A method for calculating a power supply reliability index of a highway self-consistent energy system

Abstract

A kind of method for calculating the power supply reliability index of highway self-consistent energy system, comprising: establishing highway self-consistent energy system;Real-time power model of supply side of highway self-consistent energy system and traffic load model of demand side are constructed: wind turbine generator power-speed conversion formula is used to construct wind turbine generator power model, power conversion formula is used to construct photovoltaic unit power generation power model, traffic load model is constructed by using normal distribution scene;Reliability parameter model of each component of highway self-consistent energy system is established;The power supply reliability index is calculated using three-state sequential Monte Carlo.This application not only considers from the application object, but also reflects the difference between power supply elements and other elements from the algorithm itself, considers the type of multiple faults, improves the limitation that traditional sequential Monte Carlo algorithm does not distinguish power supply fault from switch element and other element fault.

Description

Technical Field

[0001] This invention relates to a calculation method, and more particularly to a method for calculating a power supply reliability index, a method for calculating a power supply reliability index for a self-consistent energy system for highways during the planning stage, and its application. Background Technology

[0002] In calculating the power supply reliability index of a power system, the traditional sequential Monte Carlo algorithm typically uses fixed historical data on both the supply and demand sides of the research object, without considering the uncertainties on both sides.

[0003] Furthermore, regarding the algorithm itself, the traditional sequential Monte Carlo algorithm treats components as having only two states: normal operation and fault. It analyzes the relationship between power supply and load only during the normal operation time T1 and the fault repair time T2, assuming that any fault, whether in the power supply or other switching components, will only result in a shutdown, instantly losing its operational capability. For example, in their paper "Reliability Analysis of Multi-Source DC Distribution Network Based on Sequential Monte Carlo Simulation" published in the journal *Distribution Technology*, authors Guo Haomin and Liu Sai argue that when a fault occurs, regardless of the type of fault, the component will instantly shut down.

[0004] Traditional sequential Monte Carlo algorithms do not consider the diversity of fault modes in faulty components, assuming that the power source and other components have the same fault state. However, in reality, wind and solar power plants typically experience a period of operation with the fault after most common faults occur. Furthermore, minor faults in wind and solar generators do not manifest as a complete shutdown or instantaneous loss of power generation; rather, they result in a weakening of power generation capacity. During this period, to ensure economic efficiency, personnel will analyze the cause of the fault without shutting down the plant. Only after identifying the problem will maintenance and repairs be carried out.

[0005] According to Article 18 of the Wind Turbine Maintenance Management Regulations (Version A) issued by China Huaneng Group New Energy Development Co., Ltd., the turbine can only be shut down for maintenance when major auxiliary equipment or auxiliary equipment malfunctions. Malfunctions include: abnormal noises from the rotor, inability to steer or turn, inconsistent rotor speed, and rotor rotation without generator output. However, if the wind turbine exhibits a malfunction where the wind speed reaches the rated wind speed but the rotor speed is below the rated speed and the generator cannot output the rated voltage, the turbine's power generation capacity is reduced. In this case, inspection and troubleshooting can be performed, and immediate shutdown is not necessary.

[0006] Similarly, according to the National Energy Information Platform, photovoltaic (PV) system malfunctions are categorized into two types: severe malfunctions, such as overheating or abnormal bus voltage, which will immediately shut down the inverter connected to the PV unit, requiring maintenance; and general malfunctions, such as fan failures, which do not significantly impact personal safety or inverter safety. In these cases, the PV unit's output power will only decrease, and immediate shutdown is not necessary, as the time required from detection to repair can save costs. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention discloses a three-state sequential Monte Carlo calculation method for power supply reliability indicators, and particularly discloses a three-state sequential Monte Carlo calculation method for power supply reliability indicators in highway self-consistent energy system planning. The calculation scheme is as follows:

[0008] A method for calculating the power supply reliability index of a self-consistent energy system for highways, characterized by:

[0009] Step 1: The specific structure of establishing a self-sufficient energy system for highways includes: a supply side, a demand side, and a microgrid; the supply side consists of wind, solar, and energy storage power generation systems; the demand side is mainly based on the demand of highways and is connected through the intermediate medium of microgrids.

[0010] Step 2: Constructing a real-time power model for the supply side and a traffic load model for the demand side of the highway self-consistent energy system: Using different probability density distributions, the wind speed scenario for wind turbines, the solar irradiance scenario for photovoltaic units, and the traffic load demand scenario for highways are determined through sampling using the three-state sequential Monte Carlo method. A wind turbine power generation model is constructed using the power-wind speed conversion formula, and a photovoltaic power generation model is constructed using the power conversion formula. Analysis of historical traffic load data suggests that the historical load data can be fitted to a normal distribution, and the load values ​​at the same time on different days in each quarter of the year are not significantly different, thus allowing the same set of normal distributions to be used for description. Therefore, this invention adopts a typical load day fitting method based on normal distribution to approximate the load scenarios of the four quarters of the year. It is assumed that the 24-hour load of a highway approximately follows 24 different normal distributions, and the expected value of the hourly normal distribution within 24 hours is obtained through sampling as the traffic load demand value at that moment, serving as the demand-side power model.

[0011] Step 3: Establish reliability parameter models for each component of the self-consistent energy system of the highway;

[0012] Step 4: Calculate power supply reliability indices using three-state sequential Monte Carlo methods.

[0013] This invention discloses a self-consistent energy system for highways, characterized in that: the system includes a non-volatile storage medium, the non-volatile storage medium including a stored program, wherein the program, when running, controls the device where the non-volatile storage medium is located to execute the above-described method.

[0014] The present invention discloses an electronic device, characterized in that it includes a processor and a memory; the memory stores computer-readable instructions, and the processor is used to execute the computer-readable instructions, wherein the computer-readable instructions execute the method described above.

[0015] Beneficial effects

[0016] This invention, when calculating the power supply reliability index of a self-consistent energy system for highways using sequential Monte Carlo methods, fully considers the uncertainties on both the supply and demand sides when establishing the power generation model on the supply side and the load model on the demand side. It constructs probability density distributions on both the supply and demand sides and uses Monte Carlo sampling to simulate wind and solar load scenarios, rather than using fixed historical data, thus reflecting the uncertainties on both the supply and demand sides.

[0017] From the perspective of the three-state sequential Monte Carlo algorithm itself, the traditional sequential Monte Carlo algorithm only considers two states for components in its computational flow: normal operation and fault. It analyzes the relationship between power supply and load only during the normal operation time T1 and the fault repair time T2, assuming that all types of components, whether power supplies or other switching components, will only experience shutdown faults when a fault occurs, which does not completely match the actual operation of a power plant. The three-state sequential Monte Carlo algorithm proposed in this invention includes three states in its simulation time: normal operation time T1 + fault-prone operation time T2 + fault equipment repair time T3. A fault-prone operation time T2 is added to distinguish the fault states of power supply components from those of ordinary switching components. A random number x is used... j By comparing the failure probability, a failure diagnosis time T2 and a repair time T3 were determined. The photovoltaic, wind turbine, and energy storage systems on the supply side were artificially set to exhibit a reduction in power generation capacity to 60% of their normal power generation capacity (normal power calculated from real-time wind speed / solar irradiance) within the failure diagnosis time T2. Maintenance was only performed on the corresponding faulty equipment within the repair time T3, ensuring economic efficiency in actual operation. Only ordinary switching components without power generation capacity consistently exhibited a shutdown failure within both the failure diagnosis time T2 and the repair time T3.

