A method and system for collaborative layout of substation modules adapted to distributed photovoltaic power generation

Abstract

This invention relates to the field of substation layout optimization technology, and particularly to a method and system for collaborative layout of substation modules adapted to distributed photovoltaic (PV) systems. The method includes: acquiring historical operating data of the substation, extracting power fluctuation sequences, establishing a transmission sensitivity model, identifying the dominant PV fluctuation mode and extracting the dominant period through matrix decomposition; calculating cable transmission delay based on the time constant of the RL circuit composed of cable inductive reactance and resistance, and calculating the total delay time and reactive power regulation mismatch factor by combining controller delay and equipment response delay; selecting cable cross-sectional areas that meet delay constraints based on the target mismatch margin coefficient, back-calculating the maximum cable path length through inductive voltage drop constraints, constructing a layout optimization model and solving for the optimal layout coordinates of each module; placing a three-dimensional geometric model based on the optimal layout coordinates and planning the cable tray laying path. This invention can avoid bus voltage oscillations caused by regulation mismatch due to excessive layout distance.

Description

Technical Field

[0001] This invention relates to the field of substation layout optimization technology, and in particular to a method and system for collaborative layout of substation modules adapted to distributed photovoltaic power generation. Background Technology

[0002] With the continuous growth of distributed photovoltaic (PV) grid integration, the stability of substation bus voltage faces new challenges. PV output is affected by meteorological factors such as cloud cover and sudden temperature changes, resulting in significant power fluctuations on a timescale of several seconds to tens of seconds. The reactive power compensation device within the substation needs to track and adjust reactive power within the corresponding time window to maintain the bus voltage within the allowable deviation range.

[0003] Traditional substation layout methods focus on meeting electrical safety distances, maintenance access widths, and the convenience of civil construction. In engineering design practice, reactive power compensation devices and main transformers are usually arranged in their respective functional areas, based on equipment dimensions, minimum safety distances specified in regulations, and maintenance space requirements. The evaluation indicators for the layout scheme focus on two dimensions: the utilization rate of the substation area and construction feasibility.

[0004] In the prior art, Chinese patent document CN114418145B discloses a GIM-based digital twin method and system for substations. This method constructs a digital twin substation morphology by performing a three-dimensional digital mapping of the physical substation, and then conducts fusion analysis of equipment status data and inspection data based on this. However, this solution focuses on the visualization and defect warning during the substation operation and maintenance phase, and does not address the physical impact of cable transmission paths on the transmission delay of electrical control signals during the layout design phase. Therefore, it cannot provide numerical basis for the layout distance between the reactive power compensation device and the main transformer based on adjustment timing constraints.

[0005] The aforementioned layout methods and technical solutions do not consider the impact of cable connection paths between equipment on the transmission delay of reactive power regulation commands. When the physical distance between the reactive power compensation device and the main transformer is large, the inductive reactance of the long-distance cable generates additional phase lag during reactive current transmission, causing the inductive voltage drop of the cable to increase with the increase of path length, thus reducing the actual compensation effect of reactive power reaching the main transformer side. The RL time constant introduced by the cable inductive reactance, combined with the controller calculation delay and equipment response delay, forms the total delay time. Once the total delay time mismatches with the rapid fluctuation cycle of photovoltaic power output, the reactive power output of the reactive power compensation device cannot keep up with the bus voltage change demand in time, resulting in a phase mismatch between reactive power output and voltage demand. This leads to chasing oscillations in the bus voltage within the station, reducing the safety and stability of the power grid operation. Summary of the Invention

[0006] To address the technical problem that traditional substation layout methods fail to incorporate cable inductive reactance accumulation into spatial planning, leading to reactive power regulation delays and photovoltaic fluctuation cycle mismatches, and causing bus voltage oscillations, this invention provides solutions in the following aspects.

[0007] In a first aspect, the present invention provides a substation module collaborative layout method adapted to distributed photovoltaic systems, comprising:

[0008] Historical operation data of substations are acquired, power fluctuation sequences of distributed photovoltaic grid-connected points are extracted, a transfer sensitivity model is established and a corresponding sensitivity matrix is ​​constructed, matrix decomposition of the sensitivity matrix is ​​performed to identify the spatial distribution characteristics of the dominant photovoltaic fluctuation mode, and the dominant period of the power fluctuation sequence is extracted through frequency domain analysis. A model is established to measure the impact of cable inductive reactance accumulation on the transmission delay of reactive power regulation commands. The actual cable path length and total cable inductive reactance are calculated based on the distance between equipment and cable parameters. The total delay time is determined, and the ratio of the total delay time to the dominant cycle is used as the reactive power regulation mismatch factor. Based on the preset target mismatch margin coefficient, the cable cross-sectional area that meets the delay constraint is selected. Based on the inductive voltage drop constraint, the maximum allowable cable path length is calculated. A layout optimization model is constructed with minimizing the actual cable path length as the optimization objective. The model is solved under electrical safety distance constraints, maintenance passage width constraints, and cable path length constraints. The optimal layout coordinates of each module are output. Based on the optimal layout coordinates, place the three-dimensional geometric models of each device in the virtual space, plan the cable tray laying path, and generate a three-dimensional routing model and two-dimensional drawings.

[0009] This invention establishes a transmission sensitivity model and performs matrix decomposition on the sensitivity matrix to extract the spatial distribution characteristics of the dominant photovoltaic fluctuation mode. Based on this, the dominant period is extracted from the power fluctuation sequence of the photovoltaic access point with the largest coupling contribution weight. On this basis, the cable transmission delay is calculated according to the time constant of the RL circuit composed of cable inductance and resistance. The total delay time is obtained by combining the controller calculation delay and the equipment response delay. The ratio of the total delay time to the dominant period characterizes the degree of timing mismatch. Cable cross-sectional areas that meet the delay constraints are selected based on a preset target mismatch margin coefficient. The maximum allowable cable path length is calculated by back-calculating through inductive voltage drop constraints. Under engineering regulations such as electrical safety distance, the optimal layout coordinates of each module are solved, ensuring that the layout distance is controlled within the upper limit corresponding to the mismatch factor and the inductive voltage drop ratio, thus avoiding phase mismatch between reactive power output and voltage demand that could cause bus voltage oscillation.

