Distribution network equipment modeling simulation method based on digital twin application architecture

Abstract

The invention relates to a distribution network equipment modeling simulation method based on a digital twin application architecture. The method comprises the following steps: S1, constructing a distribution network digital twin space description application architecture comprising an interactive communication layer, a model building layer and a simulation analysis layer; s2, on the basis of the constructed distribution network digital twin space description application architecture, carrying out differential modeling on basic elements of the distribution network equipment; s3, based on the constructed distribution network digital twin space description application architecture, modeling is carried out on core distribution network equipment facilities; and S4, based on an electromagnetic transient algorithm, completing information interaction and simulation operation simulation between the distribution network equipment. The method has the beneficial effects of better calculation efficiency and better real-time performance.

Description

[Technical field]

[0001] The present invention relates to the technical field of power system computer simulation, and in particular to a distribution network equipment modeling and simulation method based on a digital twin application architecture. [Background technology]

[0002] As global air pollution and climate warming continue to intensify, reducing carbon emissions has become a consensus of the international community, and new energy vehicles have therefore received strong support and development. According to the New Energy Vehicle Industry Development Plan (2021-2035), the number of electric vehicles is entering a stage of rapid growth. It is expected that by 2025, the sales of new energy vehicles will account for 20% of the total vehicle sales; by 2035, pure electric vehicles will become the mainstream of sales. As an emerging form of load, electric vehicles have a great deal of randomness in load charging. The changing load of electric vehicles will change the power supply demand in a certain period of time, bringing certain negative impacts to the regional distribution network.

[0003] The integrated photovoltaic, energy storage and charging power station integrates photovoltaic, energy storage and electric vehicle charging facilities, and has the advantage of efficient use of localized clean energy. It contains multiple resources such as photovoltaic, energy storage, electric vehicles, etc. It is an important means of efficient use of localized clean energy, and it is also a new object to meet the development needs of future low-carbon distribution networks. In addition, the integration of photovoltaic, energy storage and charging can overcome the shortcomings of single resources such as photovoltaic, energy storage, and electric vehicles. It can not only smooth the uncertainty of photovoltaic output through energy storage and electric vehicles, but also combine electric vehicles and energy storage to improve the power supply capacity of the system, and has greater fault recovery potential. Therefore, from the three aspects of source, load and storage, the integrated photovoltaic, energy storage and charging power station is the best solution to this problem.

[0004] However, due to the access of renewable energy and a high proportion of power electronic equipment, the composition of distribution network equipment is becoming increasingly complex. The nonlinear characteristics of electrical equipment and the dynamic characteristics of different time scales are intertwined, which increases the difficulty of analysis and puts higher requirements on transient process modeling and real-time simulation of distribution networks. Traditional simulation methods only model the dynamic process of a local power grid at a certain time and space scale. They cannot describe multiple power grid forms in the system at the same time, and there are problems such as large computational complexity and insufficient simulation efficiency, which makes it difficult to meet the needs of accurate simulation of new power systems.

[0005] Digital twin (DT, also known as digital mirror, digital mapping) is based on the Internet of Things, data middle platform, cloud computing, digital thread, big data analysis, artificial intelligence and other technologies. It is a modeling and simulation technology that integrates multi-scale, multi-physical quantity, and multi-probability. It makes full use of physical models, sensor updates, operation history and other data, integrates multi-disciplinary, multi-physical quantity, multi-scale, and multi-probability simulation processes, and completes mapping in virtual space, thereby reflecting the full life cycle of the corresponding physical equipment. The power grid industry has the characteristics of high value and simple technical requirements, which is very suitable as a leading scenario for the application of digital twins. Although some power companies have introduced digital twin technology, it is mainly used in the digital twin of power plants and distribution plants, which is simply visualization. There are various problems in the application of other scenarios, and it has not formed a global and full life cycle application. Therefore, the digital twin technology of photovoltaic storage charging stations has obvious development potential and application value.

[0006] The present invention aims to solve the technical problems that traditional simulation methods have large amount of calculation, low efficiency, and difficulty in fully reflecting the dynamic process of power electronic equipment; the topological structure of photovoltaic storage charging stations is complex and the control methods are diverse; the solver model is time-consuming to build and real-time is difficult, and thus technical improvements are made to the distribution network equipment modeling and simulation method. [Summary of the invention]

[0007] The purpose of the present invention is to provide a distribution network equipment modeling and simulation method with better computing efficiency and better real-time performance.

[0008] To achieve the above-mentioned purpose, the technical solution adopted by the present invention is a distribution network equipment modeling and simulation method based on a digital twin application architecture, wherein the distribution network equipment is a photovoltaic storage charging station, the basic components of which include inductive components, resistive components and capacitive components, and the core distribution network equipment facilities of which include photovoltaic cells, energy storage systems, grid-connected converters and charging loads; the method comprises the following steps:

[0009] S1. Construct a distribution network digital twin spatial description application architecture including an interactive communication layer, a model building layer, and a simulation analysis layer;

[0010] S2. Based on the constructed distribution network digital twin space description application architecture, the basic components of the distribution network equipment are modeled by differentiation;

[0011] S3. Model the core distribution network equipment and facilities based on the constructed distribution network digital twin space description application architecture;

[0012] S4. Based on the electromagnetic transient algorithm, complete the information interaction and simulation operation between distribution network equipment.

[0013] Preferably,

[0014] The interactive communication layer of the distribution network digital twin space description application architecture is used to establish a data interaction mechanism in the physical entity and model building layer, input the topological information, equipment parameters, monitoring quantity, and real-time operation data of the distribution network equipment and facilities, and realize the storage and preprocessing of data;

[0015] The distribution network digital twin space describes the application architecture model building layer, with primary equipment, secondary equipment and auxiliary sensing equipment as the basis of the digital twin model;

[0016] The distribution network digital twin space description application architecture simulation analysis layer uses an object-oriented modeling method and an electromagnetic transient simulation method to read and monitor the environmental quantities, electrical quantities, physical quantities and behavioral quantities of distribution network equipment and facilities, and collects the system key node voltage, energy storage and charging equipment charge status, and active power data.