[0018] In summary, the three-state sequential Monte Carlo algorithm adopted in this invention not only considers the application objects more comprehensively and specifically, but also reflects the difference between power supply components and other types of components in the algorithm itself. Taking into account various types of faults, it improves the limitation of traditional sequential Monte Carlo algorithms that do not distinguish between power supply faults and faults of other components such as switching components, thus ensuring the economic concept in actual operation. Attached Figure Description

[0019] Figure 1 A bar chart showing the Weibull probability density distribution of wind speed;

[0020] Figure 2 This is a schematic diagram of the probability distribution of beta irradiance.

[0021] Figure 3 This is a probability distribution diagram of beta irradiance.

[0022] Figure 4 This is a schematic diagram of the normal distribution of the load;

[0023] Figure 5 The simulation flowchart for a three-state sequential Monte Carlo simulation is shown below.

[0024] Figure 6 The active power output curve of the wind turbine generator;

[0025] Figure 7 A schematic diagram of a two-state model of a component;

[0026] Figure 8 Simplified structural diagram of energy storage device;

[0027] Figure 9 This is a state transition diagram for an energy storage device;

[0028] Figure 10 This is a schematic diagram of the state transition process of a microgrid under the Well-being theory.

[0029] Figure 11 A flowchart for the three-state sequential Monte Carlo calculation of power supply reliability indicators;

[0030] Figure 12 A structural block diagram for configuring a self-consistent energy system for highways;

[0031] Figure 13 A schematic diagram of the structure of a self-consistent energy system for a highway;

[0032] Figure 14 The flowchart for the two-state sequential Monte Carlo calculation of the power supply reliability index. Detailed Implementation

[0033] A method for calculating the power supply reliability index of a self-consistent energy system for highways, characterized by:

[0034] Step 1: The specific structure of establishing a self-sufficient energy system for highways includes: a supply side, a demand side, and a microgrid; the supply side consists of wind, solar, and energy storage power generation systems; the demand side is mainly based on the demand of highways and is connected through the intermediate medium of microgrids.

[0035] Step 2: Construct a real-time power model for the supply side and a traffic load model for the demand side of the self-consistent energy system of the highway: Using different probability density distributions, the wind speed scenario of the wind turbine, the solar irradiance scenario of the photovoltaic unit, and the traffic load demand scenario of the highway are determined by sampling using the three-state sequential Monte Carlo method; the power generation model of the wind turbine is constructed using the power-wind speed conversion formula of the wind turbine, the power generation model of the photovoltaic unit is constructed using the power conversion formula, and the traffic load model is constructed by using the scenario constructed with a normal distribution and taking the expected value of the normal distribution of each hour in 24 hours as the traffic load demand value at that moment.

[0036] Monte Carlo simulation is a computer stochastic simulation method based on probability and statistics theory. Monte Carlo simulation methods are divided into non-sequential and sequential Monte Carlo simulation methods. When assessing the reliability of power systems, sequential Monte Carlo simulation methods are generally performed on an hourly basis, enabling the simulation of system states over a certain time span.

[0037] Before using the three-state sequential Carlow simulation method to calculate the power supply reliability index of the highway self-consistent energy system, it is necessary to first construct the power generation side of the highway self-consistent energy system: power models of wind turbines, photovoltaic units, and energy storage systems, as well as the load model of the traffic side, and the reliability models of each component of the highway self-consistent energy system.

[0038] 1. Real-time power model on the supply side

[0039] (1) Wind power model

[0040] The power curve of a wind turbine is related to changes in wind speed. Currently, the most widely used method is to approximate the output power of a wind turbine using a quadratic function, such as... Figure 5-6 As shown.

[0041] Given the wind speed v at a certain moment t The output power of a single wind turbine can be calculated using the power conversion formula, where the conversion formula is:

[0042]

[0043] In the formula: P r P represents the rated power of a single fan. tv is the power of the wind turbine at time t; r v ci v co These represent the rated wind speed, cut-in wind speed, and cut-out wind speed of the fan, respectively; the values ​​of A, B, and C depend on v. r and v ci Size.

[0044]

[0045] (2) Photovoltaic power model

[0046] The output power of a photovoltaic (PV) unit is related to various factors such as solar irradiance, ambient temperature, PV panel tilt angle, and PV unit conversion efficiency. Among these factors, solar irradiance has the greatest impact on PV output power. Therefore, when constructing the power model in this patent, only the impact of solar irradiance on the output power of the PV unit is considered.

[0047] When the irradiance G at a certain moment is known t The output power of a photovoltaic unit can be calculated using the power conversion formula, which is:

[0048]

[0049] In the formula: P t P represents the power of the photovoltaic unit at time t. m G is the rated power of the photovoltaic unit; std The irradiance given for a standard environment is typically taken as 1 kW / m²; R c This is a specific irradiance, typically taken as 0.15 kW / m².

[0050] (3) Energy storage charging / discharging model

[0051] The working principle of an energy storage system in a microgrid is as follows: when the power generation supply from wind and solar power exceeds the load demand, the energy storage system stores electrical energy; when the power generation supply from wind and solar power is less than the load demand, the energy storage system releases the electrical energy. There are many types of energy storage systems, including supercapacitor energy storage, battery energy storage, electrochemical energy storage, compressed air energy storage, etc. This invention uses a battery as the energy storage system.

[0052] Let B(t) be the electrical energy stored in the energy storage system at time t, and let P be the charging / discharging power. B (t), then the charging / discharging time series of the energy storage system is:

[0053] B(t+1)=B(t)+P B (t) (3-4)

[0054] The power limit of the energy storage device during charging / discharging is as follows:

[0055]

[0056]

[0057] In the formula: P B A positive value for (t) indicates that the energy storage device is charging; P B A negative value for (t) indicates that the energy storage device is discharging; P disch-max P represents the maximum discharge power of the energy storage device. ch-max B is the maximum charging power of the energy storage device. max B represents the maximum capacity of the energy storage device. min This represents the minimum capacity of the energy storage device.

[0058] 2. Demand-side transportation energy load model

[0059] For the demand-side load curve of the self-consistent energy system of highways, after comprehensively analyzing the load data of the traffic demand side, it is assumed that the load values ​​at the same time on different days are not significantly different. Therefore, this invention adopts a typical load day fitting method based on normal distribution. It is assumed that the load side of the highway approximately follows a normal distribution, and the expected value of the normal distribution for each hour within 24 hours is considered to be the traffic load demand at that time.

[0060] The normal distribution model is a probability distribution model frequently used in engineering. Also known as the Gaussian distribution, its graph is called the normal curve. If the probability density function f(z) of the random variable Z is...

[0061]

[0062] In the formula: σ>0, and μ and σ are constants, called random variables, Z follows a function with parameters μ and σ. 2 The normal distribution is denoted as Z ~ N(μ, σ). 2 To fit a random variable to a normal distribution, it is only necessary to determine μ and σ. 2 That is, where μ is the expectation of the distribution, and σ 2 Let be the variance of this distribution.

[0063] The typical load day fitting method divides a day into 24 time periods, each lasting 1 hour. Then, using a large amount of grid load data, it fits the load data for the same time on different days into a normal distribution. Finally, the expected value of the normal distribution at each time moment is used as the load value for that moment on the typical day. The final fitted typical day is obtained by calculating the load values ​​for the 24 time periods.

