Optical storage integrated UPQC multi-priority load capacity ellipsoid constraint locating and sizing method

Abstract

The invention discloses a light-storage integrated UPQC multi-priority load capacity ellipsoid constraint locating and sizing method, and belongs to the technical field of power distribution network optimization operation and electric energy quality comprehensive control. The invention aims to provide a power distribution network multi-priority load-oriented capability ellipsoid constraint UPQC-optical storage integrated device site selection and capacity optimization method. The optical storage integrated UPQC multi-priority load capability ellipsoid constraint site selection and capacity optimization method realizes coordination and balance among electric energy quality management, energy storage energy safety and system economy. According to the multi-priority load differentiation constraint, loads are divided according to key, important and general classes, and weighted compliance rate indexes and hierarchical power quality constraints are introduced. According to the method, capacity redundancy and control conflicts caused by traditional independent constraints are avoided, the feasibility and precision of device site selection and constant volume optimization are improved, and unified optimization of capacity configuration, electric energy quality management and energy supporting capacity is achieved.

Description

Technical Field

[0001] This invention belongs to the field of power distribution network optimization operation and power quality integrated control technology. Background Technology

[0002] With the increasing penetration of new energy sources and the growing level of electronic power in end-user loads, power quality management in distribution networks has become a crucial link in ensuring the safe, stable, and economical operation of the system. Distribution networks not only undertake tasks such as multi-source grid connection, power flow distribution, and voltage support, but also directly affect the power supply quality on the user side. Currently, distributed photovoltaic output is significantly affected by weather conditions, resulting in frequent power fluctuations. Simultaneously, the large-scale integration of new loads such as electric vehicle charging facilities, data centers, semiconductor manufacturing, and medical equipment has led to the system exhibiting "dual high" characteristics—high volatility and high sensitivity. This has resulted in increasingly prominent problems such as node voltage fluctuations, harmonic distortion, voltage dips, and short-term power outages, seriously affecting the safe and stable power supply to critical loads. Traditional reactive power regulation methods (such as on-load tap changers, parallel capacitors, and static var compensators, SVG) are insufficient in response to these issues. With limited speed, it is difficult to maintain voltage quality and system stability under conditions of high fluctuation and high harmonics.

[0003] To address this, the field of power quality control has proposed the "Unified Power Quality Conditioner (UPQC)," a composite compensation device. UPQC combines series and parallel compensators, simultaneously achieving voltage regulation, harmonic suppression, and power factor correction. Its core principle is to offset voltage disturbances through series compensators and suppress current harmonics through parallel compensators, thus achieving bidirectional management of voltage and current quality within the same device. Existing engineering practices have demonstrated the significant effectiveness of UPQC in suppressing voltage flicker, harmonic distortion, and voltage dip ride-through. However, traditional UPQCs lack energy support capabilities; their DC bus voltage relies entirely on grid-side balance. When a deep voltage dip or short-term power outage occurs, the UPQC often fails due to the DC voltage drop, unable to continuously maintain power supply to critical loads. This limitation causes a significant decrease in the compensation performance of UPQC under extreme operating conditions.

[0004] To overcome the lack of energy support in UPQC (Power Quality Control Center), researchers have recently proposed introducing energy storage and photovoltaic (PV) systems into the UPQC DC bus, forming an integrated UPQC-PV-storage device. This device utilizes the energy buffering capacity of the energy storage unit to stabilize the DC link, and absorbs renewable energy output through the PV port. It can perform active and reactive power coordination compensation under normal operating conditions, and independently support critical load power supply during short-term power outages or voltage dips. The introduction of PV and energy storage expands UPQC from a traditional "power quality regulation device" to a "multi-port system with energy routing and dynamic support capabilities," achieving integration of energy management and power quality governance. However, existing research mainly focuses on device-level control strategies and structural improvements, lacking quantitative analysis and optimization of its configuration location and capacity at the system level.

[0005] On the other hand, with the widespread integration of distributed photovoltaic and energy storage into the distribution network, the system's operating status is not only affected by reactive power distribution but also closely related to active power flow and energy storage dispatch. Traditional voltage / VAR optimization (VVO) models only focus on voltage constraints, making it difficult to coordinate active power output from energy storage, UPQC compensation, and voltage fluctuation control in dynamic environments. While recent active-reactive power joint dispatch methods have introduced multi-objective optimization concepts, they still have shortcomings. (1) Most studies assume that the compensation device is fixed and fail to determine its optimal location and capacity at the planning level. (2) Existing control strategies usually only target a single indicator (such as voltage or THD) for constraint optimization, and lack a comprehensive evaluation and control framework for differentiated power quality constraints for multi-priority loads; (3) Traditional UPQC and photovoltaic-storage coordinated control mostly adopt separate power constraints (i.e., independently limit active, reactive and voltage injection power), which makes it difficult to accurately reflect the dynamic capability coupling relationship of multi-port devices. Summary of the Invention

[0006] The purpose of this invention is to propose a capacity ellipsoidally constrained UPQC-PV-storage integrated device site selection and capacity optimization method for multi-priority loads in distribution networks, so as to achieve a coordinated balance between power quality management, energy storage security and system economy.