[0017] Preferably, step S2 specifically includes the following sub-steps:

[0018] Number the circuit network, and write the node information and parameters of each device in the circuit topology into the read record file to complete the pre-storage of data;

[0019] Based on the pre-stored data, the backward Euler method is selected to perform differential modeling of the basic components.

[0020] Preferably,

[0021] Inductor Component Differential Model L (t) = Y L u L (t)+I h_L (t-Δt);

[0022] Among them, Y L is the equivalent admittance of the inductor, I h_L (t-Δt) is the historical current source, u L (t) is the inductor terminal voltage, i L (t) is the inductor current at the current moment.

[0023] Preferably, step S3 specifically includes the following sub-steps:

[0024] Photovoltaic cells are distributed power sources, equivalent to a constant current source model;

[0025] The energy storage system includes an energy storage battery, which is equivalent to a controlled voltage source model with internal resistance;

[0026] The grid-connected converter uses feed-forward decoupling to achieve independent control of dq components;

[0027] The charging load is the electric vehicle battery, which is equivalent to a controlled voltage source model with internal resistance;

[0028] Output voltage of the controlled voltage source model with internal resistance of the energy storage system Among them, V 0 represents the internal electromotive force, SOC represents the state of charge of the battery, β is the characteristic constant of the battery, I is the discharge current of the energy storage battery, and r represents the internal resistance of the battery;

[0029] Mathematical Model of Grid-connected Converter Among them, K ip is the proportional coefficient of the PI regulator, K il is the integral coefficient of the PI regulator, s is a complex variable, i q is the q-axis component of the AC side current, for i q Current reference value, L is the inductance value on the AC side, R is the line impedance on the AC side, i d is the d-axis component of the AC side current, for i d Current reference value;

[0030] The three-phase voltage source grid-connected converter is updated by the average value modeling method, and the AC side node voltage equation and the DC side node voltage equation are obtained as follows:

[0031]

[0032]

[0033] Among them, I a , I b , I c is the current flowing into nodes a, b, c on the AC side, u a ,u b ,u c is the three-phase voltage of a, b, c on the AC side, u i ,u j is the DC side i, j voltage, D is the duty cycle, Y on is the conductivity, I i , I j is the current flowing into the DC side node i, j.

[0034] Preferably, step S4 specifically includes the following sub-steps:

[0035] When the core distribution network equipment and facilities perform cyclic calculations, they first call the control function, calculate the duty cycle of the current simulation step based on the control quantity at the previous moment, and update the admittance matrix based on the average value model;

[0036] In each cycle calculation process, the historical current of the current simulation step is first calculated from the branch voltage and branch current at the previous moment;

[0037] The historical current of the current simulation step is combined with the equivalent current of the independent voltage source of the current simulation step to calculate the node injection current vector In;

[0038] Solve the node voltage vector according to the network equation, and calculate the branch voltage vector and branch current vector according to the node voltage;

[0039] The simulation results of each simulation step are recorded repeatedly until the simulation process is terminated. The simulation results include node voltage, charge and discharge power, and battery SOC.

[0040] The present invention provides a distribution network equipment modeling and simulation method based on a digital twin application architecture, which has the following beneficial effects:

[0041] 1. An electromagnetic transient simulation algorithm using digital twin technology to build objects is proposed. The characteristics of object-oriented programming are fully utilized to provide a digital twin spatial description application architecture for distribution network equipment and facilities from three levels: interactive communication, model building, and simulation analysis.

[0042] 2. Using object-oriented methods, the distribution network equipment of the photovoltaic storage charging station was modeled based on the electrical topology, and the twin was combined with the control strategy to provide the equivalent model of the basic unit of the distribution network equipment and the control model of the power conversion unit;

[0043] 3. An electromagnetic transient simulation algorithm using digital twin technology to construct objects is proposed, and electromagnetic transient modeling and operation simulation of the photovoltaic charging station system as a whole are carried out; specifically, differential modeling is carried out for basic components, and average value modeling is carried out for power electronic converters. A systematic electromagnetic transient algorithm is developed on the simulator, giving full play to the hardware advantages of the simulator, and realizing real-time simulation of the photovoltaic charging station power system with millisecond steps; further, in order to verify the solution accuracy and efficiency improvement effect of the simulator, a photovoltaic microgrid power generation system control model is built on the PLECS platform, and the solution result accuracy and simulation efficiency of the method of the present invention and the traditional node analysis method are compared; the comparison results show that the electromagnetic transient simulation algorithm developed based on C++ has better computing efficiency and better real-time performance than traditional offline simulation software.

Brief Description of the Drawings

[0044] Figure 1 This is the architecture diagram of the digital twin space description application of the distribution network.

[0045] Figure 2 This is the topological structure diagram of the photovoltaic storage charging station.

[0046] Figure 3 It is the overall structure diagram of the simulation analysis layer.

[0047] Figure 4It is the overall flow chart of electromagnetic transient simulation algorithm.

[0048] Figure 5 It is the equivalent circuit diagram of the battery.

[0049] Figure 6 This is the battery charge and discharge control block diagram in constant voltage discharge mode.

[0050] Figure 7 This is the battery charging and discharging control block diagram in constant power charging mode.

[0051] Figure 8 It is the topology diagram of the grid-connected converter.

[0052] Fig. 9 It is the voltage and current dual closed-loop control block diagram of the grid-connected converter.

[0053] Fig.10 This is a comparison chart of the DC bus voltage waveform simulated by the simulator and PLECS.

[0054] Fig.11 This is a comparison chart of the charging power waveform of the energy storage battery simulated by the simulator and PLECS.