[0064] After setting the normal distribution function for the i-th time period within 24 hours using formula (3-7), the expected load value for the i-th time period is calculated using the following formula.

[0065] E(x i)=∫x i f(x i )dx i (3-8)

[0066] In the formula: E(x) i ) represents the expectation for the i-th time period; x i Let f(x) be the independent variable representing the normal distribution function for the i-th time period. i Let represent the normal distribution function for the i-th time period.

[0067] Using this expected value as the load value at that moment, and calculating the expected values ​​for all moments, we then use these expected values ​​to fit a typical load day, i.e., E(x1), E(x2), ..., E(x... 24 ) is used as the load value for a typical load day that is fitted.

[0068] Step 3: Establish component reliability parameter models for each part of the self-consistent energy system of the highway, including reliability models for energy storage devices, wind turbines, and photovoltaic units.

[0069] This invention's reliability study posits that components in a power system exist in three states: normal operation, fault, and maintenance. It utilizes a traditional Markov model to simulate the operating states of components within the system, such as... Figure 7 As shown.

[0070] The average uptime of the component is T F The average repair time is T. R Assuming the state duration follows an exponential distribution, then T F and T R The expression is:

[0071]

[0072]

[0073] In the formula: x1 and x2 are random numbers uniformly distributed within (0, 1), which can be generated using the rand() function in Matlab; λ and μ represent the equivalent failure rate and equivalent repair rate of the component, respectively.

[0074] Reliability model of energy storage devices

[0075] The components in a microgrid can be divided into power supply components and non-power supply components. Power supply components include wind turbines, photovoltaic units, and energy storage systems, while non-power supply components mainly include circuit breakers, transformers, and fuses. Reliability data for the failure rate and repair rate of non-power supply components are obtained from relevant literature. For the reliability model study of wind turbines, photovoltaic units, and energy storage systems, this invention uses the state-space method to calculate their equivalent failure rate and repair rate.

[0076] The state-space method represents a system using its states and the possible transitions between those states, and then calculates the system's equivalent failure rate and recovery rate. The specific steps are: listing all possible system states; forming a state transition diagram and state transition matrix; determining the probability of a stable state based on the failure rate and recovery rate of each component in the system; and calculating the equivalent failure rate and equivalent recovery rate of the entire system.

[0077] In this invention, the energy storage system is a battery energy storage system. The reliability model for the energy storage device only considers a two-stage battery structure consisting of four parts connected in series: a battery bank, a DC / DC converter, a DC / AC inverter, and a grid-connected filter. Its structure is as follows: Figure 8 As shown in the diagram. In this structure, the battery discharge process involves the direct current being converted by a DC / DC converter and a DC / AC inverter, and then fed into the main power grid through a three-phase filter.

[0078] Let the failure rates of each part of the energy storage system be λ. C1 , λ C2 , λ C3 , λ C4 The repair rates were μ c1 μ c2 μ c3 and μ c4 The equivalent failure rate λ of the energy storage system is then... c for:

[0079] λ c =λ c1 +λ c2 +λ c3 +λ c4 (3-10)

[0080] The steps to solve for the equivalent repair rate of an energy storage device are as follows:

[0081] (1) Since the four components of the energy storage device are a series system, if any one part fails, the entire system will stop operating and enter a fault state. Considering only first-order faults, the state transition diagram of the energy storage device is formed, as follows: Figure 9 As shown.

[0082] Figure 9 In the diagram, the number 0 represents that the device is in normal operation, while the numbers 1, 2, 3, and 4 represent the device's fault status caused by the battery pack, DC / DC converter, DC / AC inverter, and filter, respectively.

[0083] (2) Form the state transition matrix A. A is:

[0084]

[0085] (3) Solve for the steady-state probability p of the energy storage device, p = [p0, p1, p2, p3, p4], where p0 is the normal state probability of the energy storage device, and p1, p2, p3, p4 are the failure probabilities of each part of the energy storage device. The system of equations for solving p is:

[0086]

[0087] (4) Based on the relationship between p0 and the equivalent failure rate λ and the equivalent repair rate μ, solve for the equivalent repair rate μ. c Where p0 and λ c μ c The relationship is as follows:

[0088]

[0089] The equivalent repair rate μ of the energy storage device c for:

[0090]

[0091] In summary, by reviewing literature and relevant materials, the failure rates λ of the four components of the energy storage device were obtained. C1 , λ C2 , λ C3 , λ C4 and repair rate μ c1 μ c2 μ c3 and μ c4 Then, using the energy storage device reliability model established in this section, and applying formulas (3-10) and (3-14), the equivalent failure rate λ of the energy storage device can be obtained. c (times / year) and equivalent repair rate μ c (times / hour)

[0092] Reliability model of wind turbine

[0093] This invention establishes a reliability model for wind turbine generators, considering only four components: the wind turbine, AC / DC rectifier, DC / AC inverter, and filter. Taking a two-stage permanent magnet direct-drive wind turbine generator as an example, a simplified wind turbine model is adopted. The wind turbine is connected to the microgrid bus after passing through the AC / DC rectifier, DC / AC inverter, and filter. By reviewing literature and relevant data, the failure rate and repair rate of each component are obtained, thus extending the energy storage device reliability model to wind turbine generators.

[0094] Let the failure rates of the four components of the wind turbine be λ. F1 , λ F2 , λ F3 , λ F4The repair rates were μ F1 μ F2 μ F3 and μ F4 The equivalent failure rate λ of the wind turbine unit F for:

[0095] λ F =λ F1 +λ F2 +λ F3 +λ F4 (3-15)

[0096] Equivalent repair rate μ of wind turbine F for:

[0097]

[0098] Reliability model of photovoltaic units

[0099] There are two types of photovoltaic (PV) generator topologies: two-stage and single-stage. In a two-stage structure, the PV array uses a DC / DC converter to increase or decrease the DC voltage to the voltage required for inversion, while simultaneously achieving maximum power point tracking (MPPT). Then, a DC / AC inverter converts the DC voltage to AC voltage, which is then filtered and fed into the microgrid bus.

[0100] In this invention, the photovoltaic (PV) unit adopts a two-stage structure, and the reliability model of the PV unit is established by considering only four parts: the PV array, the DC / DC converter, the DC / AC inverter, and the filter. The failure rate and repair rate of each part are obtained through literature and related data, and the reliability model of the energy storage device is extended to the PV unit.

[0101] Let the failure rates of the four components of the photovoltaic unit be λ. G1 , λ G2 , λ G3 , λ G4 The repair rates were μ G1 μ G2 μ G3 and μ G4 The equivalent failure rate λ of the photovoltaic unit G for:

[0102] λ G =λ G1 +λ G2 +λ G3 +λ G4 (3-17)

[0103] Equivalent repair rate μ of photovoltaic units G for:

[0104]

[0105] Step 4: Calculate power supply reliability indices using three-state sequential Monte Carlo methods.

[0106] The most significant characteristic of self-consistent energy systems for highways is the uncertainty on both the supply and demand sides. On the supply side, natural resources like wind and solar power are significantly affected by surrounding environmental and weather factors, resulting in intermittent and fluctuating power output. Simultaneously, traffic load on the demand side is also affected by weather and environmental factors, exhibiting uncertainty. Therefore, when calculating the power supply reliability index of a self-consistent energy system for highways, the uncertainties of wind, solar, and load must be considered, which can be achieved by constructing probability density distributions for each.