[0007] The steps of this invention are: S1. Multi-priority load differentiation constraints: loads are divided into three categories: critical, important and general, and weighted compliance rate indicators and graded power quality constraints are introduced. S11. Voltage and Harmonic Distortion Constraints: The voltage deviation and harmonic distortion level of various load nodes at any time t should meet the following constraints: (10) in, This refers to the relative deviation of the node voltage. The total harmonic distortion rate of the node; The limit values ​​for different load categories are as follows: Critical load: ; Important loads: ; Typical load: ; S12, Voltage Dip Ride-Through Constraint: Introduce voltage ride-through constraints for critical load nodes: (11) in, In order to cross the lower limit, This refers to the duration of the voltage drop. S13, Introduce the node harmonic voltage sensitivity matrix Its definition is as follows: (12) in, Let i be the voltage offset of node i under the ℎth harmonic. Harmonics of the nth UPQC device Current injection amount, This represents the harmonic sensitivity of node i to the current injected into device n; S2. Construct a multi-objective optimization function: (13) in: This refers to the node priority weight coefficient. Branch resistance; These are the device's rated apparent power and energy storage capacity, respectively; C S C E This is the unit capacity cost coefficient. This is the lifetime degradation coefficient; In equation (13): The first item represents the weighted sum of squares of voltage deviation, which is used to measure the system voltage quality; The second item represents the weighted sum of squares of harmonic distortion, reflecting the quality of the electrical waveform. The third item represents line power loss, which reflects the economic efficiency of distribution network operation; Item 4 is the investment and energy storage capacity cost of the UPQC-integrated photovoltaic and energy storage device; Item 5 is the cost of energy storage battery life loss; S3. Establishment of the constraint model: The DC-side power balance of the UPQC-photovoltaic-storage integrated device at any time t is: (15) in: Photovoltaic port power; Energy storage port power; : Active power on the AC side of the parallel port; : Active power on the AC side of the series port; : Converter efficiency; System losses; Introducing the composite power component: (16) The normalized ellipsoidal constraint for the device in the three-dimensional power space of DC active power injection – parallel reactive power injection – series injection is defined as follows: (17) Where: S is the total apparent capacity of the device; This refers to the rated capacity of the parallel converter; This refers to the rated capacity of the series converter; For the reactive power of the parallel ports; For series injection voltage; Line current; S4. Set of runtime constraints: S41. Power flow balance and voltage safety constraints: (18) At the same time, the node voltage and branch current should satisfy: (19) S42, DC-side energy conservation constraint (20) in, For photovoltaic port output power, For energy storage power, To improve the energy transfer efficiency of parallel and series converters, For system losses; S43, Parallel compensation port capacity constraint (twenty one) in Rated apparent capacity for parallel ports; S44, Series Compensation Port Voltage and Capacity Constraints (twenty two) in, For injection voltage, For line current, Indicates the magnitude of the injected voltage; S45, Energy Storage State of Charge Dynamics and Emergency Floor Constraints (twenty three) And should satisfy: (twenty four) Where E is the rated energy capacity of the energy storage. For charge and discharge efficiency, SOC floor For energy storage emergency flooring; S46. Energy storage charging and discharging mutual exclusion and power smoothing constraints: Set mutual exclusion conditions: (25) Set smoothing constraints: (26) S47, Photovoltaic Output and Power Constraints The output power of the photovoltaic port is limited by irradiance and inverter capacity. (27) Light curtailment is permitted when system capacity or voltage constraints are triggered: (28) S48. Power Quality and Voltage Ride-Through Constraints Node voltage deviation and harmonic distortion should meet the following requirements: (29) Critical loads must meet ride-through constraints during voltage dips: (30) Introducing the harmonic voltage sensitivity matrix: (31) in, Let i be the voltage sensitivity of device n to node i under the ℎth harmonic; S5, Two-level optimization solution S51, Upper-level optimization, i.e., planning layer: S511, Decision Variables ∈ {0,1}: Whether the nth candidate node has a device installed; : Rated apparent power of the device; Rated energy capacity of energy storage; S512, Upper-level optimization objective function (32) in: The first item represents the equipment investment and fixed-capacity cost; the second item represents the expected operating cost under different typical operating scenarios. For scene sets; The overall intraday cost returned for optimizing lower-level operations; S513, Constraints (33) Budget refers to the upper limit of the investment budget; S52, Lower-level optimization, i.e., runtime layer: S521. Run the target function: (34) Among them, control variables It includes quantities such as parallel port reactive power output, series injection voltage, energy storage charging and discharging power, and photovoltaic output regulation; S522, Constraints In the lower-level operation optimization, all of the following constraints must be met: (1) Node power flow balance and voltage and current safety boundaries; (2) DC energy conservation and port capacity constraints; (3) Capability ellipsoid constraint; (4) Dynamic equations of energy storage SOC and emergency floor constraints; (5) Photovoltaic power output and curtailment constraints; (6) Multi-priority power quality and voltage ride-through constraints; S523, The lower-level optimization problem is formalized as follows: (35) in Let these represent the sets of equality and inequality constraints, respectively. S523, Rolling Prediction and Execution Mechanism (1) At the current moment To obtain photovoltaic and load forecast data for the next H steps; (2) Construct the corresponding SOCP optimization model and obtain the optimal control sequence. ; (3) Only execute the first step of the control quantity. The system enters the next moment; (4) Update the measurement and forecast data, rebuild the optimization model, and repeat the process. S524, Upper and Lower Layer Collaboration Relationship (1) Data exchange between the upper and lower layers is achieved through runtime costs and configuration parameters; (2) Location, capacity and energy storage scale of upper output device ; (3) The lower layer performs rolling optimization under a given configuration and feeds back the average running cost. To the upper level; (4) The upper layer iteratively updates the site selection and capacity setting scheme based on feedback to form an investment-performance balance solution.

[0008] The advantages of this invention are: (1) A capacity modeling method based on capacity ellipsoid constraints is proposed. The capacity boundary of the photovoltaic-storage integrated UPQC in the active, reactive and voltage injection dimensions is uniformly described by the ellipsoid form, which accurately describes the power coupling relationship of the multi-port, avoids the capacity redundancy and control conflict caused by traditional independent constraints, and improves the feasibility and accuracy of device site selection and capacity optimization; (2) Establish a differentiated configuration mechanism for multi-priority loads. Set graded power quality constraints and weighted objective functions according to the importance of the loads to achieve rigid guarantee of power supply quality for critical loads and economical configuration of general loads, thereby improving the system's adaptability to highly sensitive and highly fluctuating loads; (3) Achieve comprehensive optimization of power quality and energy constraints. In the site selection and capacity determination model, voltage deviation, harmonic distortion and energy storage constraints are considered simultaneously to form a system-level configuration framework for photovoltaic-storage integrated UPQC, so as to achieve unified optimization of capacity configuration, power quality management and energy support capabilities. Attached Figure Description

[0009] Figure 1 is a structural diagram of the UPQC+ integrated photovoltaic and energy storage system; Figure 2 is a flowchart of the invented algorithm; Figure 3 shows the variation curve of the minimum voltage of the entire network under typical daily operation of the UPQC-photovoltaic-storage integrated device; Figure 4 shows the reactive power output |Q| support curves of the two ports (nodes 18 and 30) on the parallel side under a typical day; Figure 5 is a comparison of the harmonic suppression performance of the benchmark scheme and the device of the present invention; Figure 6 shows the weighted sliding window compliance curves for multi-priority loads (vertical axis range 0–110%). Figure 7 shows the 24-hour state of charge (SOC) and charge / discharge power curves of the energy storage system; Figure 8 shows the investment-performance Pareto front considering hard constraints on voltage and harmonics. Detailed Implementation

[0010] This invention proposes a coordinated control and site selection method for integrated photovoltaic-storage UPQC (Upgraded Product Quality Control) based on capacity ellipsoid constraints for multi-priority loads. It achieves coordinated operation, constraint consistency, and configuration optimization of the devices at the system level, while considering voltage support, harmonic suppression, energy balance, and economic objectives. This method has significant engineering implications and widespread application value. This invention aims to overcome problems in existing technologies such as the disconnect between UPQC and energy storage operation, lack of consideration for load priorities, and insufficient system optimization in device configuration.

