A power distribution network reliability optimization scheduling method and system based on fault monitoring
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- LISHUI POWER SUPPLY COMPANY OF STATE GRID ZHEJIANG ELECTRIC POWER
- Filing Date
- 2026-03-19
- Publication Date
- 2026-06-19
Smart Images
Figure CN122243072A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of distribution network reliability assessment technology, and specifically to a distribution network reliability optimization scheduling method and system based on fault monitoring. Background Technology
[0002] With the continuous development of smart grid technology, the distribution network, as the end point of the power system, is directly connected to users, and its power supply reliability has a significant impact on users' electricity experience and economic and social development. Traditional distribution network planning projects often focus on optimizing the network structure, such as adding tie switches, sectionalizing switches, and ring main units. These measures mainly improve reliability from the perspective of the network structure, but lack refined consideration for operation and maintenance. In actual operation, operation and maintenance factors such as equipment failure rate, feeder automation level, and uninterrupted power supply capability also have a significant impact on power supply reliability and are closely related to operation and maintenance investment.
[0003] However, existing methods for assessing the reliability of power distribution networks have the following shortcomings: First, failure rate models mostly use static statistical values and fail to make full use of real-time monitoring data and historical equipment information for dynamic prediction, resulting in delayed early warnings and crude maintenance strategies. Second, they do not adequately consider the intermediate development process of equipment from normal to fault state and fail to achieve quantitative assessment of equipment health status. Third, existing methods mostly focus on post-fault maintenance analysis and lack comprehensive consideration of preventive maintenance and its economics.
[0004] For example, CN110288208A discloses a comprehensive evaluation method for the reliability and economy of a radial distribution network. This method constructs a failure rate membership function based on fuzzy theory to address the problem of limited historical outage data. However, the failure rate model still relies on static statistical values and fails to integrate real-time monitoring data for dynamic correction, nor does it quantify the impact of maintenance actions on the failure rate. Furthermore, its optimization model focuses on network structure modification and capacitor configuration, neglecting maintenance investment decisions such as feeder automation and live-line work, thus failing to achieve dynamic synergistic optimization of economy and reliability. CN110490376A proposes an intelligent soft-switching planning method for improving the reliability and economy of distribution networks, aiming to minimize the sum of investment cost and reliability cost to enhance the self-healing capability of the distribution network. However, it does not consider the synergistic optimization of conventional maintenance methods such as feeder automation and live-line work; its failure rate relies on historical statistical values, lacking dynamic correction based on real-time monitoring data and quantification of maintenance impact, thus failing to achieve dynamic assessment and early warning of equipment health status.
[0005] Therefore, there is an urgent need for a method that can dynamically assess equipment status based on multi-source data, generate hierarchical early warnings, and optimize scheduling strategies to achieve a comprehensive improvement in the reliability of power supply in the distribution network and optimize operation and maintenance costs. Summary of the Invention
[0006] To address the shortcomings of existing technologies, the present invention aims to provide a method and system for optimizing the scheduling of power distribution networks based on fault monitoring, thereby resolving the problems mentioned in the background section.
[0007] To achieve the above objectives, the present invention provides a method for optimizing the reliability of a distribution network based on fault monitoring, comprising the following steps: Step 1: Multi-source monitoring data acquisition and preprocessing, collecting multi-source monitoring data of power distribution network equipment to form an equipment status characteristic database; Step 2: Establish a dynamic model of equipment failure rate based on Weibull distribution, and use Marquardt method to determine the boundary point and corresponding parameters of equipment during the random failure period and the wear failure period; dynamically correct the basic failure rate by integrating real-time monitoring data, and introduce service life rollback factor to quantify the impact of maintenance on equivalent service life, so as to realize dynamic prediction of equipment failure rate. Step 3: Based on the predicted dynamic failure rate, combined with the equipment's operating years, maintenance history, and environmental factors, construct an equipment health index, and classify and issue early warnings for the equipment according to the health index; Step 4: Construct an optimization decision model considering the economics of operation and maintenance. With the goal of improving power supply reliability, a total cost of ownership (TCO) minimization optimization model is established by comprehensively considering equipment failure rate, feeder automation coverage, and uninterrupted power supply (UPS) capability. The optimal investment allocation scheme is solved by taking the output dynamic failure rate and health index as inputs, the investment in equipment failure rate improvement, feeder automation, and UPS as decision variables, and the investment budget and unidirectional evolution of equipment status as constraints. Step 5: Generate the optimal scheduling strategy. Based on the final investment allocation plan, convert it into executable scheduling instructions and generate a scheduling strategy that includes maintenance plans, feeder automation deployment plans, and live-line work implementation plans.