[0107] I. Analysis of the application of the Monte Carlo method.

[0108] (1) To use the Monte Carlo method, it is necessary to first construct a statistical experimental probability model that matches the actual problem by combining the actual physical properties of the problem itself, make appropriate adjustments to the model, and select appropriate probability model parameters.

[0109] (2) When sampling from a population with a known distribution, it is necessary to consider what method to use to appropriately sample from the population with a known distribution based on the distribution of each random variable in the model, and generate sufficient random numbers in the computer simulation. Since various probability models can be regarded as being composed of the probability distribution of relevant random variables, the general approach is to first generate random numbers that follow a uniform distribution, and then generate random numbers that follow a specific distribution based on actual cases, so as to continue the random simulation experiment.

[0110] (3) There is no exact formula to determine the optimal number of simulation scenes for different models. The scene reduction method is usually used, which involves selecting the minimum number of scenes when the simulation results remain basically stable through a large number of simulation experiments.

[0111] II. Simulating wind and solar load scenarios using Monte Carlo sampling

[0112] After summarizing the respective properties of wind, solar, and traffic load, this invention uses Weibull probability density distribution to construct wind speed, beta distribution to construct solar radiation intensity, and normal distribution to construct traffic demand-side load model.

[0113] Wind speed, solar irradiance, and system load can all satisfy Weibull, Beta, and normal distributions in h. 24 probability density distributions of wind speed, solar irradiance, and system load are set for 24 hours a day. Monte Carlo scene sampling is used to obtain wind and solar load scenes considering uncertain factors.

[0114] (1) The wind speed follows a Weibull probability distribution at any hour, the irradiance follows a beta distribution at any h, and the load follows a normal distribution at the same time on different days.

[0115] (2) Use MATLAB functions to simulate and generate wind speed, irradiance and load for multiple scenarios. (3) Under the condition of satisfying the distribution density function, the scenario can be reduced.

[0116] Probability density distribution of wind speed

[0117] Wind speed is a random phenomenon, and its distribution can be approximated. In this invention, the probability model for wind speed adopts the Weibull distribution. The Weibull probability density function is as follows:

[0118]

[0119] In the formula: c and k are the scale parameter (m / s) and shape parameter of the model, respectively; v is the wind speed (m / s). If the average wind speed v and the wind speed variance σ are known, the formulas for calculating parameters c and k can be obtained as follows:

[0120]

[0121]

[0122] In the formula: σ represents the average (m / s) and standard deviation (m / s) of the historical wind speed data, respectively.

[0123] Integrating equation (2-1), we obtain the probability distribution function F(v) for wind speed:

[0124]

[0125] Let U = F(v) and R = 1 - U. By taking the inverse function of equation (2-4), we can obtain the wind speed v. t for:

[0126]

[0127] In summary, after obtaining the parameters c and k of the Weibull distribution model, for any given time, a random number R belonging to (0, 1) can be generated using the rand() function in Matlab. Substituting this random number R into equation (2-5) yields the simulated wind speed value v. t .

[0128] When the number of sampled scenarios is set to 1000, the shape parameter k of the Weibull model for wind speed is 1.637; the scale parameter c is 5.218; the Weibull probability distribution of wind speed at this time is as follows. Figure 1 As shown.

[0129] Probability distribution of light radiation intensity

[0130] Factors influencing solar radiation intensity include clouds and shadows. The probability distribution of solar radiation intensity over one hour or several hours can be approximated by a Beta probability distribution. The Beta distribution is a set of continuous probability distributions defined on the interval [0,1], with two shape parameters α and β, and its probability density function is...

[0131]

[0132] In the formula: r is the solar irradiance at a certain moment in that period, W / m² 2 rmax is the maximum solar irradiance during that period, in W / m². 2 Γ(.) is the gamma function; α and β are shape parameters of the Beta distribution, and their changes will lead to changes in the shape of the probability density curve of the Beta distribution. α and β can be calculated based on the expected value μ and variance δ of the solar radiation intensity over a period of time.

[0133]

[0134]

[0135] When the number of sampled scenes is set to 1000, the shape parameters of the beta model for irradiance are: α = 0.6869, β = 2.1320. The probability distribution of the beta irradiance at this time is shown in the figure. Figure 3 As shown.

[0136] Probability distribution of the load model

[0137] Regarding the demand-side load curve of a self-consistent energy system for highways, after comprehensively analyzing the load data on the traffic demand side, it is assumed that the load values ​​at the same time on different days are not significantly different. Therefore, this patent adopts a typical load day fitting method based on normal distribution. It is assumed that the load side of the highway approximately follows a normal distribution, and the expected value of the normal distribution for each hour within 24 hours is considered to be the traffic load demand at that time.

[0138] The normal distribution model is a probability distribution model frequently used in engineering. Also known as the Gaussian distribution, its graph is called the normal curve. If the probability density curve of the random variable Z is...

[0139]

[0140] In the formula: σ>0, and μ and σ are constants, called random variables, z follows a function with parameters μ and σ. 2 The normal distribution is denoted as Z ~ N(μ,σ). 2To fit a random variable to a normal distribution, it is only necessary to determine μ and σ. 2 That is, where μ is the expectation of the distribution, and σ 2 Let be the variance of this distribution.

[0141] The typical load day fitting method divides a day into 24 time periods, each lasting 1 hour. Then, using a large amount of grid load data, it fits the load data for the same time on different days into a normal distribution. Finally, the expected value of the normal distribution at each time moment is used as the load value for that moment on the typical day. The final fitted typical day is obtained by calculating the load values ​​for the 24 time periods.

[0142] When the number of sampled scenarios is set to 1000, the normal distribution of the load model has a mean μ = 5.9 * 1e³w and a variance σ² = 0.1μ. The normal distribution diagram of the load at this time is as follows: Figure 4 As shown.

[0143] Traditional sequential Monte Carlo simulation of faults is based on a two-state reliability model. It performs Monte Carlo sampling of the fault states and fault times of components during the simulation period, analyzes the relationship between power supply and load within the influence range of the component's normal operating time and fault states, accumulates relevant load parameters, and finally uses load reliability indices to evaluate system reliability. The simulation steps are as follows:

[0144] (1) Determine the equipment reliability indicators (failure rate, repair rate).

[0145] (2) After the simulation starts, calculate the normal working time T of each device. i And by setting certain rules, faulty equipment can be identified.

[0146] Generally, we set λ. i Let x be the failure rate of component i (i = 1, 2, ..., n), and there are n components in the system. Use the rand() function in Matlab to generate n random numbers x. i Calculate the normal operating time T of each component. i =-(ln x i ) / λ i Let T1 = min(T) i If T1 is the minimum normal operating time, then component i corresponding to T1 will fail.

[0147] (3) Determine the repair time for the faulty equipment.

[0148] Generate another random number x j Calculate the fault repair time Tμ of faulty component j. j , where Tμ j =–(ln x j ) / μ jμ j Let be the repair rate of faulty component j.

[0149] (4) Determine the load that is shut down due to equipment failure, the normal working time of the accumulated load, the number of shutdowns, the downtime and other parameters.

[0150] (5) The simulation time has been added to the Monte Carlo simulation time (minimum normal operation time and fault repair time). Determine whether the simulation threshold has been reached. If so, proceed to the next step. Otherwise, return to step (2) and proceed to the next sequential Monte Carlo simulation.

[0151] (6) Statistical load reliability index and system reliability index.