[0011] To achieve the above objectives, the present invention provides a method for site selection and capacity optimization of UPQC-integrated photovoltaic-storage devices, comprising the following steps: Step 1: Obtaining Distribution Network Operation Information. Collect the node topology, line impedance parameters, typical daily load curves, and distributed photovoltaic power output data of the distribution network. Determine the voltage range and line current limit of each node, and establish an operation model reflecting the relationship between node voltage, current, and power flow.

[0012] Step Two: The Working Mechanism of the UPQC – Integrated Photovoltaic and Energy Storage Device. The UPQC – integrated photovoltaic and energy storage device is introduced into the candidate node. This device interconnects the photovoltaic port, energy storage port, grid-connected port, parallel compensation port, and series compensation port via a shared DC bus, enabling multi-port energy interaction. The coordinated operation of these multiple ports gives the UPQC – integrated photovoltaic and energy storage device active and reactive power regulation capabilities, while also providing integrated support for power quality management in the distribution network.

[0013] Step 3: Multi-priority load classification and differentiated control. Based on the importance of power supply, system loads are divided into three categories: critical, important, and general. Critical loads require continuous power supply during voltage dips or short-term power outages, and are subject to the strictest voltage deviation and harmonic limits. Important and general loads are allowed limited fluctuations and participate in the optimization objective through weighted parameters.

[0014] Step Four: Establishing Comprehensive Optimization Objectives This invention sets comprehensive optimization objectives at the distribution network level to balance power quality, operational economy, and energy storage lifespan. The optimization objectives include: 1) Voltage deviation and harmonic distortion indicators, used to quantify power quality levels. 2) Active power loss and equipment operating costs, used to reflect system economics. 3) Equipment investment and energy storage lifespan indicators, used to constrain the long-term sustainability of the configuration.

[0015] Step 5: Capacity Ellipsoid Constraint. To describe the capacity limitations of the UPQC-integrated photovoltaic-storage device under multi-port coupling conditions, this invention proposes a capacity ellipsoid constraint mechanism. This constraint simultaneously limits the device's active power output, reactive power support, and series injection capabilities in the form of an ellipsoidal envelope, thereby establishing the correlation boundary between the power of each port.

[0016] Step Six: Setting Operational Constraints for Energy Storage and Power Quality. During the comprehensive optimization process, dynamic constraints on both the energy storage system and power quality are considered. Energy storage constraints include the time evolution of the state of charge (SOC), upper and lower limits, and emergency floor restrictions, to ensure that the energy storage has energy support capabilities during normal operation and short-term power outages.

[0017] Step Seven: Two-Layer Optimization Mechanism. To achieve coordinated optimization of device configuration and operation scheduling, this invention establishes a two-layer optimization framework. The upper-layer optimization employs a multi-objective evolutionary algorithm (NSGA-II) to optimize the site selection and capacity configuration of the UPQC-integrated photovoltaic-storage device within the candidate node set, forming a Pareto front for investment cost and performance indicators. The lower-layer optimization, based on the upper-layer scheme, uses a second-order cone programming method based on rolling predictive control (SOCP-MPC) to dynamically and jointly optimize active power, reactive power, and energy storage status at each time period, achieving power quality regulation and energy balance throughout the entire operation process.

[0018] In the specific implementation of this invention, in order to achieve the site selection and capacity optimization of the UPQC-photovoltaic-storage integrated device for multiple priority loads of the distribution network, the entire invention performs the following steps in sequence.

[0019] Step 1: Obtaining Distribution Network Operation Information In an embodiment of the present invention, basic operating information of the distribution network is first obtained, including network topology, line parameters, load characteristics, and output prediction of distributed photovoltaic power generation.

[0020] 1. Network topology and electrical parameters The power distribution network is described using a node-branch structure, with N nodes and L branches. For each line... Collect line impedance And calculate the admittance matrix parameters: (1).

[0021] The allowable range of node voltage is defined as follows: (2) This ensures stable operating voltage.

[0022] 2. Load and Photovoltaic Data For each node, determine the time-series load power based on the typical daily load curve. And obtain the active power output of distributed photovoltaic power based on prediction models or monitoring data. .

[0023] In implementation, the 24 hours are discretized into 96 time periods, each 15 minutes long, to establish a rolling time frame. With optimized time-series models, photovoltaic output can be predicted using meteorological irradiance. (3) Where, η pv For photovoltaic conversion efficiency, A pv For the component area, G t This refers to irradiance.

[0024] 3. System power flow balance relationship At the node level, the active and reactive power balance of the system should satisfy: (4) In the formula: Inject active and reactive power into the nodes; For load power; θ represents the node voltage amplitude. ij The phase angle difference represents the node voltage. This equation guarantees nodal power flow balance and energy conservation, forming the basis for subsequent optimization constraints.

[0025] 4. Time Scale and Data Update Distribution network operation information is input into the optimization module in a discrete time-series format within the day, and updated every 15 minutes. Photovoltaic forecast values ​​can be dynamically adjusted based on rolling forecast results, and the load curve is corrected from historical data to form a typical daily profile. Through this mechanism, the system can achieve real-time data updates and rolling iterations of the optimization model during operation. These steps complete the acquisition of distribution network operation information and parameter initialization, providing comprehensive data support for the site selection, capacity determination, and scheduling optimization of the UPQC-PV-Storage Integrated Device.