[0008] As a second aspect of the present invention, a distribution network reliability optimization scheduling system based on fault monitoring is proposed, including a memory and a processor. The memory stores a computer program, and when the computer program is executed by the processor, it implements a distribution network reliability optimization scheduling method based on fault monitoring.
[0009] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention integrates multi-source monitoring data and historical equipment information, constructs a dynamic failure rate prediction model based on the Weibull distribution, and uses the Marquardt method to accurately delineate the boundary between accidental failure periods and wear-out failure periods. Simultaneously, it introduces real-time monitoring data correction factors and service life regression factors to quantify the impact of changes in operating status and maintenance activities on equipment failure rates, achieving dynamic and accurate assessment of equipment health status. Compared to traditional static failure rate models, this invention can identify equipment degradation trends earlier, providing a scientific basis for proactive early warning and preventative maintenance, and effectively avoiding unplanned power outages caused by sudden failures.
[0010] This invention aims to minimize total cost of ownership by constructing an optimization decision-making model that considers operational and maintenance (O&M) economics, and comprehensively manages investments in equipment failure rate improvement, feeder automation, and uninterrupted power supply (UPS) operations. By using dynamic failure rate and health index as model inputs, the optimal investment allocation scheme is solved under constraints such as investment budget and unidirectional evolution of equipment status, achieving synergistic optimization of power supply reliability improvement and O&M economics.
[0011] This invention transforms the abstract investment allocation scheme output by the optimization model into specific executable scheduling instructions that include equipment maintenance priorities, feeder automation deployment sequence, and uninterrupted power supply (UPS) capability construction sequence; and ensures the feasibility of the scheduling strategy under budget constraints through an investment budget closed-loop verification mechanism. Attached Figure Description
[0012] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts. Wherein: Figure 1 This is a schematic diagram of the overall processing flow of the distribution network reliability optimization scheduling method based on fault monitoring proposed in one embodiment of the present invention; Figure 2 This is a schematic diagram of the bathtub curve representing the equipment failure rate proposed in one embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the change in failure rate before and after maintenance, as proposed in one embodiment of the present invention. Detailed Implementation
[0013] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.
[0014] The present invention will be further described in detail below with reference to the accompanying drawings, but this is not intended to limit the scope of the invention.
[0015] As a first aspect of the present invention, such as Figures 1 to 3 As shown, a distribution network reliability optimization scheduling method based on fault monitoring is proposed, which includes the following steps: Step 1: Multi-source monitoring data acquisition and preprocessing To achieve a comprehensive understanding of the operating status of power distribution network equipment, it is necessary to collect multi-source heterogeneous data from multiple business systems, covering the entire life cycle of the equipment and its real-time operating status.
[0016] Real-time operating data is acquired from SCADA systems, power distribution automation systems, or online monitoring devices. This includes electrical quantities such as three-phase current, three-phase voltage, active power, reactive power, power factor, and harmonic content; status quantities such as circuit breaker opening and closing positions, switch status, and transformer tap positions; and thermal status quantities such as equipment temperature and ambient temperature. The data acquisition frequency is determined according to the data type. Electrical quantities are generally acquired on a minute-by-minute basis, status quantities are event-triggered, and temperature quantities are acquired on an hourly basis.
[0017] Environmental data, including ambient temperature, lightning strikes, wind speed, humidity, etc., are acquired from meteorological monitoring systems or micrometeorological devices and spatially correlated with the equipment location to analyze the impact of environmental factors on equipment failure rates.
[0018] Historical fault records are extracted from the Production Management System (PMS) or Fault Management System, including faulty equipment ID, equipment type, fault occurrence time, fault recovery time, fault type, fault cause, fault location, and fault impact information such as the number of households affected and the number of affected users at the time of power outage, thereby constructing the core data for the basic failure rate model.
[0019] Obtain equipment ledger information from the asset management system or production management system, including parameters such as equipment ID, equipment type, years of operation, manufacturer, and model, for equipment classification and aggregation analysis of similar equipment.
[0020] Extract maintenance records from the production management system or maintenance management system, including maintenance equipment ID, maintenance time, maintenance type, maintenance content, and maintenance cost information such as labor hours, material costs, and power outage time. This information is used to quantify the impact of maintenance on equipment failure rate and support the determination of service life rollback factors.