[0010] Preferably, the step of establishing a transfer sensitivity model and constructing a corresponding sensitivity matrix, and performing matrix decomposition on the sensitivity matrix to identify the spatial distribution characteristics of the dominant photovoltaic fluctuation mode, includes: Construct the system Jacobian matrix, and use the system Jacobian matrix to obtain the partial derivatives of active power and reactive power with respect to the bus voltage amplitude, thereby obtaining the active power sensitivity coefficient and reactive power sensitivity coefficient. The active power sensitivity coefficient and reactive power sensitivity coefficient are assembled into a sensitivity matrix. Singular value decomposition is performed on the sensitivity matrix to obtain the left singular matrix, the diagonal singular value matrix and the right singular matrix. The minimum number of singular values ​​that meet the preset energy proportion is selected by adopting the cumulative contribution rate criterion, and the corresponding left singular vectors are extracted to form the spatial distribution feature vector of the dominant photovoltaic fluctuation mode.

[0011] By constructing the system Jacobian matrix and performing singular value decomposition on the sensitivity matrix, the multidimensional photovoltaic fluctuation information is compressed into at least a few dominant modes, so that subsequent layout optimization only needs to focus on the fluctuation component that contributes the most to voltage stability, thus reducing the input dimension of the optimization model.

[0012] Preferably, the extraction of the dominant period of the power fluctuation sequence through frequency domain analysis includes: Calculate the power change rate sequence of the power fluctuation sequence, and calculate the root mean square value and peak factor of the power change rate sequence; The frequency domain amplitude spectrum is obtained by performing a fast Fourier transform on the power fluctuation sequence; Search for the peak frequency of the frequency domain amplitude spectrum within the preset frequency band, record the peak frequency as the dominant frequency, and use the reciprocal of the dominant frequency as the dominant period.

[0013] Preferably, the step of establishing a model of the cumulative cable inductive reactance affecting the transmission delay of reactive power regulation commands, and calculating the actual cable path length and total cable inductive reactance, includes: The ratio of the total floor space occupied by all equipment within a substation to the total area of ​​the substation area is defined as the equipment floor space density. The path detour coefficient is calculated based on the equipment footprint density. The actual cable path length is obtained by multiplying the path detour coefficient by the straight-line distance between the reactive power compensation device and the main transformer. Select a cable with a corresponding cross-sectional area based on the rated capacity and current rating of the reactive power compensation device, determine the unit inductive reactance of the cable, and multiply the unit inductive reactance of the cable by the actual cable path length to obtain the total inductive reactance of the cable.

[0014] By introducing the equipment footprint density calculation path detour coefficient, the deviation between the actual cable laying path and the straight distance is incorporated into the numerical model, making the calculation results of cable inductive reactance accumulation closer to the actual wiring conditions in the station.

[0015] Preferably, determining the total delay time and using the ratio of the total delay time to the dominant cycle as the reactive power adjustment mismatch factor includes: calculating the cable transmission delay based on the total inductive reactance and total resistance of the cable; adding the controller calculation delay, the cable transmission delay, and the power module response delay of the reactive power compensation device to obtain the total delay time; and comparing the reactive power adjustment mismatch factor with a preset mismatch factor threshold.

[0016] The total delay time is obtained by calculating and superimposing the controller calculation delay, cable transmission delay, and reactive power compensation device power module response delay separately. The reactive power regulation mismatch factor is defined by the ratio of the total delay time to the dominant cycle, so that the matching degree between the regulation timing and the fluctuation cycle can be intuitively judged with dimensionless values.

[0017] Preferably, the step of selecting cable cross-sectional areas that satisfy the delay constraint based on a preset target mismatch margin coefficient, and calculating the maximum allowable cable path length based on the inductive voltage drop constraint, includes: The maximum allowable cable transmission delay is calculated based on the target mismatch margin factor, dominant cycle, controller calculation delay, and reactive power compensation device power module response delay. Cable cross-sectional areas that meet the delay constraints are then selected based on the maximum allowable cable transmission delay. The maximum allowable cable path length can be calculated by inversely using the rated reactive current of the reactive power compensation device, the unit inductive reactance of the cable, and the allowable inductive voltage drop ratio. Using the set of two-dimensional coordinates of each module in the substation as decision variables, a layout optimization model is constructed with the goal of minimizing the actual cable path length between the reactive power compensation device and the main transformer.

[0018] Preferably, the step of outputting the optimal layout coordinates of each module includes: An improved particle swarm optimization algorithm is used to solve the layout optimization model. The two-dimensional coordinates of all equipment in the substation are spliced ​​into a particle position vector. After the particle position is updated in each iteration, it is checked whether each particle meets the electrical safety distance constraint, maintenance passage width constraint and cable path length constraint. For particles that violate electrical safety distance constraints, a penalty function method is used to superimpose a penalty term into the objective function; for particles that violate maintenance passage width constraints, they are projected along the passage normal to the nearest feasible coordinate position that meets the passage width requirement. After the iteration converges, the particle position vector with the minimum objective function value is output, and the optimal layout coordinates of each module are obtained by decoding.

[0019] The particle swarm optimization algorithm performs a superposition penalty term for particles that violate electrical safety distance constraints and a normal projection correction for particles that violate maintenance passage width constraints. This enables the algorithm to gradually guide infeasible solutions into feasible domains that satisfy various engineering specifications during the iteration process, reducing invalid iterations caused by constraint conflicts.