[0152] Traditional sequential Monte Carlo simulations assume that when any component fails—whether it's a power source, switching element, or circuit—it immediately loses its capability, entering a completely faulty open-circuit state. Furthermore, a single sequential Monte Carlo run only includes normal operation and fault repair time. In reality, any fault will have a period for fault diagnosis and analysis. Moreover, when wind and solar power generators experience minor faults, they don't necessarily exhibit a complete shutdown or instantaneous loss of power generation; rather, they show a weakening of power generation capacity.

[0153] According to Article 18 of the Wind Turbine Maintenance Management Regulations (Version A) issued by China Huaneng Group New Energy Development Co., Ltd., the turbine can only be shut down for maintenance when major auxiliary equipment or auxiliary equipment malfunctions. Malfunctions include: abnormal noises from the rotor, inability to steer or turn, inconsistent rotor speed, and rotor rotation without generator output. However, if the wind turbine exhibits a malfunction where the wind speed reaches the rated wind speed but the rotor speed is below the rated speed and the generator cannot output the rated voltage, the turbine's power generation capacity is reduced. In this case, inspection and troubleshooting can be performed, and immediate shutdown is not necessary.

[0154] Similarly, according to the National Energy Information Platform, photovoltaic (PV) system malfunctions are categorized into two types: severe malfunctions, such as overheating or abnormal bus voltage, which will immediately shut down the inverter connected to the PV unit, requiring maintenance; and general malfunctions, such as fan failures, which do not significantly impact personal safety or inverter safety. In these cases, the PV unit's output power will only decrease, and immediate shutdown is not necessary, as the time required from detection to repair can save costs.

[0155] Therefore, the three-state sequential Monte Carlo simulation time adopted in this invention includes three parts: normal operation time T1 + fault-prone operation time T2 + faulty equipment repair time T3. This is used to distinguish the fault state of the power supply from that of ordinary switching elements.

[0156] Traditional sequential Monte Carlo methods neglect the analysis of component types and do not distinguish between power supply failures and failures of other components such as switching components. In contrast, three-state sequential Monte Carlo methods can demonstrate the difference between power supply components and other types of components.

[0157] Therefore, the simulation steps for the three-state sequential Monte Carlo method are as follows:

[0158] (1) Determine the equipment reliability indicators (failure rate, repair rate).

[0159] (2) After the simulation starts, calculate the normal working time T of each device. i And by setting certain rules, faulty equipment can be identified.

[0160] Generally, we set λ. i Let x be the failure rate for i (i = 1, 2, ..., n). Use the rand() function in Matlab to generate n random numbers x. i Let n be the total number of components. Calculate the normal operating time T for each component. i =-(ln x i ) / λ i Let T1 = min(T) i If T1 is the minimum normal operating time, then component i corresponding to T1 will fail.

[0161] (3) The faulty operation time T2 + the faulty equipment repair time T3.

[0162] Use the rand() function in Matlab to generate a random integer x from 1 to n.

[0163] If 1 ≤ x ≤ a, it indicates that a component in the power supply system has failed, where a is the sum of the numbers of the wind turbine, photovoltaic unit, and energy storage system. If the failure occurs in any of the wind turbine, photovoltaic unit, or energy storage system, a random number x between 0 and 1 is generated accordingly. j Calculate the sum of the faulty component j's operating time T2 and the faulty equipment's repair time T3. μj T μj =-(lnx) j ) / μ j μ j The repair rate of faulty power supply j.

[0164] Among them, wind power, photovoltaic, and energy storage systems showed a reduction in power generation capacity to 60% of normal power generation capacity (normal power calculated from real-time wind speed / solar irradiance) within the fault diagnosis time T2, with T2 accounting for 2 / 3 of the T2 timeframe. μj Time, but only the repair time T3 = T for the corresponding faulty equipment. μj -T2 will be shut down for maintenance.

[0165] If a+1≤x≤n, it indicates that other switching elements, circuit breakers, transformers, etc. have failed: within the faulty operation time T2 + the faulty equipment repair time T3, they all manifest as faulty shutdown.

[0166] (4) Determine the load that is shut down due to equipment failure, the normal working time of the accumulated load, the number of shutdowns, the downtime and other parameters.

[0167] (5) The simulation time has been added to the Monte Carlo simulation time: (normal working time T1 + faulty running time T2 + faulty equipment repair time T3). Determine whether the simulation threshold has been reached. If yes, proceed to the next step; otherwise, return to step (2) and proceed to the next sequential Monte Carlo simulation.

[0168] (6) Statistical load reliability index and system reliability index.

[0169] Traditional sequential Monte Carlo methods neglect the analysis of component types, failing to distinguish between power supply failures and failures of other components such as switching elements, and also failing to consider the types of component failures. In contrast, the three-state sequential Monte Carlo method not only reflects the differences between power supply components and other component types but also takes into account various types of failures.

[0170] The Well-being theory will be applied to the calculation of the power supply reliability index of the self-consistent energy system of highways.

[0171] (1) Application of Well-being Theory in Microgrids

[0172] Well-being theory divides a power generation system into three states: healthy state, boundary state, and risk state; and uses probability to represent each state.

[0173] At any given moment, a healthy state indicates that the microgrid's generation supply exceeds the current load demand, and there is a certain reserve capacity. Referring to traditional power systems, the reserve capacity for a healthy state is set to 0.1 times the maximum load. A boundary state indicates that the current generation supply is exactly equal to the load demand. A risk state indicates that the current generation supply cannot meet the load demand, and some loads will experience power outages.

[0174] Because wind and solar power generation are significantly affected by external environmental factors, the operating state of a microgrid transitions between healthy, boundary, and risk states when operating independently. Figure 10 As shown, the reliability of a microgrid power generation system is evaluated using the probability values ​​of three states.

[0175] (2) Calculation formula for power supply reliability index

[0176] The expressions representing the healthy state, boundary state, and risk state of a microgrid are as follows:

[0177]

[0178] In the formula: S H This indicates that the microgrid is in a healthy state; P Lmax P represents the maximum load of the microgrid; Lt P represents the load of the microgrid at time t; t S represents the maximum power that the microgrid can output at time t; M This indicates that the microgrid is in a boundary state; S R This indicates that the microgrid is in a risky state.

[0179] The reliability assessment uses a 1-hour timeframe, assuming that the microgrid's output power and load remain constant within that hour. Using Well-being theory, the reliability indicators for the microgrid generation system are: healthy state probability P(H), boundary state probability P(M), and risky state probability P(R). The calculation expressions for each indicator are as follows:

[0180]

[0181]

[0182]

[0183] In the formula: N represents the number of simulated years in the reliability assessment process of the power generation system; n(H), n(M), and n(R) represent the cumulative number of hours the microgrid is in a healthy state, a boundary state, and a risk state during the reliability assessment process, respectively.

[0184] According to the definition of power supply reliability in the industry standard DL_T5542-2018: Power supply reliability (RS) represents the ratio of the total number of hours of effective power supply to users to the total number of hours in the statistical period, denoted as RS, and can be calculated using the following formula:

[0185]

[0186] In the formula: T user T represents the average power outage time. st Indicates the time period for the statistics.