[0055] Fig.12 This is a comparison chart of the charging power waveforms of the simulator and the PLECS simulated charging pile.

[0056] Fig.13 This is a comparison chart of the state of charge waveform of the energy storage battery simulated by the simulator and PLECS. [Specific implementation method]

[0057] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0058] Example 1

[0059] This embodiment implements a distribution network equipment modeling and simulation method based on a digital twin application architecture.

[0060] The modeling and simulation method of this embodiment is used to solve the technical problems that traditional simulation methods have large computational complexity, low efficiency, and difficulty in fully reflecting the dynamic process of power electronic equipment; the photovoltaic storage charging station has a complex topology structure and various control methods; the solver model is time-consuming to build and difficult to realize real-time. It has better computational efficiency and better real-time performance.

[0061] The specific implementation steps of the modeling and simulation method of this embodiment are as follows:

[0062] 1) Figure 1 This is a diagram of the application architecture of the digital twin space description of the distribution network. Figure 1 As shown in the figure, in order to realize real-time update of data and ensure real-time evolution of the model, the construction of the distribution network digital twin space description application architecture can be divided into three parts: interactive communication layer, model building layer, and simulation analysis layer.

[0063] The interactive communication layer establishes a data interaction mechanism in the physical entity and model building layer, enters the topological information, equipment parameters, monitoring quantities, real-time operation data, etc. of the distribution network equipment and facilities, and realizes the storage and preprocessing of data.

[0064] The model building layer is composed of three types of equipment: primary equipment, secondary equipment and auxiliary sensing equipment as the basis of the digital twin model. Figure 2 This is the topological structure diagram of the photovoltaic charging station. Figure 2 As shown in the figure, taking the photovoltaic storage charging station distribution network equipment as an example, it mainly includes AC power grid, distributed power supply, energy storage and charging load equipment. The AC power grid uses a three-phase voltage source converter to access the DC bus, and the energy storage and charging equipment are connected to the system using different power electronic equipment according to their own charging and discharging characteristics.

[0065] Figure 3 This is the overall structure diagram of the simulation analysis layer. Figure 3 As shown in the figure, the simulation analysis layer uses an object-oriented modeling method to read and monitor the environmental, electrical, physical and behavioral quantities of distribution network equipment and facilities based on the electromagnetic transient simulation algorithm, and collects data such as voltage at key system nodes, state of charge of energy storage and charging equipment, and active power. On the one hand, it improves the accuracy of modeling complex physical mechanisms; on the other hand, it provides a data basis for system status evaluation and analysis and decision-making deduction.

[0066] Figure 4 This is the overall flow chart of the electromagnetic transient simulation algorithm. Figure 4 As shown in the figure, an electromagnetic transient algorithm based on electrical topology is designed to perform object-oriented system modeling and simulation operation simulation on the photovoltaic charging station equipment. Among them, the initialization part first completes the pre-storage of data and the modeling of basic components. On this basis, the core distribution network equipment facilities are modeled and the electrical quantities of the equipment are initialized; the loop part first calls the control function to update the node admittance matrix, calculates the historical current and the injected current, and after the node voltage equation is solved, the node voltage and branch current are updated, and finally the timer is triggered to enter the next loop; the simulation data results are saved at the end of the loop.

[0067] 2) Based on the constructed distribution network digital twin space description application architecture, the basic components of the distribution network equipment are differentially modeled.

[0068] First, the circuit network is numbered, and the node information and parameters of each device in the circuit topology are written into the read record file to complete the pre-storage of data. The pre-stored data is used as the modeling parameters of basic components and core equipment facilities.

[0069] Secondly, the basic components are modeled by differential analysis, such as the capacitors, inductors, resistors and their series-parallel combinations on the DC side, the series-parallel combinations of three-phase capacitors, three-phase inductors and three-phase resistors in the AC circuit, and AC and DC voltage sources. The backward Euler method is selected for differential analysis to ensure the numerical stability of the system. Taking the DC inductor as an example, the differential process is briefly described below.

[0070] Assume that the node i and the node j are connected by an inductor L, and the characteristic equation can be expressed as:

[0071]

[0072] Differencing it using the backward Euler method:

[0073]

[0074] After finishing:

[0075] i L (t) = Y L u L (t)+I h_L (t-Δt)

[0076]

[0077] I h_L (t-Δt)=i L (t-Δt) (3)

[0078] Where Y L is the equivalent admittance of the inductor, I h_L (t-Δt) is the historical current source.

[0079] After the element is differentially modeled, the initial node admittance matrix of the element can be written out from the equivalent admittance of the element, and then the loop body is entered to prepare for the loop to solve the node voltage equation. For basic elements composed of resistors, inductors, capacitors, etc., their admittance matrix does not change with time during the loop process, so there is no need to reconstruct it.

[0080] 3) Based on the constructed distribution network digital twin space description application architecture, model the core distribution network equipment and facilities.

[0081] Taking the photovoltaic energy storage charging station equipment as an example, the grid-connected converter, photovoltaic, energy storage, and charging station equipment are modeled, including the model part and the control part. The model part can be divided into unit components and interface converters for modeling respectively. The interface converter model calls the algorithm of the control part to improve the scalability and versatility of the model.

[0082] Photovoltaic cells are distributed power sources that transmit power to the microgrid via the DC bus. Usually, photovoltaic cells work in maximum power point tracking mode. Since the DC bus voltage is approximately constant, photovoltaic cells can be equivalent to a constant current source model.

[0083] Energy storage systems generally include energy storage batteries and necessary power conversion units. Figure 5 is the equivalent circuit diagram of the battery. Figure 5 As shown, the energy storage battery unit is generally a storage battery. In order to reflect the charge state of the battery and the change of the battery terminal voltage, the storage battery needs to be modeled in detail, which can be equivalent to a controlled voltage source with internal resistance.