[0020] Preferably, the three-dimensional geometric model for placing each device in virtual space based on optimal layout coordinates includes: By calling the building information modeling software interface, the standard three-dimensional geometric models of each piece of equipment in the substation are placed in the virtual space according to the optimal layout coordinates; Generate collision detection bounding boxes in virtual space that match the device geometry, and perform collision detection on all device pairs; If geometric interference exists, mark the conflicting device pair and return to adjust the optimal layout coordinates.

[0021] Preferably, the planning of the cable tray laying path, generating a three-dimensional routing model and two-dimensional drawings, includes: The ground inside the substation is divided into grids, the access status of grid nodes is marked, and a path search algorithm is used to search for the shortest collision-free path that satisfies the cable bending radius constraint on the grid map. A 3D model of the cable tray is generated in virtual space along the node sequence of the shortest collision-free path, and models of cable tray bends or tees are inserted at the turning points of the path. The planar layout is generated by projecting the three-dimensional geometric model from directly above, and the cross-sectional view is generated from the section plane where the cable tray is located. A performance report table containing the reactive power adjustment mismatch factor and the actual cable path length is output.

[0022] By searching for the shortest collision-free path that satisfies the cable bending radius constraint on the grid map and generating a 3D orientation model along the path, the layout optimization results are transformed into drawing files that can be directly used for construction. A performance report table containing reactive power adjustment mismatch factor and actual cable path length is also output, which facilitates engineers to check the compliance of the layout scheme item by item.

[0023] Secondly, the present invention provides a substation module collaborative layout system adapted to distributed photovoltaics, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned substation module collaborative layout method adapted to distributed photovoltaics is implemented.

[0024] The method for collaborative layout of substation modules adapted to distributed photovoltaics is stored in memory as a computer program, loaded and executed by the processor, so that each step of the above method can be run through computer equipment.

[0025] The beneficial effects of this invention are as follows: This invention introduces cable inductive reactance accumulation into the substation spatial layout optimization model. By establishing an RL circuit time constant model, the cable transmission delay is calculated and the reactive power regulation mismatch factor is derived. Based on the target mismatch margin coefficient, the cable cross-sectional area that meets the delay constraint is selected. On this basis, the maximum allowable cable path length is deduced by inversely using the inductive voltage drop constraint. The response timing constraint of the reactive power compensation device to photovoltaic power fluctuations and the reactive power compensation efficiency constraint are combined into the physical upper limit of the cable path length. Under the condition of meeting the engineering specifications such as electrical safety distance and maintenance passage width, the layout optimization model constrains the layout distance between the reactive power compensation device and the main transformer within the range corresponding to the mismatch factor threshold and the inductive voltage drop ratio threshold. This avoids bus voltage oscillation caused by the mismatch between reactive power regulation delay and photovoltaic fluctuation cycle or excessive inductive voltage drop leading to a decrease in compensation efficiency.

[0026] Furthermore, this invention extracts the dominant photovoltaic fluctuation modes by performing singular value decomposition on the sensitivity matrix, compressing multidimensional fluctuation information into at least a few dominant components, thus reducing the input scale of the optimization model. Based on the optimal layout coordinates, it calls the building information modeling software interface to place the three-dimensional geometric model and combines the path search algorithm to plan the cable tray laying path, transforming the optimization results into construction documents containing plan layout drawings, section drawings, and performance report tables. This enables the design team to verify the responsiveness and compliance of the layout scheme item by item at the drawing stage. Attached Figure Description

[0027] Figure 1 This is a flowchart of a substation module collaborative layout method adapted to distributed photovoltaic power generation; Figure 2 It is a curve showing the percentage of inductive voltage drop as a function of cable path length. Detailed Implementation

[0028] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0029] This invention discloses a substation module collaborative layout method adapted to distributed photovoltaic power generation, referring to... Figure 1 This includes steps S1-S4: S1. Modeling of photovoltaic fluctuation characteristics and sensitivity.

[0030] Historical operational data for the past three months was obtained from the substation's SCADA system, with a sampling interval of 1 second. Active power was extracted for each distributed photovoltaic grid-connected point within the substation during typical periods of dramatic solar radiation fluctuations. and reactive power Fluctuation sequence. For the first... For each photovoltaic (PV) grid connection point, calculate the power fluctuation: ; ; In the formula, For the first Each photovoltaic access point is located at Active power at any given time For the first Each photovoltaic access point is located at Reactive power at any given moment The sampling time step is defined as follows. The power change rate at each photovoltaic access point during the period of rapid change is statistically analyzed, and the root mean square value of the power change rate sequence is recorded. and peak factor The root mean square (RMS) value reflects the average amplitude of power fluctuations, while the gust factor reflects the peak intensity of the fluctuations. At the first sampling time, active power data from the previous sampling time is unavailable, and the power fluctuation is... Set it to 0.

[0031] A transfer sensitivity model is established to assess the impact of each photovoltaic access bay on the bus voltage of each section within the substation. A power flow calculation program is then used to construct the system Jacobian matrix under the current operating conditions. , for the Active power injected by each photovoltaic access point and reactive power The active and reactive power are obtained by using the system Jacobian matrix to determine the first... Busbar voltage amplitude The partial derivatives are used to obtain the active sensitivity coefficient. and reactive power sensitivity coefficient All Each photovoltaic access point The sensitivity coefficients of the busbar segments are assembled as follows: 3D sensitivity matrix and Matrix elements and The magnitude of the value directly reflects the first The power fluctuation at the photovoltaic access point affects the first The intensity of the impact of the bus voltage.

