similar portions between the various embodiments can be referenced to each other. Each embodiment focuses on the differences between the other embodiments. In particular, the hardware + program embodiments are generally similar to the method embodiments, so their description is relatively simple. For relevant portions, refer to the description of the method embodiments.

[0125] The foregoing description of this specification describes specific embodiments. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0126] Although the present application provides method operation steps such as embodiments or flowcharts, more or fewer operation steps may be included based on conventional or non-creative work. The order of steps listed in the embodiments is only one way of executing the steps among many steps and does not represent the only execution order. When an actual device or client product is executed, it can be executed in the order shown in the embodiments or the drawings or in parallel (for example, in a parallel processor or multi-threaded processing environment).

[0127] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0128] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1The function specified in one or more boxes.

[0129] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0130] The present invention is not limited to the embodiments described above. The above description of the specific embodiments is intended to describe and illustrate the technical solutions of the present invention. The above specific embodiments are merely illustrative and not restrictive. Without departing from the scope of the present invention and the scope of protection of the claims, those skilled in the art may make various specific modifications based on the teachings of the present invention, all of which fall within the scope of protection of the present invention.

Claims

1. A planning method for an electric-hydrogen coupling system based on adaptive time granularity, characterized in that: include: Based on the Standard Template Library algorithm STL, the annual net load time series is decomposed to extract the intraday component and seasonal component; Perform vertical granularity division on the intraday components to generate typical days, then perform horizontal granularity division on the typical days to obtain time granularity combinations of different typical days; perform horizontal granularity division on the seasonal components to obtain time granularity combinations of seasonal components; and establish a time granularity structure that includes intraday components and seasonal components. Determine the specific temporal granularity structure, including the optimal number of particles in each typical day and seasonal component; Establish a collaborative planning model for the electric-hydrogen coupling system based on a specific time granularity structure; Solve the collaborative planning model of the electric-hydrogen coupling system to obtain the capacity values and scheduling plans of various equipment in the electric-hydrogen coupling system.

2. The method for planning an electric-hydrogen coupling system based on adaptive time granularity according to claim 1, characterized in that: The standard template library algorithm STL can extract the long-term fluctuation characteristics and short-term variation characteristics of the annual net load; Long-cycle fluctuation characteristics are used to match hydrogen energy, and short-cycle change characteristics are used to match battery energy storage.

3. The method for planning an electric-hydrogen coupling system based on adaptive time granularity according to claim 1 or 2, characterized in that: The formula for decomposing the annual net load time series based on the Standard Template Library algorithm STL is as follows: P Im (t) represents the annual net load time series, represents the intraday component of hourly t throughout the year, represents the seasonal component of hourly t throughout the year, Represents the spurious component at each hour t throughout the year.

4. The method for planning an electric-hydrogen coupling system based on adaptive time granularity according to claim 1, characterized in that: The granularity division of intraday components, typical day components and quarterly components all uses the Ward clustering method based on the sum of squared deviations.

5. The method for planning an electric-hydrogen coupling system based on adaptive time granularity according to claim 1, characterized in that: The specific time granularity structure is determined based on the improved particle swarm optimization algorithm PSO, including: When using Ward clustering to divide the intraday components into vertical granularity, first divide the intraday components into Converted into a 365×24 matrix, each row is an initial particle x i , i=1,2...,365; each initial particle is regarded as an original cluster and merged recursively to minimize the variance within the cluster until the convergence condition is met; During the granulation process, the centers of clusters I and J are determined by formula (2): and Formula (3) calculates the dissimilarity D(I,J) between each pair of adjacent clusters I and J; where N I and N J are the number of particles in cluster I and cluster J respectively. The two nearest neighbor clusters are merged according to the dissimilarity matrix; the cluster center is taken is the particle cluster S after longitudinal granulation k The typical particle is the typical day k; particle cluster S k The number of particles in is the frequency of the typical day. After the daily components are divided into vertical granularity to generate typical days, a suitable granularity combination needs to be found horizontally for each typical day. Then, the same method as the vertical granulation is used to recursively granulate the 24 hours within the typical day. Seasonal component Ward clustering is used to perform horizontal granulation to form S particles, each of which contains several hours. The value of each particle in the seasonal component after clustering is defined as P season (s), s=1,2,…,S,P season (s) represents the seasonal component of the net load in the particle, and the value of each hour in the same particle is the same; the number of particles obtained after horizontal granulation on a typical day k is defined as D k , k=1,2,...K; each particle also contains several hours, and the value of each particle in the intraday component after clustering is defined as P day (k,n)k,n=1,2,3...K, represents the value of the intraday component of the net load in the nth particle on the kth typical day; In the same particle of the intraday component, the equipment output status remains unchanged in short cycles within 5 hours; in the same particle of the quarterly component, the equipment output status remains unchanged in long cycles of more than 5 hours.