[0021] After data acquisition, the raw data needs to be preprocessed to eliminate data quality issues. First, data cleaning is performed to remove outliers that are significantly outside the reasonable range, interpolation or labeling is used to complete missing data, and duplicate records are eliminated. Second, time alignment is performed to unify data from different sources and with different acquisition frequencies to the same time scale. Then, data normalization is performed to map feature parameters of different dimensions to the [0,1] interval to eliminate the influence of different dimensions. Finally, data association is performed to link different types of data from the same device according to device ID and timestamp, forming a device status time series.
[0022] After the above preprocessing, a device status feature database is formed. The database contains the device's basic attributes, real-time operating features, environmental features, fault history features, and maintenance history features, providing high-quality data support for subsequent dynamic failure rate prediction models.
[0023] Step 2: Construct a dynamic failure rate prediction model A dynamic model of equipment failure rate is established based on the Weibull distribution. The model parameters are fitted using the least squares method or the Marquardt method and dynamically updated according to real-time monitoring data.
[0024] like Figure 2 The diagram shows a bathtub curve illustrating the equipment failure rate. This curve reflects the change in failure rate over time throughout the equipment's lifespan, and can be divided into three stages: the early failure period, the random failure period, and the wear-out failure period. During the early failure period, the failure rate is high initially but decreases rapidly; during the random failure period, the equipment operates stably with a low and relatively constant failure rate; during the wear-out failure period, the equipment ages, and the failure rate increases rapidly with operating time. This invention fits the curve based on a Weibull distribution, describing the characteristics of each stage through different values of the shape parameter β: β < 1 corresponds to the early failure period, β = 1 corresponds to the random failure period, and β > 1 corresponds to the wear-out failure period. In practical engineering applications, the parameters are mainly fitted for the random failure period and the wear-out failure period separately.
[0025] Specifically, in practical engineering applications, Weibull distribution fitting is typically performed using equipment failure time point data or failure interval time data. For simplicity, historical failure records are converted into annual failure indicator variables (0 indicates no failures this year, 1 indicates at least one failure this year) and approximated as the annual failure rate. For cases with less than one year of operation, a proportional calculation is required. In this embodiment, each year represents a complete operation, so 0 / 1 is used directly for counting.
[0026] The Marquardt method is used to iteratively solve for the boundary points and corresponding parameters of the two stages. The specific steps are as follows: Collect n sets of failure rate data Arrange the data in chronological order; randomly select two sets of data as the initial values for the first iteration to fit the preliminary Weibull parameters. and ; Calculate the residual sum of squares Q for each data point under the current parameters; construct the damping matrix by calculating the partial derivatives, update the parameter estimates, and repeat the iteration until the residual sum of squares converges.
[0027] At the same time, for each possible boundary point k, calculate the expression that minimizes the following: ,in, The average failure rate is the rate of random failures. The actual failure rate in year ti. The value is the Weibull fit during the wear and tear failure period; the k that minimizes Sk is selected as the optimal boundary point.
[0028] For the data in this example, the optimal boundary point was obtained through iterative calculation using the Marquardt method. =10 years. That is, the first 10 years are the period of occasional failures, and the period of wear and tear failures begins from the 11th year.
[0029] The data is fitted using a Weibull distribution, with known data points: (11, 1.0), (12, 0), (13, 1.0), (14, 0), (15, 1.0). Because the data is sparse and contains zero values, in practical engineering, cumulative distribution, average failure rate, or clustering of similar equipment data are often used. This is a conceptual demonstration, using a failure time point fitting method. The fault time interval follows a Weibull distribution, and its cumulative distribution function is: The failure time points (commissioning years) were selected as 11, 13, and 15, and the reliability values for each failure point were obtained through Kaplan-Meier estimation. Then, the parameters are solved by linear regression using the Weibull distribution linearization formula. and : .
[0030] Parameter fitting results: obtained through calculation Substituting into the Weibull failure rate function, we obtain the wear-out failure period failure rate function: .
[0031] Example: Instantaneous failure rate calculation when t=13: .
[0032] Furthermore, based on the basic failure rate model, a multi-source monitoring data correction factor is introduced to make the model reflect the current actual operating status of the equipment. The correction factor is derived from multi-source monitoring indicators: Electrical quantities: load current, voltage fluctuations, harmonic content, etc. For example, prolonged overload will accelerate insulation aging, and correction factors... ; Mechanical quantities: Number of circuit breaker operations and cumulative breaking current. As the number of operations approaches the mechanical lifespan, the failure rate increases, requiring a correction factor. ; Environmental factors: ambient temperature, humidity, and lightning strike density. High temperature and humidity accelerate insulation degradation; correction factor. ; State variables: partial discharge intensity, vibration, gas content in oil, etc. For example, when partial discharge exceeds a threshold, .