[0187] Based on the Well-being theory and the DL_T5542-2018 industry standard, the formula for calculating the power supply reliability index of a self-consistent energy system for highways is as follows:

[0188]

[0189] (3) Flowchart for calculating power supply reliability index using the three-state sequential Monte Carlo method

[0190] The supply side consists of wind, solar, and energy storage, the demand side is the highway, and the intermediate architecture is a microgrid. The energy storage system strategy in the self-consistent energy system of the highway is as follows: when the supply-side output is greater than the load demand at time t, the energy storage device charges; when the supply-side output is less than the load demand, the energy storage device discharges. The specific process for calculating the power supply reliability index using a three-state sequential Monte Carlo simulation of system fault conditions, considering single-order component failures, is as follows: ① Constructing power models for the supply and demand sides of the self-consistent energy system of the highway. Constructing wind speed Weibull distribution model parameters and solar irradiance Beta model, sampling N=1000 8760h wind and solar scenarios using Monte Carlo, and then using the power conversion formula to calculate the unit output power under the 8760h wind and solar scenarios. For the load side, a probability model is constructed using a normal distribution, and then sampling N=1000 8760h load scenarios using Monte Carlo. ② Establishing reliability parameter models for each component of the self-consistent energy system of the highway. The three-state sequential Monte Carlo method posits that the operating state of all components consists of three sub-states: normal operating time T1, operating time with faults T2, and fault repair time T3. To obtain the specific times of these three states for each component, it is necessary to determine the equivalent failure rate λ and equivalent repair rate μ of all components in the system that require consideration of fault conditions. For circuit breakers, a key component in microgrids, the specific parameters λ and μ can be obtained by consulting relevant literature. For wind power, photovoltaic, and energy storage systems, the internal structures are complex and diverse, and simplified two-stage structures are adopted. First, the failure rate λ and repair rate μ of the internal components of the two-stage structure are obtained by consulting relevant literature. Then, the equivalent failure rate λ and equivalent repair rate μ of the overall wind power, photovoltaic, and energy storage system are calculated using the state-space method.

[0191] ③ After establishing the power model and reliability model, simulation can be performed. Set the total number of simulation years N = 1000, the simulation start year MT = 1, and HT = 0 to indicate simulation from hour 0 to 8760 hours. At the beginning, all components are in normal condition; the microgrid considers the failure scenarios of circuit breaker components, configures the capacity and number of wind, solar, and energy storage systems, and formulates strategies for energy storage charging / discharging processes. During the simulation, the supply and demand relationship of the system is compared, and the reliability parameters provided by well-being theory are used: the cumulative hours n(H), n(M), and n(R) of the healthy state, boundary state, and risk state, as well as the cumulative load reduction L, are accumulated hourly. A set of 8760 hours of wind and solar load data is read in for a three-state sequential Monte Carlo simulation. At the beginning, all components are in normal condition: n(H) = 0, n(M) = 0, and n(R) = 0, with a cumulative load reduction L = 0.

[0192] ④ First, determine the component that failed based on the minimum normal operating time T1. Set the total number of components in the system to n, and use the Matlab function rand() to generate n random numbers x. i Calculate the normal operating time T of each power supply. i =-(ln x i ) / λ i , λ i Let x be the failure rate of each component i (i = 1, 2, ..., n) in the system. i Let T1 = min(T) be a random number that follows a uniform distribution from 0 to 1. i If the power supply i corresponding to T1 fails, then the component i with the shortest normal operating time is considered to fail first.

[0193] ⑤ Determine the faulty component's operating time T2 + the faulty equipment's repair time T3.

[0194] Use the `rand()` function in Matlab to generate a random integer `x` between 1 and n. If 1 ≤ x ≤ a, it indicates a component failure in the power supply system (wind turbine, photovoltaic unit, or energy storage system). `x` is a random integer uniformly distributed between 1 and n, `a` is the sum of the number of components in the wind turbine, photovoltaic unit, and energy storage system, and `n` is the total number of components in the system. When a power supply component fails, another random number `x` is generated. j Calculate the sum of the fault-tolerant operating time T2 of power supply component j and the repair time T3 of the faulty device. μj T μj =-(ln x j ) / μ j μ j x represents the repair rate of faulty power supply component j. j To conform to a uniformly distributed random number from 0 to 1, the faulty power supply element is artificially set to exhibit a reduction in power generation capacity to 60% of its normal power generation capacity (normal power derived from the power conversion formula based on real-time wind speed / solar irradiance) within the fault diagnosis time T2. T2 occupies 2 / 3 of the T... μj Time, but only the repair time T3 = T for the corresponding faulty equipment. μj -T2 will be shut down for maintenance.

[0195] If a+1≤x≤n, it indicates that other switching elements, circuit breakers, transformers, etc. have failed. Since these components do not have the ability to generate electricity, but only play the role of transmitting electrical energy, after such components fail, they will all be out of service within the fault operation time T2 + the repair time of the faulty equipment T3.

[0196] ⑥ Perform power balance analysis on the system within the time period HT~HT+T1+T2+T3 (start time h~start time h+normal working time T1+fault-prone operation time T2+faulty equipment repair time T3). Let t=0:

[0197] 1) t' = HT + t, based on the power models and load models of wind power, photovoltaic power, and micro-sources, calculate the wind and solar power at time t' as P, respectively. W (t'), P S (t'), load power is P L (t'). At this time, the maximum power that the microgrid can output is P(t') = P W (t')+P S (t')+P B (t')(P B (t') represents the maximum discharge power of the energy storage device at time t'.

[0198] 2) If P W (t')+P S (t')≥P L If (t'), the energy storage device is charged, and step 4 is executed; otherwise, proceed to step 3).

[0199] 3) If P W (t')+P S (t')+P B (t')≥P L (t'), then proceed to step 4); otherwise L = L + P L (t')-P(t'), n(R)=n(R)+1, go to step 5).

[0200] 4) If P(t') ≥ 1.1P Lmax If n(H) = n(H) + 1, proceed to step 5); otherwise n(M) = n(M) + 1, proceed to step 5).

[0201] 5) If t <T λj +T μj If -1, then t = t + 1, return to step 1); otherwise, proceed to step 6.

[0202] ⑦ Advance the simulation time to HT = HT + T λj +T μj If HT < 8760, return to step ③; otherwise, execute MT = MT + 1.

[0203] Repeat the above process, each hour must not exceed 8760 hours. Once 8760 hours are exceeded, the above cycle terminates and proceeds to step ⑦ to determine whether to start a new year of 8760 hours of fault simulation.

[0204] ⑧ When MT≤N, read in the second set of 7680h data for wind and solar load, return to step ③, and perform fault simulation for the next year. Otherwise, exit the loop and calculate the power supply reliability index of the highway self-consistent energy system based on the cumulative hours of health status, boundary status, and risk status n(H)=0, n(M)=0, and n(R)=0; and the cumulative load reduction L=0.

[0205] In summary, the calculation process for the power supply reliability index of a self-sufficient energy system for highways is as follows: Figure 11 As shown.

[0206] Example

[0207] First, before calculating the power supply reliability index of the self-sufficient energy system for highways, the following data is required:

[0208] I. To construct an uncertain wind and solar load scenario, the following data is required:

[0209] (1) The average wind speed and data of 24 hours per hour on typical days in four quarters were used to construct the scale parameter c and shape parameter k of the Weibull distribution.

[0210] (2) The mean and variance of the hourly solar irradiance in the typical solar irradiance of the four quarters in 24h were used to construct the shape parameters α and β of the Beta distribution.

[0211] (3) The expected value and variance of the hourly load of the typical daily load model for four quarters are used to construct the normal distribution of the load.