[0032] To identify the dominant photovoltaic fluctuation mode that has the greatest impact on the bus voltage, the sensitivity matrix is... Perform singular value decomposition to obtain In the formula, for Left-singular matrix for 3D diagonal singular value matrix, for A right-singular matrix. A diagonal singular value matrix. diagonal elements Sort in descending order for each singular value. This corresponds to the energy contribution of a photovoltaic fluctuation mode. The cumulative singular value energy percentage is calculated. : ; In the formula, For the first A singular value, This represents the total number of singular values. Using the cumulative contribution rate criterion, values ​​are selected that make... The minimum energy percentage exceeded the preset limit for the first time. Value, before extraction The left singular vector corresponding to the maximal singular value .forward The left singular vector corresponding to the maximal singular value The spatial distribution feature vector constituting the dominant photovoltaic wave mode, the vector's first... Each component Indicates the first The photovoltaic access point is at the first The coupling contribution weight in each dominant mode. The preset energy percentage range is [0.9, 0.98], and in this embodiment, it is 0.95. If the energy percentage is too low, secondary dominant modes that have a substantial impact on voltage stability will be missed; if it is too high, weak modes dominated by noise will be included, increasing the complexity of subsequent calculations. In the dimensionality reduction analysis of similar power systems, 0.95 is a commonly used engineering threshold.

[0033] The calculation is performed using a 110kV substation connected to 5 photovoltaic points as an example. After singular value decomposition, the first two singular values ​​are... , All other singular values ​​are less than 0.5. Cumulative energy percentage Since the value exceeds 0.95, the first two left singular vectors are extracted. Each vector contains five components, and the photovoltaic access points with larger component values ​​have a stronger impact on the voltage stability within the station.

[0034] Time-domain features are extracted from the photovoltaic power fluctuation sequence. The power change rate sequence is calculated. Root mean square value: ; In the formula, The number of sampling points. For the first Each photovoltaic access point is located at The rate of change of power at time t. Calculate the peak factor. A larger peak factor indicates more severe and irregular fluctuations. The dominant photovoltaic (PV) access point is determined by the component with the largest coupling contribution weight in the spatial distribution characteristic vector. The power fluctuation sequence of the dominant PV access point is then analyzed. Perform a fast Fourier transform to obtain the frequency domain amplitude spectrum. Search for the peak frequency of the frequency domain amplitude spectrum within the 0.1Hz to 10Hz frequency band, and denote the peak frequency as the dominant frequency. The reciprocal of the dominant frequency is the dominant period. Physically, this reflects the typical repetitive cycle of photovoltaic power output fluctuations.

[0035] Assuming that under conditions of drastic solar radiation variation, the power fluctuation sequence at a photovoltaic grid connection point, after Fast Fourier Transform (FFT), exhibits a spectral peak with the largest amplitude at 2.5 Hz, and the dominant frequency is... Hz, dominant cycle s, meaning the photovoltaic output completes one fluctuation cycle approximately every 0.4 seconds. Dominant cycle. This will serve as the benchmark parameter for the response time constraint of subsequent reactive power compensation devices.

[0036] S2, Modeling of cable inductive reactance accumulation and transmission delay.

[0037] A numerical model is established to illustrate the impact of cable inductive reactance accumulation on the transmission delay of reactive power regulation commands, considering the cable connection path between the reactive power compensation device and the main transformer within the substation. The reactive power compensation device is assumed to be installed at a two-dimensional plane coordinate position within the substation. The main transformer is located at the coordinate position. The straight-line distance between the two is: ; In the formula, , The two-dimensional plane coordinates of the reactive power compensation device. , This refers to the two-dimensional plane coordinates of the main transformer. The actual cable laying path must bypass obstacles such as switchgear, supports, and fences within the station along the cable trays. The total floor area occupied by all equipment within the station is calculated. Total area of ​​the station area The ratio is defined as the equipment floor space density. Since the equipment footprint density increases approximately linearly with the cable path detour, the path detour coefficient is used to characterize the extension factor of the actual cable laying path relative to the straight-line distance. ; In the formula, The path detour coefficient. This refers to the equipment footprint density. The constant 0.5 in the coefficient comes from the statistical regularity of cable laying in typical substations: when the equipment footprint density... As the value increases from 0 to 0.4, the cable path lengthens by an average of approximately 20%, and the slope of the linear fit is 0.5. The value range is [0, 0.6], and 0.3 is used in this embodiment. In similar substation projects, the equipment footprint density is usually between 0.2 and 0.5, and 0.3 is a typical value for medium density. Actual cable path length ,when When the coordinates are less than half the sum of the maximum dimensions of the two devices, it indicates that the two devices have partially overlapped physically. This coordinate combination does not meet the electrical safety distance constraint and is excluded from the subsequent layout optimization solution by the constraint conditions.

[0038] Verification was conducted using actual operational data from the station: the straight-line distance between the reactive power compensation device and the main transformer is 15m, and the equipment density within the station is [data missing]. The path detour coefficient is 0.3. Actual cable path length m.

[0039] Inductive reactance per unit length of cable The inductive reactance depends on the cable cross-sectional area, conductor material, and laying method. For a typical 10kV substation copper-core cross-linked polyethylene cable, the unit inductive reactance is 0.08 Ω / km for a cross-sectional area of ​​240 mm², 0.1 Ω / km for a cross-sectional area of ​​150 mm², and 0.12 Ω / km for a cross-sectional area of ​​95 mm². The appropriate cable cross-sectional area should be selected based on the rated capacity and current rating of the reactive power compensation device in the substation to determine the unit inductive reactance. Total inductance of the cable This inductive reactance generates an additional inductive voltage drop and phase lag during reactive current transmission. The larger the impedance, the greater the inductive impedance introduced per unit length of cable, and the more severe the phase shift during reactive current transmission.