6. The method for planning an electric-hydrogen coupling system based on adaptive time granularity according to claim 1, characterized in that: When solving the coordinated planning model of the electric-hydrogen coupling system, the input includes a time granularity combination of quarterly components and a time granularity combination of typical days, which are used for seasonal peak regulation and adjustment of intraday source-load mismatch.

7. The method for planning an electric-hydrogen coupling system based on adaptive time granularity according to claim 1, characterized in that: In the collaborative planning model of the electric-hydrogen coupling system, the annual operation and maintenance cost of the electric-hydrogen coupling system (EHCS) is used as the objective function, and the constraints include long-cycle equipment constraints, short-cycle equipment constraints, and electric-hydrogen balance constraints.

8. A planning device for an electric-hydrogen coupling system based on adaptive time granularity, characterized in that: include: Net load decomposition module, used to decompose the annual net load time series and extract the intraday component and seasonal component; The time granulation module is used to divide the daily components into vertical granularity to generate typical days; It also performs horizontal granularization on typical days to obtain time granularity combinations of different typical days; it is also used to perform horizontal granularity division on seasonal components to obtain time granularity combinations of seasonal components; To establish a time granularity structure that includes intraday and seasonal components; Optimization solution module, used to determine the specific time granularity structure; The electric-hydrogen coupling system collaborative planning module is used to establish an electric-hydrogen coupling system collaborative planning model based on a specific time granularity structure; The calculation module is used to solve the collaborative planning model of the electric-hydrogen coupling system.

9. The electric-hydrogen coupling system planning device based on adaptive time granularity according to claim 8, characterized in that: The standard template library algorithm STL is used in the net load decomposition module to decompose the annual net load time series; the granularity division of intraday components, typical day components and quarterly components is all performed using the systematic clustering method Ward based on the sum of squared deviations.

10. The electric-hydrogen coupling system planning device based on adaptive time granularity according to claim 8, characterized in that: The specific time granularity structure is determined based on the improved particle swarm optimization algorithm PSO, including: When using Ward clustering to divide the intraday components into vertical granularity, first divide the intraday components into Converted into a 365×24 matrix, each row is an initial particle x i , i=1,2...,365; each initial particle is regarded as an original cluster and merged recursively to minimize the variance within the cluster until the convergence condition is met; During the granulation process, the centers of clusters I and J are determined by formula (2): and Formula (3) calculates the dissimilarity D(I,J) between each pair of adjacent clusters I and J; where N I and N J are the number of particles in cluster I and cluster J respectively. The two nearest neighbor clusters are merged according to the dissimilarity matrix; the cluster center is taken is the particle cluster S after longitudinal granulation k The typical particle is the typical day k; particle cluster S k The number of particles in is the frequency of the typical day. After the daily components are divided into vertical granularity to generate typical days, a suitable granularity combination needs to be found horizontally for each typical day. Then, the same method as the vertical granulation is used to recursively granulate the 24 hours within the typical day. Seasonal component Ward clustering is used to perform horizontal granulation to form S particles, each of which contains several hours. The value of each particle in the seasonal component after clustering is defined as P season (s), s=1,2,…,S,P season (s) represents the seasonal component of the net load in the particle, and the value of each hour in the same particle is the same; the number of particles obtained after horizontal granulation on a typical day k is defined as D k , k=1,2,...K; each particle also contains several hours, and the value of each particle in the intraday component after clustering is defined as P day (k,n)k,n=1,2,3...K, represents the value of the intraday component of the net load in the nth particle on the kth typical day; In the same particle of the intraday component, the equipment output status remains unchanged in short cycles within 5 hours; in the same particle of the quarterly component, the equipment output status remains unchanged in long cycles of more than 5 hours.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the electric-hydrogen coupling system planning method based on adaptive time granularity according to any one of claims 1 to 7 are implemented.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for planning an electric-hydrogen coupling system based on adaptive time granularity according to any one of claims 1 to 7 are implemented.

Images