[0212] 2. Since the system considers two types of wind turbines, it is necessary to obtain the turbine unit parameters separately: cut-in wind speed, cut-out wind speed, and rated wind speed; capacity and number of wind power, photovoltaic, and energy storage units; and maximum / minimum state of charge and maximum discharge / charge power of the energy storage units.

[0213] III. Reliability parameters of wind power, photovoltaic and energy storage systems, as well as other important equipment and switching components in microgrids (failure rate λ of all components in the system). i and repair rate μ j ).

[0214] In this simulation, an uncertain wind and solar load scenario was constructed. The wind speed probability density function for each hour of a typical day in four quarters was constructed using the Weibull probability density function. The solar irradiance probability density function for each hour of a typical day in four quarters was constructed using the Beta model. The load probability density function for each hour of a typical day in four quarters was constructed using the normal distribution.

[0215] The system considers two types of fans, and the fan unit parameters include: cut-in velocity, cut-out velocity, and rated velocity. Power capacity and reliability parameters are configured as follows: Figure 12 As shown in Table 3-1, the wind power, photovoltaic, and energy storage units adopt simplified unit structures. The overall system equivalent failure rate and repairability rate are obtained by consulting references.

[0216] Table 3-1 Power (capacity) configuration and reliability parameters of each power supply

[0217]

[0218]

[0219] Microgrids consider the different fault recovery times of wind turbines, photovoltaic units, energy storage units, and circuit breakers. Micro-generation (DG) and common loads can be connected to or disconnected from the grid via their respective circuit breakers, depending on the operational needs of the microgrid system. The constructed microgrid structure is as follows: Figure 13 As shown:

[0220] Combination Figure 14 , 13 Substitute the configuration from Table 3-1. Figure 11 The key characteristic indicators of the self-consistent energy system for highways are evaluated and calculated as follows.

[0221] The results of the simulation using MATLAB are shown in Table 3-2.

[0222] Table 3-2 Calculation Results of Power Supply Reliability

[0223]

[0224] 6. Comparison with the two-state sequential Monte Carlo example

[0225] The specific process for calculating the power supply reliability index by substituting the above input variables into a two-state sequential Monte Carlo simulation is as follows:

[0226] ① Based on historical wind and solar load data, we constructed the parameters for the Weibull distribution model of wind speed, the Beta model of solar irradiance, and the normal distribution probability model of the load. We then used Monte Carlo sampling to generate N = 1000 wind and solar load scenarios with a duration of 8760 hours each.

[0227] ② Parameter Initialization. Set the total number of simulation years N = 1000, the simulation start year MT = 1, and HT = 0 to indicate that the simulation starts from hour 0. The cumulative hours for the healthy state, boundary state, and risk state are n(H) = 0, n(M) = 0, and n(R) = 0, respectively; the cumulative number of times for the healthy state, boundary state, and risk state are m(H) = 0, m(M) = 0, and m(R) = 0, respectively; the cumulative load reduction L = 0; set the number of components to n, and all components are in normal state at the start; power limitations are considered during the energy storage charging / discharging process. Read in a set of 8760h wind and solar load data.

[0228] ③ Identify the faulty component. Let λ i Given the failure rate of i (i = 1, 2, ..., n), generate n random numbers x using the rand() function in Mat / ab. i Calculate the normal operating time T of each power supply. i =-(ln x i ) / λ i Let T1 = min(T) i If T1 corresponds to power supply i, then T1 is faulty. ④ Determine the repair time T2 of the faulty equipment, where T2 = T μj =-(ln x j ) / μ j μ j T represents the repair rate of faulty power supply j. μj x represents the repair time for equipment with component failure. j These are random numbers that conform to a uniform distribution from 0 to 1.

[0229] ⑤ Perform power balance analysis on the microgrid within the time period HT~HT+T1+T2 (start time h~start time h+normal working time T1+repair time of faulty equipment T2). Let t=0:

[0230] 1) t′=HT+t, based on the power models and load models of wind power, photovoltaic power, and micro-power sources, calculate the wind and solar power at time t′ as P, respectively. W (t′), P S (t′), load power is P L (t′). At this time, the maximum power that the microgrid can output is P(t′) = P W (t′)+P S (t′)+P B (t′)(P B (t′) represents the maximum discharge power of the energy storage device at time t′.

[0231] 2) If P w (t′)+P S (t′)≥P LIf (t′) occurs, the energy storage device is charged, and step 4 is executed; otherwise, proceed to step 3).

[0232] 3) If P W (t′)+P S (t′)+P B (t′)≥P L (t′), then proceed to step 4); otherwise L = L + P L (t′)-P(t′), n(R)=n(R)+1, go to step 5).

[0233] 4) If P(t′) ≥ 1.1P Lmax If n(H) = n(H) + 1, proceed to step 5); otherwise n(M) = n(M) + 1, proceed to step 5).

[0234] 5) If t < T λj +T μj If -1, then t = t + 1, return to step 1); otherwise, proceed to step 6.

[0235] ⑥ Advance the simulation time to HT = HT + T1 + T2. If HT < 8760, return to step ③; otherwise, execute MT = MT + 1.

[0236] Repeat the above process, each hour must not exceed 8760 hours. Once 8760 hours are exceeded, the above cycle terminates and proceeds to step ⑦ to determine whether to start a new year of 8760 hours of fault simulation.

[0237] ⑦ When MT≤N, read in the second set of 7680h data for wind and solar load, return to step ③, and perform fault simulation for the next year. Otherwise, exit the loop and calculate the power supply reliability index of the highway self-consistent energy system based on the cumulative hours of health status, boundary status, and risk status n(H)=0, n(M)=0, and n(R)=0; and the cumulative load reduction L=0.

[0238] In summary, the calculation process for the power supply reliability index of a self-sufficient energy system for highways is as follows: Figure 12 As shown.

[0239] Figure 14 Two-state sequential Monte Carlo calculation flowchart for power supply reliability index

[0240] Combination Figure 14 , 13 Substitute the configuration from Table 3-1. Figure 14 The key characteristic indicators of the self-consistent energy system for highways are evaluated and calculated as follows.

[0241] The results of the simulation using MATLAB are shown in Table 3-3.

[0242] Table 3-3 Calculation Results of Power Supply Reliability

[0243] Power supply reliability RS ％ 90.0462％ Expected power shortage EENS kWh / year 2872.287 kWh / year

[0244] A comparison of the performance indicators calculated by the two-state sequential Monte Carlo algorithm and the three-state sequential Monte Carlo algorithm (Table 3-2) shows that the three-state algorithm, which considers the fault-prone operating time on the generator side, calculates less expected power supply shortage (i.e., load reduction) than the two-state sequential Monte Carlo algorithm. This is because the three-state algorithm takes into account the fault-prone operating time on the generator side. During this time, the generator's power generation capacity decreases, but it does not immediately shut down for maintenance. Consequently, the three-state sequential Monte Carlo algorithm better reflects the actual fault handling situation in electric fields.

[0245] This invention addresses the calculation of power supply reliability indicators for self-consistent energy systems on highways during the planning phase, and innovates a three-state sequential Monte Carlo algorithm that considers multiple component failure scenarios. When performing sequential Monte Carlo calculations for power supply reliability indicators in self-consistent energy systems on highways, the uncertainties on both the supply and demand sides are fully considered when establishing the power generation model on the supply side and the load model on the demand side. Probability density distributions are constructed on both the supply and demand sides, and Monte Carlo sampling is used to simulate wind and solar load scenarios.