[0026] Step 2: Operation Mechanism of UPQC-Integrated Photovoltaic and Energy Storage Device In an embodiment of the present invention, the UPQC-photovoltaic-storage integrated device is used as a multi-port power electronic system with dual functions of power quality regulation and energy management, as shown in Figure 1. Its topology includes: a photovoltaic port, an energy storage port, a grid-connected port, a parallel compensation port, and a series compensation port. Each port interacts with energy through a shared DC bus. The device internally contains an energy storage unit, a voltage-source converter, and a controller, realizing active and reactive power joint regulation, harmonic suppression, and voltage support functions.

[0027] 1. DC side energy balance relationship To ensure consistent energy flow under different operating conditions, the DC-side power balance should meet the following requirements: (5) In the formula: Output power for the photovoltaic port; This represents the power of the energy storage port (positive for discharging, negative for charging). These are the AC active power of the parallel and series ports, respectively; The efficiencies of parallel and series converters are respectively; This refers to system losses.

[0028] This formula shows that the DC input power (photovoltaic + energy storage) of the device is equal to the sum of the AC output power and system losses, thus ensuring energy conservation. During normal operation, the photovoltaic port mainly provides active power during the day, while the energy storage port is used for peak shaving and valley filling and smoothing photovoltaic power. When the grid voltage drops or the load changes abruptly, the energy storage port discharges in reverse to support the PQC DC bus voltage, ensuring that the device has continuous voltage regulation capability.

[0029] 2. Well-connected compensation port Parallel compensation ports are used to suppress harmonic currents, provide reactive power support, and correct power factors. Their operating capacity is limited by the rated apparent power. (6) In the formula S sh Rated capacity for parallel ports.

[0030] 3. Series compensation port The series compensation port is mainly used to offset voltage fluctuations, three-phase imbalances, and voltage drops. Its voltage injection and capacity constraints are as follows: (7) In the formula: ρ is the upper limit of the injection voltage ratio, which is taken as 0.2; Line voltage; S represents the line current. se This refers to the rated capacity of the series converter.

[0031] 4. Dynamic Constraints of Energy Storage SOC The energy storage unit provides energy buffering and emergency support for the device, and its state of charge (SOC) meets the following requirements: (8) And set: (9) Where: E is the rated energy capacity of the energy storage; These represent charge / discharge efficiency and SOC, respectively. floor This indicates the energy storage emergency floor (generally 10%).

[0032] This constraint ensures that the energy storage system maintains energy balance during charge-discharge cycles, while providing short-term power outage support for critical loads. Through the above operating mechanism, the UPQC-PV-energy storage integrated device can achieve active power distribution and reactive power support during normal operation, and maintain system voltage stability through energy storage discharge and series voltage injection during faults or voltage disturbances, thereby achieving integrated operation of energy management and power quality governance.

[0033] Step 3: Multi-priority load grading and constraint differentiation settings In embodiments of the present invention, to reflect the sensitivity of different types of loads to power quality and their power supply guarantee requirements, loads in the distribution network are divided into three categories according to their importance: critical loads, important loads, and general loads. This classification method is used to implement differentiated voltage and harmonic constraints during the optimization process, thereby prioritizing the power supply quality of critical users under limited resource conditions.

[0034] 1. Load type and priority coefficient The definitions and priority coefficients of various load types are shown in Table 1: .

[0035] Priority coefficient By introducing a weighting factor for voltage and harmonic deviations into the optimization objective function, the optimization model can automatically increase the constraint priority of critical loads during calculation.

[0036] 2. Voltage and Harmonic Distortion Constraints The voltage deviation and harmonic distortion level of various load nodes at any time t should meet the following constraints: (10) in, This refers to the relative deviation of the node voltage; This represents the total harmonic distortion (THD) of the node. The limit values ​​for different load categories are as follows: Critical loads: Important loads: General load: .

[0037] Constraints for critical loads are hard constraints that must be satisfied; constraints for important and general loads are soft constraints that can be satisfied in specific situations. The objectives are reflected in the form of penalty factors, so that the system can take into account the overall economy as much as possible while meeting the quality requirements of key nodes.

[0038] 3. Voltage sag pass-through constraint In extreme conditions such as voltage dips or short-term power outages, to ensure continuous power supply to critical loads, this invention introduces voltage ride-through constraints for critical load nodes: (11) In the formula, To cross the lower limit, we use 0.95 pu; This refers to the duration of the voltage drop.

[0039] This constraint requires that during voltage dips, the voltage at critical load nodes must not fall below a lower limit. The constraint is implemented by... The UPQC-PV-storage integrated device uses a series compensation port and a parallel reactive power support port together. The series port injects additional voltage to raise the load-side voltage, while the parallel port provides reactive power to improve the power factor.

[0040] 4. Harmonic sensitivity relationship and distribution influence To reflect the harmonic mitigation effect of the UPQC-integrated photovoltaic-storage device in a multi-node system, this invention introduces a node harmonic voltage sensitivity matrix. Its definition is as follows: (12) In the formula, Let i be the voltage offset of node i under the ℎth harmonic. Let n be the harmonic current injection amount of the nth UPQC device. This represents the harmonic sensitivity of node i to the current injected into device n.

[0041] The sensitivity matrix can be derived from the harmonic impedance matrix of the distribution network and is used to establish the node THD and device parameters. The linear approximation relationship between the input currents facilitates incorporating the harmonic mitigation effect into the capacity and power constraints during subsequent optimization solutions.

[0042] Step 4: Comprehensive optimization of the objective function In this invention, the site selection and capacity determination are comprehensively considered under multiple objectives, including active and reactive power coordination control, power quality improvement, energy-efficient operation and equipment life extension of the UPQC-photovoltaic-storage integrated device.

[0043] This invention constructs the following multi-objective optimization function: (13) In the formula: the first term represents the weighted sum of squares of voltage deviation, used to measure system voltage quality; the second term represents the weighted sum of squares of harmonic distortion, reflecting the quality of power waveform; the third term represents line power loss, used to reflect the economic efficiency of distribution network operation; the fourth term represents the investment and energy storage capacity cost of the UPQC-photovoltaic-energy storage integrated device; and the fifth term represents the cost of energy storage battery lifespan degradation. This is the node priority weight coefficient, corresponding to the load category (critical, important, and general are assigned 3, 2, and 1 respectively); Branch resistance; These are the device's rated apparent power and energy storage capacity, respectively. This is the unit capacity cost coefficient.

[0044] In the energy storage lifespan loss item, The lifespan degradation factor, considering the effects of SOC range and depth of discharge (DOD), is set with reference to the "Guideline for Lifespan Assessment of Power Storage Batteries (2023)" as follows: (14).