[0084] The output voltage of the general equivalent circuit model of the battery is:

[0085] U=V-Ir (4)

[0086] In the formula, U represents the output voltage, V represents the output voltage of the controlled voltage source, r represents the equivalent internal resistance of the model, and I is the discharge current of the energy storage battery.

[0087] According to the discharge characteristic curve of the battery, the output voltage V of the controlled voltage source is:

[0088]

[0089] Where V 0 represents the internal electromotive force; SOC represents the state of charge of the battery; β is the characteristic constant of the battery, which can be calculated by the charge AH ​​of the battery when it is not fully charged. 1 and the corresponding terminal voltage V 1 Calculated.

[0090] The battery state of charge SOC is changed from initial SOC 0 , the battery capacity AH and discharge current I jointly determine:

[0091]

[0092] At the same time, the state of charge SOC can also be determined by the battery charge AH 1 The ratio of its capacity AH is:

[0093]

[0094] The charge of the battery in a certain state is known.1 and terminal voltage V 1 , the corresponding state of charge SOC can be calculated, and then the characteristic constant β of the battery can be calculated according to formula (2).

[0095] The power conversion unit of the energy storage system adopts a bidirectional DC / DC converter, which realizes the bidirectional flow of electric energy through the bidirectional DC / DC converter. The bidirectional DC / DC converter can work in boost mode or buck mode. When the electric energy generated by the distributed power source is in excess, the battery stores the electric energy, and the bidirectional buck / boost converter works in buck mode, at which time the power flows from the grid side to the battery side; when the electric energy generated by the distributed power source is not enough to meet all loads, the battery releases the electric energy, and the bidirectional buck / boost converter works in boost mode, at which time the power flows from the battery side to the grid side. To achieve the above purpose, the bidirectional DC / DC converter has two working modes: constant voltage discharge and constant power charging.

[0096] Figure 6 This is the battery charge and discharge control block diagram in constant voltage discharge mode. Figure 6 As shown in the figure, during constant voltage discharge, the battery's goal is to adjust the DC bus voltage so that the DC bus voltage returns to the rated value and maintains the power balance in the entire microgrid. b ,I dis-ref They are the actual value of battery charge and discharge current and the reference value of battery discharge current in the inner loop obtained by the voltage outer loop. Constant voltage discharge is achieved by voltage and current dual loop control, with the outer loop being the voltage loop and the inner loop being the current loop. In the voltage outer loop, the actual DC bus voltage V is collected. dc Compare it with the reference voltage V dc-ref The difference is sent to the PI regulator and used as the inner loop current reference value I after a current limiting link. dis-ref Enters the current loop and compares with the actual current I b After comparison, the difference is passed through the PI regulator and then enters the PWM modulation link to form a control signal.

[0097] Figure 7 This is the battery charge and discharge control block diagram in constant power charging mode. Figure 7 As shown in the figure, during constant power charging, the battery acts as a power terminal to store the electric energy generated by the distributed power source and the power grid. Constant power charging can be achieved with only one control loop. In the control loop, the battery terminal voltage V and the discharge current I are collected. b , calculate the instantaneous charging power P of the battery b , and compare it with the battery charging reference power P b-ref After comparison, the difference is sent to the PI regulator and then enters the PWM modulation link to form a control signal.

[0098] The energy storage power converter uses a switching control strategy to switch between constant voltage discharge and constant power charging control modes. The control switching signal can be given externally or according to the battery's state of charge. When the energy storage battery is in the low limit area, the energy storage should be avoided from continuing to discharge. When conditions permit, it can be switched to charging to restore the energy storage SOC; when the energy storage battery is in the high limit area, its charging power should be controlled to avoid overcharging of the energy storage; only when the energy storage battery SOC is within the normal range can it be charged and discharged normally.

[0099] The charging load is the battery of the electric vehicle, which is connected to the DC bus through a buck converter to achieve unidirectional charging. The battery unit of the charging load is similar to the energy storage battery unit, so it can be modeled according to the equivalent model of the energy storage battery. The power conversion unit of the charging load is a buck converter. Since the battery terminal voltage is lower than the DC bus voltage, a step-down buck converter is used for charging. At the same time, it can also prevent reverse current input and avoid the problem of reverse charging of the car battery.

[0100] Figure 8 is the topology diagram of the grid-connected converter. Figure 8 As shown in the figure, the grid-connected converter, as the only interface between the DC microgrid and the distribution network, has the following characteristics: it can effectively isolate AC grid disturbances; maintain the stability of the DC bus voltage; and realize the two-way flow of energy. When the output of the distributed power source is greater than the power consumed by the charging load, the energy flows to the grid side through the grid-connected converter, and the excess electric energy is consumed by the distribution network side; when the output of the distributed power source is less than the power consumed by the charging load, the energy flows to the DC bus through the grid-connected converter, thereby supporting the DC bus voltage and transmitting power to the charging load.

[0101] Fig. 9 This is the voltage and current double closed-loop control block diagram of the grid-connected converter. Fig. 9 As shown, the grid-connected converter uses a voltage-current dual closed-loop control method. The voltage outer loop is used to stabilize the DC bus voltage, modulate the working state of the system according to the voltage, and output the command current signal to the current inner loop. The function of the current inner loop is to enable the input current of the system to follow the command current signal, thereby realizing the bidirectional conversion of electric energy under the unit power factor of the system. In the three-phase static coordinate system, since the control system is difficult to design and difficult to achieve zero-static error control, it is necessary to transform it to a two-phase synchronous rotating coordinate system. At the same time, due to the existence of the inductance on the AC side of the system, the dq components of the system (dq components, that is, the two components obtained in the dq transformation (transformation from rectangular coordinate system to polar coordinate system), are commonly used concepts in the field of power system and motor control) are mutually coupled and mutually influenced. Therefore, a feedforward decoupling method is used to realize independent control of the dq components to improve the dynamic response performance of the system. Based on the topology of the LC filter grid-connected interface converter, its mathematical model in the dq coordinate system can be obtained as follows:

[0102]

[0103] Performing Laplace transform on it yields:

[0104]

[0105] After analysis, it is found that in the two-phase synchronous rotating coordinates, there is a coupling phenomenon between the dq components of the mathematical model of the system, so it is necessary to decouple the mathematical model of the system. Because feedforward decoupling has simple control and does not affect the stability of the system, feedforward decoupling control is introduced in the closed-loop system. Add the feedforward amount ωLi to the output voltage d and -ωLi q , so that it is coupled with the coupling term -ωLi in the circuit topology mathematical model d and ωLi q The dq components of the system are converted into two independent parts after decoupling. At this time, the AC control of the system is changed to DC control. In the above rotating coordinate transformation, the d axis represents the active component, the q axis represents the reactive component, and the q axis is specified to lag the d axis by 90°. In this system, the q axis is set to 0, so that the reactive power of the system is zero, thereby realizing the unity power factor rectification of the system.

[0106] Under the condition of three-phase power grid balance, assuming the voltage command formula under three-phase synchronous rotating coordinates, we have:

[0107]

[0108] Combining the above two equations, we can get:

[0109]

[0110] It can be concluded that the coupling component in the two-phase synchronous rotating coordinate system has been completely decomposed, and the proportional coefficient K of the PI regulator is reasonably designed. ip and the integration coefficient K il , the voltage and current dual closed-loop control can be achieved.

[0111] After determining the mathematical model of the core distribution network equipment and facilities and the control method, its electromagnetic transient simulation model can be established.

[0112] In the entire simulation algorithm, for the simulation of the power electronics system level, a large step simulation is used to reduce the number of calculations per unit time, thereby reducing the amount of calculations and facilitating real-time. Therefore, the traditional binary resistance model and constant admittance model are no longer applicable, and the average value model, dynamic phasor model, equivalent circuit model, etc. are often used to improve the simulation efficiency. Among them, the average value model is based on the switch function model, and the average value of the variable in the switch cycle is used to replace its actual value to obtain the functional relationship between its control signal duty cycle and the port voltage and current vector, and then the node admittance matrix and node voltage equation are updated to complete the electromagnetic transient calculation. The average value modeling method only needs to describe the power frequency characteristics of the input / output port of the device, ignoring the high-order harmonics generated by pulse width modulation (PWM), thereby allowing a larger simulation step size to be used, which greatly improves the simulation efficiency by sacrificing a certain degree of accuracy.

[0113] The following takes the three-phase voltage source grid-connected converter as an example to model its average value.

[0114] Assume that the DC port nodes of the converter are i and j, and the AC port nodes are a, b, and c. The bridge arm on-resistance is Ron, the conductance is Yon, and the bridge arm off-resistance is infinite. The duty cycle of the bridge arm on-time in one switching cycle is D k , k=a,b,c.

[0115] According to the KVL law, the AC port voltage is averaged:

[0116]

[0117] Where u a ,u b ,u c is the three-phase voltage of a, b, c on the AC side, u i ,u j is the voltage at DC side i, j, I a , I b , I c is the current flowing into the AC side nodes a, b, c.

[0118] According to the KCL law, the DC port current is averaged:

[0119] I i =-D a I a -D b I b -D c I c

[0120] I j =-(1-D a )Ia -(1-D b )I b -(1-D c )I c (13)

[0121] Where I i , I j is the current flowing into the DC side node i, j.

[0122] The node voltage equation of the AC side of the grid-connected converter is:

[0123] I a =-Y on u a +D a Y on u i +(1-D a )Y on u j

[0124] I b =-Y on u b +D b Y on u i +(1-D b )Y on u j

[0125] I c =-Y on u c +D c Y on u i +(1-D c )Y on u j (14)

[0126] The node voltage equation of the DC side of the grid-connected converter is:

[0127]

[0128] I j =(1-D a )Y on u a -D a (1-D a )Y on u i -(1-D a ) 2 Y on u j +(1-D b )Yon u b -D b (1-D b )Y on u i -(1-D b ) 2 Y on u j +(1-D c )Y on u c -D c (1-D c )Y on u i -(1-D c ) 2 Y on u j (15)

[0129] After forming the node voltage equation, the node admittance matrix of the grid-connected converter is reconstructed according to the duty cycle generated by the control code to complete the electromagnetic transient modeling.

[0130] According to this method, other power electronic devices such as buck converters, bidirectional DC / DC converters, etc. can be modeled based on the port voltage-current relationship, which will not be described in detail here.

[0131] 4) Based on the electromagnetic transient algorithm, complete the information interaction and simulation operation between the photovoltaic storage charging station equipment.

[0132] After the component differential modeling and the establishment of the electromagnetic transient model of the core equipment and facilities, the initial node admittance matrix of the component can be written out, and then the loop body is entered to prepare for the cyclic solution of the node voltage equation. Since the node admittance matrix of some components is time-varying, such as the interface converter, it is necessary to rebuild the admittance matrix. For the basic components composed of resistors, inductors, capacitors, etc., their admittance matrix has been determined before the simulation begins, so no modification is required, thereby saving computing resources and improving simulation efficiency.

[0133] At the beginning of the EMTP program calculation, the historical current item is set according to different initialization methods, and the initial value of the historical current item is usually 0.

[0134] The loop part is implemented by a timing function, and the timing period is set to 1ms, which is equal to the actual simulation step length. When the time taken for each calculation step does not exceed the timing period, real-time simulation of the system can be achieved.

[0135] Different from the calculation process of basic components, the core distribution network equipment and facilities need to call the control function first during the cyclic calculation, calculate the duty cycle of the current simulation step according to the control quantity at the previous moment, and update the admittance matrix based on the average value model.

[0136] In each cycle of calculation, the historical current of the current simulation step is first calculated from the branch voltage and branch current at the previous moment.

[0137] The historical current of the current simulation step is combined with the equivalent current of the independent voltage source of the current simulation step to calculate the node injection current vector In.