[0246] From the perspective of the three-state sequential Monte Carlo algorithm itself, the traditional sequential Monte Carlo algorithm's calculation process treats components as having only two states: normal operation and fault. It analyzes the relationship between power supply and load only during the normal operation time T1 and the fault repair time T2, assuming that all types of components, whether power supplies or other switching components, will only experience shutdown faults when a fault occurs. The three-state sequential Monte Carlo simulation time includes three states: normal operation time T1 + fault-prone operation time T2 + faulty equipment repair time T3. This distinguishes the fault states of power supply components from those of ordinary switching components. The random number x is used to... jBy comparing the fault probability with the fault diagnosis time T2 and the repair time T3, the photovoltaic, wind turbine, and energy storage systems on the supply side exhibit a reduction in power generation capacity to 60% of their normal power generation capacity (normal power calculated from real-time wind speed / solar irradiance) within the fault diagnosis time T2, and only require shutdown maintenance within the corresponding faulty equipment's repair time T3. Ordinary switching components, however, consistently exhibit shutdown faults within both the fault diagnosis time T2 and the repair time T3. The three-state sequential Monte Carlo algorithm not only considers the application objects more comprehensively and specifically, but also reflects the difference between power supply components and other component types within the algorithm itself. It takes into account various fault types, overcoming the limitation of traditional sequential Monte Carlo algorithms that do not distinguish between power supply faults and faults in other components such as switching components.

[0247] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A method for calculating the power supply reliability index of a self-consistent energy system for highways, characterized by: Step 1: The specific structure of establishing a self-sufficient energy system for highways includes: a supply side, a demand side, and a microgrid; the supply side consists of wind, solar, and energy storage power generation systems; the demand side is mainly based on the demand of highways and is connected through the intermediate medium of microgrids. Step 2: Construct the power model on the supply side and the traffic load model on the demand side of the self-consistent energy system for highways: Using different probability density distributions, the wind speed scenario of wind turbines, the solar irradiance scenario of photovoltaic units, and the traffic load demand scenario of highways are determined by sampling using the three-state sequential Monte Carlo method. The power generation model of wind turbines is constructed using the power-wind speed conversion formula of wind turbines, and the power generation model of photovoltaic units is constructed using the power conversion formula. The typical load day fitting method based on normal distribution is adopted to approximate the load scenarios of the four quarters of the year. The expected value of the normal distribution of each hour within 24 hours is obtained by sampling as the demand value of traffic load at that time, which is used as the demand-side traffic load model. Step 3: Establish reliability parameter models for the components of the self-consistent energy system of the highway; Step 4: Calculate power supply reliability indices using three-state sequential Monte Carlo methods; The power model of the real-time power supply on the supply side includes: Wind power model: Wind speed at a certain moment The output power of a single wind turbine can be calculated using the power conversion formula, where the conversion formula is: (3-1) In the formula: P r P represents the rated power of a single fan. t v is the power of the wind turbine at time t; r v ci v co These represent the rated wind speed, cut-in wind speed, and cut-out wind speed of the fan, respectively; the values ​​of A, B, and C depend on... and Size; (3-2)； Photovoltaic power model: Irradiance G at a certain moment t The output power of a photovoltaic unit can be calculated using the power conversion formula, where the conversion formula is: (3-3) In the formula: P t P represents the power of the photovoltaic unit at time t. m G is the rated power of the photovoltaic unit; std The irradiance given for the standard environment is taken as 1 kW / m²; R c The irradiance is taken as 0.15 kW / m². Energy storage charging / discharging model Suppose The energy stored in the instantaneous energy storage system is B(t), and the charging / discharging power is P. B (t), then the charging / discharging time series of the energy storage system is: (3-4) The power limit of the energy storage device during charging / discharging is as follows: (3-5) (3-6) In the formula: P B A positive value for (t) indicates that the energy storage device is charging; P B A negative value for (t) indicates that the energy storage device is discharging; P disch-max P represents the maximum discharge power of the energy storage device. ch-max B is the maximum charging power of the energy storage device. max B represents the maximum capacity of the energy storage device. min This is the minimum capacity of the energy storage device; A typical load day fitting method based on normal distribution is adopted to approximate the load scenarios of the four quarters of a year. Since the load of a highway in 24 hours follows 24 different normal distributions, the expected value of the normal distribution of each hour within 24 hours is obtained by sampling as the traffic load demand value at that time, which serves as the traffic load model on the demand side. The reliability parameter model of the components includes: the reliability model of the energy storage device: since the energy storage system adopts a two-stage battery structure, the components consist of four parts: battery pack, DC / DC converter, DC / AC inverter, and grid-connected filter. The reliability model is determined by obtaining the failure rate of each of the four parts of the energy storage device. , , , and repair rate , , and Using the state-space method and the equivalent failure rate formula Equivalent Repair Rate Formula Obtain the overall equivalent failure rate of the energy storage device. times / year, and equivalent repair rate times / hour; Reliability model of wind turbine Because wind turbine generators adopt a two-stage structure, they consist of four parts: a wind turbine generator, an AC / DC rectifier, a DC / AC inverter, and a filter. The failure rates of each of these four parts are then obtained. , , , The repair rates were respectively , , and Using the state-space method and the equivalent failure rate formula Equivalent Repair Rate Formula Obtain the overall equivalent failure rate of the wind turbine unit. times / year and equivalent repair rate times / hour; Reliability model of photovoltaic units Because photovoltaic (PV) generators adopt a two-stage structure, they consist of four parts: a photovoltaic array, a DC / DC converter, a DC / AC inverter, and a filter. The failure rates of each of these four parts are obtained. , , , The repair rates were respectively , , and Then, using the state-space method and the equivalent failure rate formula... Equivalent Repair Rate Formula Obtain the overall equivalent failure rate of the photovoltaic unit. times / year, and equivalent repair rate , times / hour.

2. The method for calculating the power supply reliability index of a self-consistent energy system for highways according to claim 1, characterized in that: Step 4 further includes the following: (1) Establish power models for the supply and demand sides of a self-consistent energy system for highways; (2) Establish reliability parameter models for each component of the self-consistent energy system of the expressway; (3) Perform simulation: Set the total number of years of simulation N=1000, the simulation start year MT=1, and HT=0 to indicate that the simulation starts from hour 0 and continues until 8760 hours; (4) Determine the component that failed by using the minimum normal operating time T1; (5) Determine the faulty component's operating time T2 + the faulty equipment's repair time T3; (6) Perform power balance analysis on the system during the time period HT~HT+T1+T2+T3; (7) Advance the simulation time to ,if Return to step (3); otherwise execute MT = MT + 1; let ,but Corresponding power supply A malfunction occurred. µ j The repair rate of faulty power supply component j. These are random numbers that conform to a uniform distribution between 0 and 1. (8) When MT≤N, read in the second set of 7680h data of wind and solar load, return to step (3), and perform fault simulation for the next year.

3. A self-consistent energy system for highways, characterized by: The system includes a non-volatile storage medium containing a stored program, wherein the program, when executed, controls the device where the non-volatile storage medium resides to perform the method of claim 1.

4. An electronic device, characterized in that, It includes a processor and a memory; the memory stores computer-readable instructions, and the processor is used to execute the computer-readable instructions, wherein the computer-readable instructions, when executed, perform the method of claim 1.

Images