[0045] By introducing this multi-objective function, this invention achieves improved power quality and economic efficiency within a unified optimization model. A multi-dimensional balance between operation and energy storage life protection is achieved, and the active-reactive power regulation and energy storage management of the UPQC-PV-energy storage integrated device are integrated and optimized.

[0046] Step 5: Capability Ellipsoid Constraints In embodiments of the present invention, to accurately reflect the comprehensive power capability boundary of the UPQC-PV-storage integrated device under multi-port collaborative operation and to prevent capacity redundancy and control imbalance caused by traditional port linear limits, a Capability Ellipsoidal Constraint model is proposed. This constraint incorporates "DC-side active power, parallel port reactive power, and series port injected power" into the same coordinate system in a unified geometric form, realizing dynamic coupling modeling of the device's multi-port capabilities.

[0047] 1. Establishment of the constraint model The DC-side power balance of the UPQC-integrated photovoltaic and energy storage device at any time t is: (15) in: Photovoltaic port power (output is positive); Energy storage port power (positive for discharging, negative for charging); : Active power on the AC side of the parallel port; : Active power on the AC side of the series port; : Converter efficiency; System losses.

[0048] Based on this, the present invention introduces a comprehensive power component: (16) This item represents the equivalent active input power of the DC bus.

[0049] To characterize the capacity coupling relationship between multiple ports, the normalized ellipsoidal constraint of the device in the three-dimensional power space of DC active power – parallel reactive power – series injection is defined as: (17) S: Total apparent capacity of the equipment; Rated capacity of parallel converters; Rated capacity of series converter; Reactive power at parallel ports; : Series injection voltage; Line current.

[0050] This constraint geometrically defines the joint feasible power domain of the UPQC-photovoltaic-storage integrated device. Power Domain ensures that the multi-port power output point is always located within the ellipsoid.

[0051] (1) Coupling constraint: Instead of treating active power, reactive power, and injected voltage independently, they are enclosed by an ellipsoid. The overall constraints on their combination relationship ensure dynamic coordination among the three within the total capacity.

[0052] (2) Physical interpretability: The major and minor axes of the ellipsoid are determined by... The values ​​represent the total capacity of the device, the capacity of the parallel branches, and the injection capability in the series, respectively. Their shapes reflect the topology parameters and operating characteristics of the device, and all points within the envelope are physically realizable.

[0053] (3) Safety Boundary: When the load on any port increases, the available capacity of other ports automatically shrinks, thereby... Prevent overload and thermal imbalance of components to ensure safe operation of the system under multiple operating conditions.

[0054] (4) Dynamic coordination: In scenarios involving voltage dips or load fluctuations, energy storage can... adjust The parallel and series ports respectively undertake the tasks of reactive power and voltage support, and the power changes of the three are automatically constrained by the ellipsoidal constraint to achieve fast and stable multi-port dynamic balance.

[0055] Step Six: Run the set of constraints In embodiments of the present invention, to ensure the safe, stable, and economical operation of the UPQC-integrated photovoltaic-storage device and the distribution network, in addition to the comprehensive optimization objectives and capacity ellipsoid constraints, the following set of operational constraints is established. This set of constraints covers system power flow, voltage safety, port capacity, energy storage conservation, and power quality requirements, ensuring the physical feasibility and engineering feasibility of the model.

[0056] 1. Power flow balance and voltage safety constraints Distribution network nodes should satisfy the balance between active and reactive power: (18).

[0057] At the same time, the node voltage and branch current should satisfy: (19) in, This constraint ensures power flow balance in the system and prevents node overvoltage or line overload.

[0058] 2. DC-side energy conservation constraint The DC bus power flow of the UPQC-integrated photovoltaic-storage device satisfies the energy conservation principle: (20) in, For photovoltaic port output power, This represents the energy storage power (positive for discharging and negative for charging). To improve the energy transfer efficiency of parallel and series converters, This constraint accounts for system losses. It ensures a balanced energy flow within the device, providing a basis for coordinated multi-port operation.

[0059] 3. Parallel compensation port capacity constraints During the reactive power support and harmonic suppression process, the active and reactive power of the parallel compensation port should be limited to the rated capacity range. (twenty one) in The rated apparent capacity for the parallel ports. This constraint ensures that the parallel converters operate within a safe power range, preventing overload.

[0060] 4. Series compensation port voltage and capacity constraints The injection voltage and capacity of the series compensation port should meet the following requirements: (twenty two) in, For injection voltage, For line current, This indicates that the injected voltage amplitude does not exceed 20% of the line's rated voltage. This constraint limits the amplitude and capacity of series compensation during voltage dip support, ensuring device insulation and thermal safety.

[0061] 5. Dynamics of Energy Storage State of Charge (SOC) and Emergency Floor Constraints The state-of-charge evolution equation of the energy storage system is: (twenty three) And should satisfy: (twenty four) Where E is the rated energy capacity of the energy storage. For charging and discharging efficiency, For the energy storage emergency floor (typically 10%), this constraint ensures the conservation of energy storage and preserves the energy supporting critical loads during short-term power outages.

[0062] 6. Mutual exclusion of energy storage charging and discharging and power smoothing constraints To prevent energy storage systems from charging and discharging simultaneously, mutual exclusion conditions are set: (25).

[0063] To suppress power fluctuations and reduce cycling losses, a smoothing constraint is set: (26) This constraint ensures continuous and stable energy storage output, improving the quality of the system's dynamic response.

[0064] 7. Photovoltaic output and constraints on power generation The output power of the photovoltaic port is limited by irradiance and inverter capacity. (27).

[0065] Light curtailment is permitted when system capacity or voltage constraints are triggered: (28) By explicitly introducing the curtailment variable, the energy loss caused by photovoltaic reduction can be quantitatively reflected in the optimization process.

[0066] 8. Power Quality and Voltage Ride-Through Constraints To ensure power quality for different types of loads, node voltage deviation and harmonic distortion should meet the following requirements: (29) The limits are set at 2%-7% and 3%-8% depending on the load level.

[0067] Critical loads must meet ride-through constraints during voltage dips: (30).