[0040] The total delay time from the issuance of the reactive power regulation command by the station controller to the actual generation of reactive power at the output of the reactive power compensation device. It consists of three parts. The controller calculates the delay. This includes the time for acquiring the bus voltage signal, executing the PI control algorithm, and generating the PWM modulation command. Its value ranges from [30, 50] ms; in this embodiment, it is set to 40 ms. The actual time should be determined by those skilled in the art based on the actual controller hardware performance. Cable transmission delay. The time constant of the RL circuit, composed of cable inductance and resistance, determines the response delay of the power module in the reactive power compensation device. This includes the IGBT switching action and the output filter settling time, with a value range of [10, 20] ms. In this embodiment, 15 ms is used. Total delay time .

[0041] According to the RL circuit theory of power systems, the inductance corresponding to the inductive reactance of a cable is... In the formula This is the power frequency. Total cable resistance. In the formula The resistance per unit length of the cable is determined by the cable's cross-sectional area and conductor material. For copper core cables, the resistance per unit length is... It is inversely proportional to the cross-sectional area; when the cross-sectional area is 150 mm² When the resistance is 0.117 Ω / km and the cross-sectional area is 95 mm² The time constant of the RL circuit is 0.184 Ω / km. This time constant characterizes the time required for the current to rise from zero to 63.2% of its steady-state value. Considering that the reactive power regulation command requires the current to reach more than 95% of the target value, the cable transmission delay is taken. : ; In the formula, Inductive reactance per unit length of cable The resistance per unit length of the cable. This is the power frequency. Because... , Cable path length Canceling the cable transmission delay in the numerator and denominator. It depends only on the ratio of unit inductive reactance to unit resistance corresponding to the cable cross-sectional area, and is independent of the cable path length. The larger or The smaller the value, the longer the RL time constant, and the greater the cable transmission delay.

[0042] Taking a 150mm² copper core cable as an example, It is 0.1Ω / km. It is 0.117Ω / km. ms. Taking a 95mm² copper core cable as an example, It is 0.12Ω / km. It is 0.184Ω / km. For cables with smaller cross-sectional areas, although the inductive reactance increases from 0.1Ω / km to 0.12Ω / km, the resistance increases from 0.117Ω / km to 0.184Ω / km. The increase in resistance exceeds the increase in inductive reactance, resulting in a shorter RL time constant. Therefore, both current carrying capacity and RL delay must be considered when selecting cable cross-sectional areas.

[0043] Although the cable path length does not affect the RL time constant, it does affect the total inductive reactance of the cable. Inductive voltage drop generated during reactive current transmission In the formula This represents the reactive current amplitude. The inductive voltage drop increases linearly with the path length, causing the actual compensation effect of the reactive power reaching the main transformer to decrease as the path length increases. This inductive voltage drop margin will be used as a constraint to back-calculate the maximum allowable cable path length in the subsequent layout optimization solution.

[0044] reactive power regulation mismatch factor Defined as total delay time With photovoltaic fluctuations dominating the cycle The ratio of reactive power regulation mismatch factor to the fluctuation period is used to characterize the degree of mismatch between regulation delay and fluctuation period. In the formula The dominant cycle is photovoltaic fluctuation. According to the Nyquist sampling theorem in control theory, to track a frequency of... For fluctuating signals, the response time of the control system must be less than 1 / 2 of the signal period, that is... ,correspond The reactive power compensation device can still achieve partial reactive power tracking when the total delay time slightly exceeds half a fluctuation period. Based on this, the Nyquist theoretical boundary is relaxed using an engineering margin factor of 1.2, and the mismatch factor threshold is set accordingly. The engineering margin coefficient ranges from [1.1, 1.3]. If the coefficient is too small, the judgment will be too strict and the actually feasible layout scheme will be excluded. Conversely, if the value is too large, the judgment boundary will be too close to the adjustment failure point. In this embodiment, 1.2 is used. When the voltage approaches or exceeds 0.6, the reactive power compensation device has not yet completed its adjustment action while the photovoltaic output has already entered the next round of reverse fluctuation. The reactive power and the bus voltage demand are mismatched, and the bus voltage in the station falls into a chasing oscillation.

[0045] Substitute the aforementioned values ​​into the calculation. Photovoltaic fluctuations dominate the cycle. The controller delay is 0.4s. The response delay of the reactive power compensation device is 40ms. The ms time is 15ms. This applies when using 150mm² copper core cable. The total latency is 8.2ms. ms, mismatch factor The value is less than the threshold of 0.6. When using 95mm² copper core cable... The total latency is 6.2ms. ms, mismatch factor Under the parameters described in this example, cable transmission delay accounts for approximately 10% to 13% of the total delay, and the mismatch factor is primarily dominated by controller delay and equipment response delay. This is especially relevant when photovoltaic fluctuations dominate the cycle. When the time is shortened to 0.1s, the mismatch factor using a 150mm² cable If the threshold of 0.6 is exceeded, a cable cross-sectional area with a shorter RL time constant or a shortened controller delay should be selected to meet the mismatch factor constraint.

[0046] S3, Module coordinate optimization solution.

[0047] To meet the reactive power regulation delay constraint, the target mismatch margin coefficient is first determined. Select qualified cable cross-sectional areas. (By...) Determine the maximum allowable cable transmission delay. In the formula This represents the target mismatch margin coefficient. Because... Depends solely on the cable cross-sectional area, assigning each candidate cross-sectional area to... and Compare and select The cross-sectional area of ​​the cable. The value range is [0.3, 0.5], and 0.5 is used in this embodiment. When the target mismatch margin coefficient is too small, the allowable cable transmission delay margin is narrowed, and the selection of cable cross-sectional area is more restricted; when it is too large, the margin is sufficient but close to the adjustment failure boundary. In engineering practice, it is usually taken as 0.5 to balance layout flexibility and adjustment reliability.

[0048] Substitute the above parameters into the calculation: It is 0.5. It takes 0.4 seconds. It takes 40ms. It takes 15ms. ms. 150mm² copper core cable For 8.2ms, 95mm² copper core cable The response time is 6.2 ms, and both cross-sectional areas meet the requirements. Delay constraints.