[0138] The node voltage vector is solved according to the network equation, and the branch voltage vector and branch current vector are calculated according to the node voltage.

[0139] This cycle repeats and records the simulation results of each simulation step until the simulation process is terminated. After the simulation process is terminated, the simulation results can be written into the record file, including node voltage, charge and discharge power, battery SOC, etc.

[0140] Furthermore, in order to verify the solution accuracy and efficiency improvement effect of the algorithm, a photovoltaic energy storage microgrid power generation system control model was built on the PLECS platform, and the accuracy of the results obtained by the method of this embodiment and the traditional node analysis method was compared. The electromagnetic transient algorithm based on object-oriented average value modeling has a basically consistent response curve with the node analysis method in both steady state and transient conditions, which verifies the correctness of this method as an electromagnetic transient simulation algorithm.

[0141] Example 2

[0142] This embodiment implements a distribution network equipment modeling and simulation method based on a digital twin application architecture.

[0143] For ease of understanding, the method of Example 1 will be described below using a photovoltaic charging station system as an example, but this should not limit the scope of protection of the present invention.

[0144] This embodiment provides a distribution network equipment modeling and simulation method based on a digital twin application architecture. In view of the problems that traditional simulation methods have large computational complexity and low efficiency, and are difficult to fully reflect the dynamic process of power electronic equipment; the topological structure of photovoltaic charging stations is complex, and the control methods are diverse; the solver model is time-consuming to build and difficult to achieve real-time, a distribution network digital twin space description application architecture is designed, and an electromagnetic transient simulation algorithm for building objects using digital twin technology is provided. By utilizing the object-oriented programming characteristics of the C++ language, a digital twin of the photovoltaic charging station distribution network equipment is quickly constructed based on electrical topology, and the equipment modeling and simulation operation simulation of the photovoltaic charging and distribution network are completed. The accuracy of the results obtained by the method of this embodiment is compared with that obtained by the traditional node analysis method on the PLECS platform. The developed electromagnetic transient algorithm has better computational efficiency and better real-time performance than traditional offline simulation software.

[0145] The object-oriented distribution network equipment model is the core of the distribution network digital twin and the basis for realizing simulation calculation and data analysis. The object-oriented modeling method uses objects as the basic unit to build the entire system. Objects are defined as a combination of data and operations. Objects can be understood as containers that encapsulate attributes and methods. Attributes are used to describe the state and characteristics of objects, and methods are used to describe the behavior and operation of objects. Some objects are abstracted according to their commonalities and aggregated into classes. Classes represent the abstraction of the above-mentioned class of objects, and each object represents an instance of the class. The object-oriented structure has three main characteristics: encapsulation, inheritance, and polymorphism. These characteristics effectively protect the security of internal data and improve the reusability and scalability of the code. For the distribution network system, a device-centric approach can be adopted, that is, a specific device is taken as an object. Then such objects are abstracted into commonalities to form corresponding device classes, and their related attribute parameters and general operation methods (functions) are encapsulated into an indivisible unit to hide information details that do not need to be known by the outside world.

[0146] The definition of a class is essentially the definition of class members. The attributes of a class correspond to member variables, and the behavior of a class corresponds to member functions. Members are classified according to their allowed access levels, which can be public, protected, or private. Public members are not subject to access restrictions and can be accessed both within and outside the class. Protected members can be accessed by this class and its derived classes. Private members can only be accessed by the class itself. These access level restrictions enable different degrees of protection and concealment of class members, thereby reducing the correlation between programs and preventing the "butterfly effect".

[0147] Following is the "Battery" class member structure;

[0148]

[0149] The "Battery" class includes two types: public members and protected members. Public members are functions included in the electromagnetic transient simulation algorithm process. Each device class needs to call these functions, while protected members are properties unique to rechargeable batteries, such as the positive and negative poles of the battery, capacity, internal resistance, initial state of charge, etc.

[0150] In order to ensure universality, it is necessary to select a single type of equipment as the smallest unit at a lower level. For example, it is better to create "BuckConverter" and "Battery" classes separately, rather than just creating a "charging station" class. On this basis, different classes can be linked. The charging station is composed of a buck converter and a rechargeable battery, while the energy storage device is composed of a half-bridge converter and a rechargeable battery. The "Battery" class is reused, and only the parameters of the rechargeable battery part of the two types of equipment need to be modified, which is convenient for code management and expansion. In addition, in order to enhance the reusability and structural integrity of the code, an abstract class should be created at a higher level to store the information shared by all its derived classes. The abstract class Sim_CATL is at the top level of the entire model structure, defining the outermost electromagnetic transient calculation function, instances of each distribution equipment class, and key electrical quantity attributes such as voltage and current of the complete electrical topology. Some of the function bodies then call the function members of each distribution equipment subclass instantiation. The final execution of the simulation program is implemented by three functions: simInit, simLoop, and simPlot, which correspond to the program initialization, loop body, and drawing link at the end of simulation, respectively, which include various electromagnetic transient calculation functions defined in the Sim_CATL class.

[0151] An object-oriented electromagnetic transient simulation program based on the average value model was written in C++ in visual studio.