[0068] To describe the harmonic suppression effect of the UPQC-integrated photovoltaic-storage device, a harmonic voltage sensitivity matrix is ​​introduced: (31) in, Let be the voltage sensitivity of node i to device n under the ℎth harmonic. This constraint is used to linearize the response of node THD to harmonic injection current, facilitating its co-solution with capacity constraints in optimization.

[0069] Through the above set of operational constraints, this invention establishes a complete operational boundary for the distribution network and the UPQC-photovoltaic-storage integrated device at the physical level, ensuring that the system operation always meets the requirements of power flow balance, safety limits and dynamic feasibility during power quality management, active and reactive power joint regulation and energy storage dispatch.

[0070] Step 7: Two-level optimization solution In an embodiment of the present invention, to balance the global optimal configuration and real-time optimal operation of the UPQC-integrated photovoltaic-storage device, a two-layer optimization solution mechanism integrating planning and operation is proposed. This mechanism consists of upper-layer site selection and capacity optimization and lower-layer rolling jointness optimization, achieving synergy between long-term investment decisions and short-term operational control through information transmission and iterative updates. The specific implementation process is shown in Figure 2.

[0071] 1. Upper-level optimization (planning layer) The goal of upper-level optimization is to determine the optimal installation nodes and capacity configuration of the UPQC-photovoltaic-storage integrated device, so that the system can achieve the overall optimization of investment and operating costs while meeting power quality and reliability requirements.

[0072] Decision variables: ∈ {0,1}: Whether the nth candidate node has a device installed; : Rated apparent power of the device; Rated energy capacity of energy storage.

[0073] Upper-level optimization objective function (32) Where: Item 1 represents the equipment investment and fixed-capacity cost; Item 2 represents the expected operating cost under different typical operating scenarios. This is a set of scenarios (including typical day, photovoltaic fluctuation ±10%, load fluctuation ±5%, etc.). The total intraday cost returned for optimizing the lower-level operation.

[0074] Constraints (33) Budget refers to the upper limit of the investment budget. Power quality and voltage ride-through constraints.

[0075] 2. Lower-level optimization (runtime layer) The goal of lower-level optimization is to optimize the active-reactive power distribution, harmonic compensation, and energy storage SOC trajectory of the devices in real time, given the device location and capacity configuration at the upper level, by combining the rolling forecast information of photovoltaics and loads, so that the system can maintain the optimal balance between power quality and economy under dynamic conditions.

[0076] Run the target function The lower layer employs a Model Predictive Control (MPC) framework to minimize runtime costs within the rolling prediction window H. (34) Among them, control variables It includes quantities such as reactive power output at parallel ports, series injection voltage, energy storage charging and discharging power, and photovoltaic output regulation. The optimization objective is the same as in step four, but only the short-term operation within the rolling time window is solved.

[0077] Constraints In the lower-level operation optimization, all of the following constraints must be met; (1) Node power flow balance and voltage and current safety boundaries; (2) DC energy conservation and port capacity constraints; (3) Capability ellipsoid constraint; (4) Dynamic equations of energy storage SOC and emergency floor constraints; (5) Photovoltaic power output and curtailment constraints; (6) Multi-priority power quality and voltage ride-through constraints.

[0078] Therefore, the lower-level optimization problem can be formalized as: (35) in Let represent the sets of equality and inequality constraints, respectively.

[0079] Rolling prediction and execution mechanism (1) At the current moment To obtain photovoltaic and load forecast data for the next H steps; (2) Construct the corresponding SOCP optimization model and obtain the optimal control sequence. ; (3) Only execute the first step of the control quantity. The system enters the next moment; (4) Update the measurement and prediction data, rebuild the optimization model, and repeat the process.

[0080] Upper and lower layer collaboration (1) Data exchange between the upper and lower layers is achieved through runtime costs and configuration parameters; (2) Location, capacity and energy storage scale of upper output device ; (3) The lower layer performs rolling optimization under a given configuration and feeds back the average running cost. To the upper level; (4) The upper layer iteratively updates the site selection and capacity setting scheme based on feedback to form an investment-performance balance solution.

[0081] Through multiple interactive iterations, the upper and lower layers achieve convergence and consistency, ensuring that the configuration scheme balances global economy and local feasibility. Ultimately, in engineering applications, the runtime control executes the lower-level MPC strategy in real time, while the upper-level configuration results are fixed and used to guide long-term scheduling.

[0082] Example verification and implementation effect description 1. Case Study System and Data The IEEE 33-node distribution network is adopted, with a base voltage of 12.66 kV. A typical 24-hour day is discretized into 96 time periods of 15 minutes each. The typical daily active / reactive power curves for each node (evening peak 19:00-21:00) show a total daily peak of approximately 3.7 MW / 2.3 Mvar. Distributed photovoltaic (PV) grid connection points are deployed at nodes 6, 18, and 30, with output based on sunny weather conditions. Node 18 (hospital campus) is designated as a critical load, node 25 (industrial) as an important load, and the rest as general loads. Power quality limits are set for critical loads (±2%, THD≤3%), important loads (±5%, THD≤5%), and general loads (±7%, THD≤8%). The operational layer MPC prediction window H=4 (i.e., 1 hour) is rolled over every 15 minutes.

[0083] 2. Candidate Access Points and Cost Calculation Candidate access: Nodes {6, 18, 24, 30}; Cost (equivalent annualized): UPQC rated apparent capacity S: 0.9 million yuan / kVA; Energy storage power Price: 0.4 million yuan / kW; Energy storage capacity E: 0.2 million yuan / kWh; Photovoltaic 0.35 million yuan / kW; Budget ceiling: 500,000 yuan; Energy storage efficiency: =0.95, =0.95, Emergency Floor SOC floor = 10%; Maximum tandem injection ratio: = 0.2; Parallel / Series Port Efficiency = = 0.97.

[0084] 3. Upper-level optimization of site selection and sizing results (representative solution from the Pareto front) Under a multi-scenario set Ω = {typical day, PV ±10%, load ±5%}, NSGA-II + lower-level SOCP-MPC evaluation was used to obtain three representative solutions. For ease of engineering selection, a "capacity-budget-effect" comparison is provided:

[0085] Here, A represents economic priority, B represents balanced recommendation, and C represents quality priority, with the annualized investment of option C exceeding the budget.