[0049] Based on the cable cross-sectional area that satisfies the delay constraint, the maximum allowable cable path length is further calculated by inversely applying the inductive voltage drop constraint. Reactive current Through the total inductance of the cable Inductive voltage drop The inductive voltage drop must not exceed the rated voltage of the bus. Permissible ratio ,Right now Solving for: ; In the formula, For the allowable inductive voltage drop ratio, The rated voltage of the busbar This refers to the rated reactive current of the reactive power compensation device. The inductive reactance per unit length of the selected cable. The value range is [0.02, 0.05], and 0.03 is used in this embodiment. Although a higher inductive voltage drop ratio relaxes the layout space, it will lead to a decrease in reactive power compensation efficiency, while too small a ratio will impose excessive constraints on the layout. In reactive power compensation projects of distribution networks, 0.03 is usually used to balance compensation efficiency and layout flexibility.

[0050] With the station's reactive power compensation device rated at 6 Mvar and the bus rated voltage... Taking 10kV as an example, reactive current A. Select 150mm² copper core cable. It is 0.1Ω / km, which is converted to Ω / m, m, approximately 8.67km. In conventional substation layouts, the cable path length between the reactive power compensation device and the main transformer generally does not exceed 200m. Inductive voltage drop constraints impose limitations on the layout coordinates when the substation area is large, the reactive current level is high, or small-section cables are used. When the rated capacity of the reactive power compensation device increases to 20Mvar, Increased to 1155A, Narrowed to 2598m; if at the same time Tightened to 0.02, It was further shortened to 1732m.

[0051] Construct a layout optimization model, with the decision variables being the set of two-dimensional coordinates of each module within the station. , In the formula This represents the total number of equipment within the station, including the main transformer, reactive power compensation devices, switchgear, capacitor banks, and station service transformers. The optimization objective is to minimize the actual cable path length between the reactive power compensation devices and the main transformer. : ; In the formula, , For the coordinate decision variables of the reactive power compensation device, , The coordinate decision variables of the main transformer. This is the path detour coefficient determined by the current equipment footprint density. Constraints include three types: electrical safety distance constraints require that the Euclidean distance between the center coordinates of any two pieces of equipment within the station is not less than the minimum distance specified in the regulations. ,Right now , The width of the maintenance passageway is determined by the equipment type combination and voltage level, referring to the regulations; the net width of the main passageway formed by the equipment arrangement should not be less than 3.5m, and the net width of the side passageway should not be less than 1.2m; the cable path length is also subject to constraints. .

[0052] An improved particle swarm optimization algorithm was used to solve the above optimization model. The two-dimensional coordinates of all equipment in the station were stitched together. dimensional particle position vector Particle velocity vector This indicates the adjustment direction and magnitude of each coordinate component. The particle swarm size is 50 to 100 particles; in this embodiment, 80 particles are used. The initial swarm generation method is as follows: using the boundary coordinates of the available area of ​​the station as constraints, and under the premise of meeting electrical safety distances, Latin hypercube sampling is performed on the coordinates of each device to generate 80 sets of initial particle position vectors. The maximum number of iterations is 200. The inertia weight decreases linearly from 0.9 to 0.4, and the individual learning factor... Group learning factor .

[0053] After each iteration of particle position update, each particle is checked to see if it satisfies the three types of constraints. For particles that violate electrical safety distance constraints, a penalty function method is used to add a penalty term to the objective function: if the equipment... Distance between The amount of penalty added Penalty coefficient Constructed as an adaptive dynamic coefficient, based on the maximum diagonal length of the station area space. The shortest safe distance from the current procedure It is calculated dynamically. This allows the magnitude of the penalty term to adaptively scale with the physical scale of the substation and match the dimensions of the objective function of the cable length. After the objective function value of the defaulting particle deteriorates, it is eliminated in the iteration.

[0054] For particles that violate the maintenance channel width constraint, project them along the channel normal to the nearest feasible coordinate position that meets the channel width requirement. After projection correction, it is necessary to check whether the new coordinate position triggers an electrical safety distance conflict with other equipment: if a conflict occurs, abandon the projection correction, return the particle to the previous valid iteration position, and trigger local re-random initialization to avoid falling into a spatial interference dead loop.

[0055] After iteration, the particle swarm converges to near the global optimum, outputting the position vector of the particle with the minimum objective function value. Decoding yields the optimal layout coordinates of each module. To verify the convergence of the algorithm, record the globally optimal objective function value for each iteration. Within 20 consecutive iterations The algorithm is considered converged when the relative change is less than 0.1%. If the convergence criterion is not triggered within 200 iterations, the particle swarm size is increased to 150 and the algorithm is run again. The optimal coordinate set is then input into the output. The formula is used to verify the actual cable path length, confirming that it does not exceed [the specified value]. And calculate the corresponding mismatch factor. Confirmed to be no more than 0.6.

[0056] S4, 3D visualization and construction drawing generation.

[0057] Obtain the optimal layout coordinates of each module Then, the Building Information Modeling (BIM) software interface is invoked to place the standard 3D geometric models of each piece of equipment within the station into the virtual space according to the optimal layout coordinates. The 3D model of the main transformer includes geometric components such as the main tank, radiator assembly, high and low voltage bushings, and grounding grid connection points. The model origin is aligned with the optimal layout coordinates. Alignment. The 3D model of the reactive power compensation device includes components such as power cabinet, AC busbar connection ports, and cooling ducts. The center of the model's mounting surface is aligned with the optimal layout coordinates. Alignment. The remaining devices are placed similarly according to their optimal layout coordinates. After placement, a collision detection bounding box matching the device geometry is generated in virtual space. Collision detection is performed on all device pairs; if geometric interference exists, the conflicting device pairs are marked, and the module coordinate optimization solution is returned to readjust the optimal layout coordinates.