[0152] Set the simulation step, simulation time and total number of system nodes:

[0153] SimulationTime(s),TimeStep(uS),NodeNumber

[0154] 10,1000,26

[0155] Import parameter information of each node and branch of the system:

[0156] AC line impedance:

[0157] n1,n2,R,L

[0158] 7,8,0.01,0.0003

[0159] 7,11,0.01,0.001

[0160] 7,14,0.01,0.001

[0161] Three-phase voltage source:

[0162] n1,V

[0163] 7,311.08

[0164] Battery:

[0165] n1(positive electrode), n2(negative electrode), Vo, R, k, S0C0, AH(A·h) 22, 2, 600, 0.01, 0.888888889, 0.7, 0.74

[0166] 24,2,600,0.01,0.888888889,0.7,0.74

[0167] 26,2,600,0.01,0.888888889,0.7,0.74

[0168] 18,13,600,0.01,0.888888889,0.7,0.74

[0169] 20,16,600,0.01,0.888888889,0.7,0.74

[0170] 6,2,600,0.01,0.888888889,0.7,0.74

[0171] Buck converter:

[0172] n1(Uin+), n2(Uin-), n3(Uo+)

[0173] 1,2,21

[0174] 1,2,23

[0175] 1,2,25

[0176] 12,13,17

[0177] 15,16,19

[0178] DC line impedance:

[0179] n1,n2,R,L,C1,3,0.0224,0,0

[0180] 1,9,0.0224,0,0

[0181] 2,4,0.0224,0,0

[0182] 2,10,0.0224,0,0

[0183] 17,18,0,0.001,0

[0184] 19,20,0,0.001,0

[0185] 21, 22, 0, 0.001, 0

[0186] 23,24,0,0.001,0

[0187] 25, 26.0, 0.001, 0

[0188] 1,2,1e5,0,0.001

[0189] 5,6,0,0.001,0

[0190] DC load:

[0191] n1,R(no 1e5),C9,1.00E+05,2.00E-02

[0192] 10, 1.00E+05, 2.00E-02

[0193] 12,1.00E+05,2.00E-03

[0194] 13,1.00E+05,2.00E-03

[0195] 15,1.00E+05,2.00E-03

[0196] 16,1.00E+05,2.00E-03

[0197] Bidirectional DC / DC converter: n1(Uin+), n2(Uin-), n3(Uo+), numSOC

[0198] 1,2,5,6

[0199] Photovoltaic modules:

[0200] n1(inflow),n2(outflow),I

[0201] 3,4,40

[0202] Three-phase voltage source grid-connected converter:

[0203] n1(DC+), n2(DC-), n3(AC), n4(L node)

[0204] 9,10,8,7

[0205] 12,13,11,7

[0206] 15,16,14,7

[0207] After the model parameters are initialized, the initial node admittance matrix of the system is constructed;

[0208] Enter the loop body, call the control function to output the duty cycle, update the node admittance matrix based on the average value model, calculate the historical current and injection current, solve the node voltage equation, update the node voltage and branch current, update the node voltage and branch current, and wait for the timer to trigger the next loop. If the simulation time reaches the set time, stop the loop, output the node voltage, DC bus voltage drop, energy storage battery power, charging battery power, battery state of charge, etc. and save them in the simlog.txt file.

[0209] After the simulation is completed, the drawing function is called to draw the DC bus voltage curve, energy storage charging and discharging power curve, AC charging pile charging power curve, and energy storage battery charge state curve.

[0210] The timing performance of the simulation is shown below. When the total number of cycles is 4000, the number of single cycles exceeding 1.1 ms does not exceed 500 times, and the number of cycles below 0.9 ms does not exceed 600 times, which indicates that the present invention has good timing performance.

[0211]

[0212] In order to verify the accuracy and efficiency improvement of the simulator, a control model of the photovoltaic energy storage microgrid power generation system was built on the PLECS platform. The simulation parameters are the same as those of the electromagnetic transient simulator:

[0213] In the photovoltaic and energy storage microgrid, the rated voltage of the DC bus is 750V, the AC side grid line voltage is 380V, the rated charging power of the energy storage battery is 250kW, and the rated charging power of the charging pile is 120kW.

[0214] Photovoltaic micro power supply simulation parameters: Use 40A constant current source.

[0215] Energy storage micro power supply simulation parameters: bidirectional DC-DC circuit inductance L = 0.001H, battery internal resistance 0.01Ω, full SOC voltage V = 600V, initial SOC = 0.7.

[0216] AC power supply simulation parameters: phase voltage effective value: 220V, frequency: 50Hz, AC side inductance of grid-connected inverter: 0.3mH, AC side resistance: 0.01Ω, DC side filter capacitor: 0.01F.

[0217] Charging pile simulation parameters: buck circuit inductance L = 0.001H, battery internal resistance 0.01Ω, full SOC voltage V = 600V, initial SOC = 0.7.

[0218] The simulation operation conditions are as follows:

[0219] From 0 to 1s, the DC bus voltage is maintained at 750V by the grid-connected inverter, the energy storage is charged at 250kW, and the charging pile is charged at 120kW.

[0220] 1-4s, the energy storage is changed to constant voltage control, and the DC bus voltage is controlled by the energy storage and the grid-connected converter at 750V. The charging pile is charged at 120kW.

[0221] By comparing with the photovoltaic storage charging station model built in PLECS, the DC bus voltage curve, energy storage charging and discharging power curve, AC charging pile charging power curve, and energy storage battery charge state curve of the two platforms are compared to verify the accuracy requirements of this embodiment.

[0222] Fig.10 This is a comparison chart of the DC bus voltage waveform simulated by the simulator and PLECS. Fig.11 This is a comparison chart of the charging power waveform of the energy storage battery simulated by the simulator and PLECS. Fig.12 This is a comparison chart of the charging power waveforms of the simulator and the PLECS simulated charging pile. Fig.13 This is a comparison chart of the state of charge waveform of the energy storage battery simulated by the simulator and PLECS. Fig.10 , Fig.11 , Fig.12 , Fig.13 As shown in Figure 1, the accuracy of the results obtained by the method of this embodiment is compared with that of the traditional node analysis method. It can be seen that the object-oriented average value modeling method has a basically consistent response curve with the node analysis method in both steady-state and transient conditions, which verifies the correctness of this method as an electromagnetic transient simulation algorithm. The simulator solution performance comparison and efficiency verification results are shown in Table 1.