[0086] Figure 3 shows the investment-performance Pareto front considering hard constraints on voltage and harmonics. As can be seen from the figure, the full... Feasible solutions with sufficient hard constraints form a clear "frontier zone," with an investment amount of 45–55 × 10⁻⁶. 4 Daily operating costs can be minimized within the CNY range. The selected solution's corresponding point is marked on the graph, ensuring 100% compliance of critical loads while reducing operating energy consumption by approximately 8%. This demonstrates that the proposed site selection, capacity determination, and coordinated control method achieves a good trade-off and optimal configuration between overall economic efficiency and power quality performance. Under a budget constraint of 500,000 CNY, Scheme B falls at the Pareto front inflection point, showing significant improvement in power quality. With investment approaching the upper limit, it offers the best overall cost-effectiveness and is therefore the recommended configuration.

[0087] The capacity parameters of Scheme B (used for solving the capacity ellipsoid constraints in (5) and the operational constraints in (6)): Total capacity S = 150 kVA, parallel port = 120kVA, serial port = 90kVA; PV power supply 80kW (distribution: N6 30 kW, N18 20kW, N30 30 kW); energy storage PCS 100kW, battery 200kWh, SOC min = 10%, SOC max = 90%.

[0088] 4. Optimization effect of scrolling in the lower running layer (Solution B) Figure 4 shows the variation curve of the minimum voltage of the entire grid under typical daily operation of the UPQC-PV-storage integrated device of the present invention. As can be seen from the figure, under the baseline scheme, affected by PV power fluctuations and load disturbances, the voltage at point PCC drops significantly at 10–11h and 8–18.5h, with the lowest voltage being approximately 0.93 pu. However, the scheme of the present invention effectively suppresses the voltage drop through the synergistic effect of series-side voltage compensation and parallel-side reactive power support, keeping the minimum voltage of the entire grid above 0.95 pu, thus verifying the voltage support capability of the device under heavy voltage drop and strong harmonic conditions.

[0089] Figure 5 shows the reactive power output |Q| support curves of the two parallel ports (nodes 18 and 30) under a typical day. As can be seen from the figure, both ports achieve significant reactive power output improvements during the 10–11h and 18–18.5h periods, with maximum support capacities reaching 85% and 75% of the rated capacity, respectively. This indicates that the device can dynamically allocate reactive power support based on node voltage deviations, achieving zoned voltage coordinated control. The total reactive power support of the parallel ports is concentrated during disturbance periods and automatically returns to zero power during stable periods, demonstrating the adaptability and high responsiveness of the control strategy.

[0090] Figure 6 compares the harmonic suppression performance of the baseline scheme and the device of this invention. This figure illustrates the performance comparison of the UPQC-integrated photovoltaic-storage device of this invention and the baseline scheme in terms of harmonic mitigation. As can be seen from the figure, the PCC voltage THD decreased from 7.2% to 2.8%, the parallel-side current THD decreased from 6.8% to 2.5%, the critical load voltage THD decreased from 5.6% to 2.3%, and the average THD of the entire network decreased from 6.1% to 2.6%, with an overall harmonic content reduction of approximately 60%. This result demonstrates that this invention achieves significant harmonic suppression while maintaining voltage, exhibiting excellent comprehensive power quality management performance.

[0091] Figure 7 shows the weighted sliding window compliance curves for multi-priority loads (vertical axis range 0–110%). As can be seen from the figure, during the disturbance periods of 10–11h and 18–18.5h, the compliance rate of critical loads under the baseline scheme is approximately 85%, important loads are approximately 92–95%, and general loads are approximately 97–99%; while the compliance rate of all three types of loads in the proposed scheme (solid line) remains at 100%. This indicates that the UPQC-photovoltaic-storage coordinated control method proposed in this invention can simultaneously meet voltage deviation and harmonic limit constraints, achieving hierarchical power quality assurance under multi-priority load distribution conditions.

[0092] Figure 8 shows the 24-hour state of charge (SOC) and charge / discharge power curves of the energy storage system. As can be seen from the figure, the energy storage system is in a charging state during the peak photovoltaic power generation period (10–15h) during the day, with the SOC rising from 60% to approximately 80%. During the high load period in the evening (18–21h), the energy storage system discharges to support the grid, and the SOC drops to around 60%. The charge / discharge power curves are smooth and without abrupt changes, indicating that the control strategy of this invention can achieve orderly energy allocation while ensuring SOC constraints (0.3–0.9), supporting both voltage and DC link stability, demonstrating high coordination in energy management.

[0093] 5. Check the effectiveness of the capability ellipsoid constraint. Envelope utilization: During the evening peak hours of 18:00-21:00, the mean radial ellipsoid is 0.86, and the peak value is 0.96; Port coupling: When Optimize automatic contraction when approaching 0.9 seconds. , To avoid overload; Safety: No control reversal or DC link drop caused by individual port saturation was observed, verifying the necessity and sufficiency of the "joint feasible domain" description.

[0094] Within a budget of 500,000 yuan, Solution B (18+30 nodes, S=150 kVA, PV 80 kW, BESS 100 kW / 200kWh) achieves: voltage deviation at key nodes ≤±1.6%, THD reduced to 2.9%, and voltage ride-through 100% compliance; daily network loss reduced by 8.2%, grid connection ramping under control, 1.9% SOC for curtailment meeting emergency floor requirements, and limited equivalent lifetime consumption; no overstepping of capacity ellipsoid constraints throughout the day, and automatic coordination of port power according to envelope.

[0095] This invention proposes a photovoltaic-storage integrated UPQC (Up-Level Quality Control) site selection and capacity determination method that combines capacity ellipsoid constraints and multi-priority load optimization for distribution networks with high-penetration renewable energy and multiple types of power electronic loads. This method offers the following advantages: (1) Significantly improves the rationality and accuracy of site selection and capacity determination. This invention introduces a capacity ellipsoid constraint model, which unifies the active, reactive and voltage injection capabilities of the UPQC-photovoltaic-storage integrated device in the form of an ellipsoid, so as to realize the accurate expression of the device's power capability and the consistency of physical constraints, avoid the capacity distortion and infeasible solution problems caused by traditional separate constraints, and make the site selection and capacity determination calculation results more consistent with the actual operating capability of the equipment.

[0096] (2) Achieve differentiated configuration for multi-priority loads. By establishing a multi-priority load weighted compliance index, the power quality requirements of critical loads, important loads and general loads are considered in the site selection and capacity determination stage, so as to achieve accurate matching between device capacity and load importance, and significantly improve the power supply guarantee level and configuration targeting of critical loads.