[0058] The cable tray laying path is planned in virtual space. The ground within the station is divided into a 0.5m × 0.5m grid, and each grid node is marked with its accessibility status: equipment-occupied areas and basic fence areas are marked as obstacles, while other areas are marked as passable. A path search algorithm, namely A... The path search algorithm starts at the grid node where the AC busbar connection port of the reactive power compensation device is located and ends at the grid node where the low-voltage bushing terminal of the main transformer is located. It searches the grid map for the shortest collision-free path that satisfies the cable bending radius constraint. The algorithm's heuristic function is the Manhattan distance from the current node to the destination, and the node movement cost is the sum of the actual movement distance and the bending penalty. When the path turns at the current node, the bending penalty cost is added. In the formula The cable outer diameter is used. The bending penalty cost dynamically calculates the equivalent distance based on the selected cable outer diameter and the minimum permissible bending radius required by the specification, to suppress excessive right-angle bends in the path that result in insufficient bending radius. After the search is complete, the shortest collision-free cable path node sequence from the reactive power compensation device to the main transformer is output, and the total path length is extracted. Compared with the aforementioned actual cable path length Perform comparison and verification.

[0059] A 3D model of the cable tray is generated in virtual space along the path node sequence. The width of the cable tray is determined according to the total cross-sectional area of ​​the laid cables and the fill rate requirements: according to DL / T5221, the total cross-sectional area of ​​the cables in the cable tray shall not exceed 40% of the cross-sectional area of ​​the cable tray. The required cross-sectional area of ​​the cable tray is calculated based on the selected cable cross-sectional area and the number of cables laid, and a standard cable tray specification is selected. Cable tray bends or tees are inserted at the turning points of the path. Cable tray supports are placed every 1.5m to 2m along the path. The height of the supports is determined according to the clearance requirements from the ground to the bottom of the cable tray, and is usually not less than 2.2m.

[0060] Generate a plan view from the top projection of the 3D geometric model, labeling the center coordinates of each piece of equipment, equipment number, foundation outline dimensions, and foundation center-to-center spacing. Generate a sectional view from the section plane containing the cable tray, labeling the cable tray installation height, support spacing, and cable tray specifications. Along A The path is drawn with thick solid lines to represent the cable tray route, with bending radii and angles marked at key turning points. A performance report of the layout scheme is output, summarizing reactive power regulation mismatch factors. Actual cable path length and its relationship with threshold The percentage margin, the total area of ​​the station, the minimum clearance between each equipment pair and its comparison with the requirements of the electrical safety regulations, and the measured value of the net width of the main maintenance passage.

[0061] Reference Figure 2 The three straight lines in the figure represent the linear increase in the proportion of inductive voltage drop with increasing cable path length under three combinations of reactive power compensation device capacity and cable cross-sectional area. The horizontal line represents the threshold value for the inductive voltage drop ratio. Under the condition of 6 Mvar / 150 mm² copper core cable, the proportion of inductive voltage drop increases slowly with increasing path length, intersecting the threshold line at approximately 8671 m. Under the condition of 12 Mvar / 150 mm² copper core cable, due to the increase in reactive current, the rate of increase in the proportion of inductive voltage drop accelerates, intersecting the threshold line at approximately 4329 m. Under the condition of 20 Mvar / 95 mm² copper core cable, the reactive current further increases and the cable unit inductive reactance is higher, resulting in the steepest increase in the proportion of inductive voltage drop, intersecting the threshold line at approximately 2165 m. The horizontal axis corresponding to each intersection point represents the maximum allowable cable path length under that condition. The larger the rated capacity of the reactive power compensation device or the higher the cable unit inductive reactance, the shorter the maximum allowable cable path length, and the tighter the cable path length constraint in the layout optimization model. This feature verifies the technical effectiveness of the present invention in deriving the maximum allowable cable path length through inductive voltage drop constraints and transforming reactive power compensation efficiency constraints into upper limits of layout space.

[0062] This invention also discloses a substation module collaborative layout system adapted to distributed photovoltaics, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a substation module collaborative layout method adapted to distributed photovoltaics according to the present invention is implemented.

[0063] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0064] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise expressly and specifically defined.

Claims

1. A substation module collaborative layout method adapted to distributed photovoltaic power generation, characterized in that, include: S1. Obtain historical operating data of the substation, extract the power fluctuation sequence of the distributed photovoltaic grid connection point, establish a transmission sensitivity model and construct the corresponding sensitivity matrix, perform matrix decomposition on the sensitivity matrix to identify the spatial distribution characteristics of the dominant photovoltaic fluctuation mode, and extract the dominant period of the power fluctuation sequence through frequency domain analysis. S2. Establish a model of the cumulative cable inductive reactance affecting the transmission delay of reactive power regulation commands. Calculate the actual cable path length and total cable inductive reactance based on the distance between equipment and cable parameters, determine the total delay time, and use the ratio of the total delay time to the dominant cycle as the reactive power regulation mismatch factor. S3. Select cable cross-sectional areas that meet the delay constraints based on the preset target mismatch margin coefficient, back-calculate the maximum allowable cable path length based on the inductive voltage drop constraint, construct a layout optimization model with minimizing the actual cable path length as the optimization objective, solve it under electrical safety distance constraints, maintenance passage width constraints and cable path length constraints, and output the optimal layout coordinates of each module. S4. Based on the optimal layout coordinates, place the three-dimensional geometric models of each device in the virtual space, plan the cable tray laying path, and generate a three-dimensional routing model and two-dimensional drawings.