[0223] Table 1 Comparison of simulation efficiency between simulator and PLECS

[0224]

[0225] In order to verify the solving accuracy and efficiency improvement effect of the simulator, a control model of a photovoltaic energy storage microgrid power generation system was built on the PLECS platform. The solution accuracy and simulation efficiency of the method of the present invention were compared with those of the traditional node analysis method. Compared with traditional offline simulation software, the method has better computing efficiency. After being compiled into a dynamic library, it can be called and run in other development environments, and has better real-time performance.

[0226] In summary, this embodiment is a distribution network equipment modeling interaction method and system based on the digital twin application architecture. It performs differential modeling on basic components and average value modeling on power electronic converters. A systematic electromagnetic transient method is developed on the simulator, giving full play to the hardware advantages of the simulator and realizing real-time simulation of the power system of the photovoltaic storage charging station with a millisecond step size.

[0227] Those skilled in the art will appreciate that all or part of the steps to implement the above embodiments may be accomplished by hardware, or may be accomplished by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, wherein the storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0228] The above is only a preferred embodiment of the present invention. It should be pointed out that ordinary technicians in this technical field can make several improvements and supplements without departing from the principle of the present invention. These improvements and supplements should also be regarded as the scope of protection of the present invention.

Claims

1. A distribution network equipment modeling and simulation method based on a digital twin application architecture, wherein the distribution network equipment is a photovoltaic charging station, the basic components of which include inductance components, resistance components and capacitance components, and the core distribution network equipment facilities of which include photovoltaic cells, energy storage systems, grid-connected converters and charging loads; characterized in that The following steps are involved: S1. Construct a distribution network digital twin spatial description application architecture including an interactive communication layer, a model building layer, and a simulation analysis layer; S2. Based on the constructed distribution network digital twin space description application architecture, the basic components of the distribution network equipment are modeled by differentiation; S3. Model the core distribution network equipment and facilities based on the constructed distribution network digital twin space description application architecture; S4. Based on the electromagnetic transient algorithm, complete the information interaction and simulation operation between distribution network equipment.

2. According to a distribution network equipment modeling and simulation method based on a digital twin application architecture according to claim 1, it is characterized in that: The interactive communication layer of the distribution network digital twin space description application architecture is used to establish a data interaction mechanism in the physical entity and model building layer, input the topological information, equipment parameters, monitoring quantity, and real-time operation data of the distribution network equipment and facilities, and realize the storage and preprocessing of data; The distribution network digital twin space describes the application architecture model building layer, with primary equipment, secondary equipment and auxiliary sensing equipment as the basis of the digital twin model; The distribution network digital twin space description application architecture simulation analysis layer uses an object-oriented modeling method and an electromagnetic transient simulation method to read and monitor the environmental quantities, electrical quantities, physical quantities and behavioral quantities of distribution network equipment and facilities, and collects the system key node voltage, energy storage and charging equipment charge status, and active power data.

3. According to a distribution network equipment modeling and simulation method based on a digital twin application architecture according to claim 2, it is characterized in that: Step S2 specifically includes the following sub-steps: Number the circuit network, and write the node information and parameters of each device in the circuit topology into the read record file to complete the pre-storage of data; Based on the pre-stored data, the backward Euler method is selected to perform differential modeling of the basic components.

4. A distribution network equipment modeling and simulation method based on a digital twin application architecture according to claim 3, characterized in that: Inductor Component Differential Model L (t) = Y L u L (t)+I h_L (t-Δt); Among them, Y L is the equivalent admittance of the inductor, I h_L (t-Δt) is the historical current source, u L (t) is the inductor terminal voltage, i L (t) is the inductor current at the current moment.

5. According to a distribution network equipment modeling and simulation method based on a digital twin application architecture according to claim 3, it is characterized in that: Step S3 specifically includes the following sub-steps: Photovoltaic cells are distributed power sources, equivalent to a constant current source model; The energy storage system includes an energy storage battery, which is equivalent to a controlled voltage source model with internal resistance; The grid-connected converter uses feed-forward decoupling to achieve independent control of dq components; The charging load is the electric vehicle battery, which is equivalent to a controlled voltage source model with internal resistance; Output voltage of the controlled voltage source model with internal resistance of the energy storage system Among them, V0 represents the internal electromotive force, SOC represents the state of charge of the battery, β is the characteristic constant of the battery, I is the discharge current of the energy storage battery, and r represents the internal resistance of the battery; Mathematical Model of Grid-connected Converter Among them, K ip is the proportional coefficient of the PI regulator, K il is the integral coefficient of the PI regulator, s is a complex variable, i q is the q-axis component of the AC side current, for i q Current reference value, L is the inductance value on the AC side, R is the line impedance on the AC side, i d is the d-axis component of the AC side current, for i d Current reference value; The three-phase voltage source grid-connected converter is updated by the average value modeling method, and the AC side node voltage equation and the DC side node voltage equation are obtained as follows: Among them, I a , I b , I c is the current flowing into nodes a, b, c on the AC side, u a ,u b ,u c is the three-phase voltage of a, b, c on the AC side, u i ,u j is the DC side i, j voltage, D is the duty cycle, Y on is the conductivity, I i , I j is the current flowing into the DC side node i, j.

6. A distribution network equipment modeling and simulation method based on a digital twin application architecture according to claim 5, characterized in that: Step S4 specifically includes the following sub-steps: When the core distribution network equipment and facilities perform cyclic calculations, they first call the control function, calculate the duty cycle of the current simulation step based on the control quantity at the previous moment, and update the admittance matrix based on the average value model; In each cycle calculation process, the historical current of the current simulation step is first calculated from the branch voltage and branch current at the previous moment; The historical current of the current simulation step is combined with the equivalent current of the independent voltage source of the current simulation step to calculate the node injection current vector In; Solve the node voltage vector according to the network equation, and calculate the branch voltage vector and branch current vector according to the node voltage; The simulation results of each simulation step are recorded repeatedly until the simulation process is terminated. The simulation results include node voltage, charge and discharge power, and battery SOC.

Images