[0097] (3) Improve capacity utilization and investment efficiency. This invention determines both device location and capacity configuration based on the SOCP optimization framework, achieving coordinated resource allocation across the entire network. Optimization results show that device capacity utilization is increased by 10%–15%, reducing system investment and operation and maintenance costs while ensuring power quality, achieving optimal technical performance and economic benefits.

Claims

1. A method for multi-priority load capacity ellipsoidal-constrained site selection and capacity determination in integrated photovoltaic-storage UPQC, characterized in that: S1. Multi-priority load differentiation constraints: loads are divided into three categories: critical, important and general, and weighted compliance rate indicators and graded power quality constraints are introduced. S11. Voltage and Harmonic Distortion Constraints: The voltage deviation and harmonic distortion level of various load nodes at any time t should meet the following constraints: （10） in, This refers to the relative deviation of the node voltage. The total harmonic distortion rate of the node; The limit values ​​for different load categories are as follows: Critical load: ; Important loads: ; Typical load: ; S12, Voltage Dip Ride-Through Constraint: Introduce voltage ride-through constraints for critical load nodes: （11） in, In order to cross the lower limit, This refers to the duration of the voltage drop. S13, Introduce the node harmonic voltage sensitivity matrix Its definition is as follows: （12） in, Let i be the voltage offset of node i under the ℎth harmonic. Harmonics of the nth UPQC device Current injection amount, This represents the harmonic sensitivity of node i to the current injected into device n; S2. Construct a multi-objective optimization function: （13） in: This refers to the node priority weight coefficient. Branch resistance; These are the device's rated apparent power and energy storage capacity, respectively; C S C E This is the unit capacity cost coefficient. This is the lifetime degradation coefficient; In equation (13): The first item represents the weighted sum of squares of voltage deviation, which is used to measure the system voltage quality; The second item represents the weighted sum of squares of harmonic distortion, reflecting the quality of the electrical waveform. The third item represents line power loss, which reflects the economic efficiency of distribution network operation; Item 4 is the investment and energy storage capacity cost of the UPQC-integrated photovoltaic and energy storage device; Item 5 is the cost of energy storage battery life loss; S3. Establishment of the constraint model: The DC-side power balance of the UPQC-photovoltaic-storage integrated device at any time t is: （15） in: Photovoltaic port power; Energy storage port power; : Active power on the AC side of the parallel port; : Active power on the AC side of the series port; : Converter efficiency; System losses; Introducing the composite power component: （16） The normalized ellipsoidal constraint for the device in the three-dimensional power space of DC active power injection – parallel reactive power injection – series injection is defined as follows: （17） Where: S is the total apparent capacity of the device; This refers to the rated capacity of the parallel converter; This refers to the rated capacity of the series converter; The reactive power of the parallel port; For series injection voltage; Line current; S4. Set of runtime constraints: S41. Power flow balance and voltage safety constraints: （18） At the same time, the node voltage and branch current should satisfy: （19） S42, DC-side energy conservation constraint （20） in, For photovoltaic port output power, For energy storage power, To improve the energy transfer efficiency of parallel and series converters, For system losses; S43, Parallel compensation port capacity constraint （21） in Rated apparent capacity for parallel ports; S44, Series Compensation Port Voltage and Capacity Constraints （22） in, For injection voltage, For line current, Indicates the magnitude of the injected voltage; S45, Energy Storage State of Charge Dynamics and Emergency Floor Constraints （23） And should satisfy: （24） Where E is the rated energy capacity of the energy storage. For charge and discharge efficiency, SOC floor For energy storage and emergency flooring; S46. Energy storage charging and discharging mutual exclusion and power smoothing constraints: Set mutual exclusion conditions: （25） Set smoothing constraints: （26） S47, Photovoltaic Output and Power Constraints The output power of the photovoltaic port is limited by irradiance and inverter capacity. （27） Light curtailment is permitted when system capacity or voltage constraints are triggered: （28） S48. Power Quality and Voltage Ride-Through Constraints Node voltage deviation and harmonic distortion should meet the following requirements: （29） Critical loads must meet ride-through constraints during voltage dips: （30） Introducing the harmonic voltage sensitivity matrix: （31） in, Let i be the voltage sensitivity of device n to node i under the ℎth harmonic; S5, Two-level optimization solution S51, Upper-level optimization, i.e., planning layer: S511, Decision Variables ∈ {0,1}: Whether the nth candidate node has a device installed; : Rated apparent power of the device; Rated energy capacity of energy storage; S512, Upper-level optimization objective function （32） in: The first item represents the equipment investment and fixed-capacity cost; the second item represents the expected operating cost under different typical operating scenarios. For scene sets; The overall intraday cost returned for optimizing lower-level operations; S513, Constraints （33） Budget refers to the upper limit of the investment budget; S52, Lower-level optimization, i.e., runtime layer: S521. Run the target function: （34） Among them, control variables It includes quantities such as parallel port reactive power output, series injection voltage, energy storage charging and discharging power, and photovoltaic output regulation; S522, Constraints In the lower-level operation optimization, all of the following constraints must be met: (1) Node power flow balance and voltage and current safety boundaries; (2) DC energy conservation and port capacity constraints; (3) Capability ellipsoid constraint; (4) Dynamic equations of energy storage SOC and emergency floor constraints; (5) Photovoltaic power output and curtailment constraints; (6) Multi-priority power quality and voltage ride-through constraints; S523, The lower-level optimization problem is formalized as follows: （35） in Let these represent the sets of equality and inequality constraints, respectively. S523, Rolling Prediction and Execution Mechanism (1) At the current moment To obtain photovoltaic and load forecast data for the next H steps; (2) Construct the corresponding SOCP optimization model and obtain the optimal control sequence. ; (3) Only execute the first step of the control quantity. The system enters the next moment; (4) Update the measurement and forecast data, rebuild the optimization model, and repeat the process. S524, Upper and Lower Layer Collaboration Relationship (1) Data exchange between the upper and lower layers is achieved through runtime costs and configuration parameters; (2) Location, capacity and energy storage scale of upper-level output device ; (3) The lower layer performs rolling optimization under a given configuration and feeds back the average running cost. To the upper level; (4) The upper layer iteratively updates the site selection and capacity setting scheme based on feedback to form an investment-performance balance solution.

Images