2. The substation module collaborative layout method adapted to distributed photovoltaic power generation according to claim 1, characterized in that, The process of establishing a transfer sensitivity model and constructing a corresponding sensitivity matrix, and performing matrix decomposition on the sensitivity matrix to identify the spatial distribution characteristics of the dominant photovoltaic fluctuation mode, includes: Construct the system Jacobian matrix, and use the system Jacobian matrix to obtain the partial derivatives of active power and reactive power with respect to the bus voltage amplitude, thereby obtaining the active power sensitivity coefficient and reactive power sensitivity coefficient. The active power sensitivity coefficient and reactive power sensitivity coefficient are assembled into a sensitivity matrix. Singular value decomposition is performed on the sensitivity matrix to obtain the left singular matrix, the diagonal singular value matrix and the right singular matrix. The minimum number of singular values ​​that meet the preset energy proportion is selected by adopting the cumulative contribution rate criterion, and the corresponding left singular vectors are extracted to form the spatial distribution feature vector of the dominant photovoltaic fluctuation mode.

3. The substation module collaborative layout method adapted to distributed photovoltaic power generation according to claim 1, characterized in that, The extraction of the dominant period of the power fluctuation sequence through frequency domain analysis includes: Calculate the power change rate sequence of the power fluctuation sequence, and calculate the root mean square value and peak factor of the power change rate sequence; The frequency domain amplitude spectrum is obtained by performing a fast Fourier transform on the power fluctuation sequence; Search for the peak frequency of the frequency domain amplitude spectrum within the preset frequency band, record the peak frequency as the dominant frequency, and use the reciprocal of the dominant frequency as the dominant period.

4. The substation module collaborative layout method adapted to distributed photovoltaic power generation according to claim 1, characterized in that, The establishment of a model for the cumulative cable inductive reactance affecting the transmission delay of reactive power regulation commands, and the calculation of the actual cable path length and total cable inductive reactance, include: The ratio of the total floor space occupied by all equipment within a substation to the total area of ​​the substation area is defined as the equipment floor space density. The path detour coefficient is calculated based on the equipment footprint density. The actual cable path length is obtained by multiplying the path detour coefficient by the straight-line distance between the reactive power compensation device and the main transformer. Select a cable with a corresponding cross-sectional area based on the rated capacity and current rating of the reactive power compensation device, determine the unit inductive reactance of the cable, and multiply the unit inductive reactance of the cable by the actual cable path length to obtain the total inductive reactance of the cable.

5. A substation module collaborative layout method adapted to distributed photovoltaic power generation according to claim 1, characterized in that, The determination of the total delay time and the use of the ratio of the total delay time to the dominant cycle as the reactive power regulation mismatch factor includes: The cable transmission delay is calculated based on the total inductive reactance and total resistance of the cable; the controller's calculated delay, the cable transmission delay, and the reactive power compensation device's power module response delay are added together to obtain the total delay time; the reactive power adjustment mismatch factor is compared with the preset mismatch factor threshold.

6. The substation module collaborative layout method adapted to distributed photovoltaic power generation according to claim 5, characterized in that, The step of selecting cable cross-sectional areas that meet the delay constraint based on a preset target mismatch margin coefficient, and then calculating the maximum allowable cable path length based on the inductive voltage drop constraint, includes: The maximum allowable cable transmission delay is calculated based on the target mismatch margin factor, dominant cycle, controller calculation delay, and reactive power compensation device power module response delay. Cable cross-sectional areas that meet the delay constraints are then selected based on the maximum allowable cable transmission delay. The maximum allowable cable path length can be calculated by inversely using the rated reactive current of the reactive power compensation device, the unit inductive reactance of the cable, and the allowable inductive voltage drop ratio. Using the set of two-dimensional coordinates of each module in the substation as decision variables, a layout optimization model is constructed with the goal of minimizing the actual cable path length between the reactive power compensation device and the main transformer.

7. A substation module collaborative layout method adapted to distributed photovoltaic power generation according to claim 1, characterized in that, The optimal layout coordinates of each output module include: An improved particle swarm optimization algorithm is used to solve the layout optimization model. The two-dimensional coordinates of all equipment in the substation are spliced ​​into a particle position vector. After the particle position is updated in each iteration, it is checked whether each particle meets the electrical safety distance constraint, maintenance passage width constraint and cable path length constraint. For particles that violate electrical safety distance constraints, a penalty function method is used to superimpose a penalty term into the objective function; for particles that violate maintenance passage width constraints, they are projected along the passage normal to the nearest feasible coordinate position that meets the passage width requirement. After the iteration converges, the particle position vector with the minimum objective function value is output, and the optimal layout coordinates of each module are obtained by decoding.

8. A substation module collaborative layout method adapted to distributed photovoltaic power generation according to claim 1, characterized in that, The three-dimensional geometric model for placing each device in virtual space based on optimal layout coordinates includes: By calling the building information modeling software interface, the standard three-dimensional geometric models of each piece of equipment in the substation are placed in the virtual space according to the optimal layout coordinates; Generate collision detection bounding boxes in virtual space that match the device geometry, and perform collision detection on all device pairs; If geometric interference exists, mark the conflicting device pair and return to adjust the optimal layout coordinates.

9. A substation module collaborative layout method adapted to distributed photovoltaic power generation according to claim 1, characterized in that, The planned cable tray laying path generates a three-dimensional routing model and two-dimensional drawings, including: The ground inside the substation is divided into grids, the access status of grid nodes is marked, and a path search algorithm is used to search for the shortest collision-free path that satisfies the cable bending radius constraint on the grid map. A 3D model of the cable tray is generated in virtual space along the node sequence of the shortest collision-free path, and models of cable tray bends or tees are inserted at the turning points of the path. The planar layout is generated by projecting the three-dimensional geometric model from directly above, and the cross-sectional view is generated from the section plane where the cable tray is located. A performance report table containing the reactive power adjustment mismatch factor and the actual cable path length is output.

10. A substation module collaborative layout system adapted to distributed photovoltaic power generation, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement the substation module collaborative layout method adapted to distributed photovoltaics according to any one of claims 1 to 9.

Images