[0033] The overall correction factor K is the product of the components: The product form implies the assumption that each correction factor is independent. This assumption is based on experience in decoupling factors in engineering practice. If actual data shows that there is coupling between factors, weighted sum or other forms can be used for correction.
[0034] For the circuit breaker in this example, the real-time monitoring data for 2023 (the 15th year) is as follows: Load current: 85% of rated current (normal). ;
[0035] Number of operations: 90% of the machine's lifespan, b=0.5, then ;
[0036] Ambient temperature: 10℃ higher than the reference temperature, c=0.03, then ;
[0037] Partial discharge: If the intensity is 0.8 times the threshold, and e=0.2, then... ;
[0038] Calculate the overall correction factor:
[0039] Therefore, the real-time failure rate in 2023 is: Considering the real-time status, the circuit breaker has a failure probability of about 25% in the current year, which is classified as a medium-risk device.
[0040] Furthermore, considering the impact of equipment maintenance on the failure rate, a service life rollback factor is introduced to describe the equivalent service life of the equipment after maintenance. For example... Figure 3The figure shows the change in equipment failure rate before and after maintenance. The horizontal axis represents the equipment's service life, and the vertical axis represents the failure rate. Curve ① is the basic failure rate curve under the condition of no maintenance, and curve ② is the failure rate curve after maintenance. According to the statistical data of this embodiment, the failure rates of the equipment before maintenance at service life of t1=2 years, t2=3.5 years, and t3=5 years were 0.10 times / year, 0.15 times / year, and 0.20 times / year, respectively. After the first major overhaul of the equipment in the 3rd year, the equivalent service life decreased from 3 years to 2.25 years. After maintenance, the failure rates of the equipment at the corresponding service life decreased to 0.08 times / year, 0.12 times / year, and 0.16 times / year, respectively, with an overall decrease of approximately 20%. As can be seen from the figure, maintenance shifts the failure rate curve downwards overall, but maintenance cannot completely restore the equipment to its original condition, and the failure rate will continue to increase with the increase of service life.
[0041] In this embodiment, the process of quantifying this process through the service age regression factor is as follows: In the formula, ta is the equivalent service life after maintenance. Let be the service life rollback factor for the i-th overhaul, and c be the adjustment factor. The impact of overhaul on the failure rate is quantified using the equivalent service life method. This means that the failure rate of the equipment after overhaul no longer follows the curve before overhaul, but is re-determined from the basic failure rate function based on the equivalent service life. Specifically, the failure rate in the s-th year after overhaul is... By continuously adjusting model parameters through real-time data monitoring, dynamic prediction of failure rates can be achieved.
[0042] Taking a certain circuit breaker as an example, the actual service life during maintenance... Years; historical data were fitted to obtain the first major overhaul service age regression factor. (Equivalent service life reduced by 25%); Basic failure rate before overhaul ; equivalent service age after maintenance The corresponding failure rate ; Reduction in maintenance failure rate .
[0043] If the failure rate is assessed in the 13th year (the second year after maintenance), then In practical applications, the final failure rate needs to be obtained by combining real-time correction factor K.
[0044] Step 3: Quantify the health status of the equipment Based on dynamic failure rate prediction results, and combined with equipment operating years, maintenance history, environmental factors, etc., an equipment health index is constructed. The health index is defined as a normalized indicator of the current equipment failure rate relative to the maximum allowable failure rate, specifically: ,in, For real-time equipment failure rate, This represents the maximum allowable failure rate for the equipment. The value can be determined comprehensively based on the equipment type, operating environment, and power supply reliability requirements. For example, for a 10kV circuit breaker, 0.1 times / year can be used. Specific values can be set with reference to industry standards or historical statistical data. The health index value is between 0 and 1, with a higher value indicating better equipment condition.
[0045] The health index is divided into four levels: normal ( ),Notice( ), early warning ( ),serious( When the health index falls below the warning threshold, the system automatically generates a warning message.
[0046] Step 4: Construct an optimization decision model With the goal of improving power supply reliability, a total cost of ownership (TOC) minimization optimization model is established, comprehensively considering factors such as equipment failure rate, feeder automation coverage, and uninterrupted power supply (UPS) capability. TOC includes investment cost and outage loss cost. Where: C represents the total investment cost, including investment in improving equipment failure rates (C1) and investment in feeder automation (C2). 1. Investment in uninterrupted power outages (C3); if is the discount rate considering interest rates and inflation; n is the system lifespan (usually 25 years); b is the ratio of the electricity price per unit of power outage to the average electricity price, calculated using the average electricity price multiplier method (usually 25); d is the average electricity price (yuan / kWh); EENS is the expected power shortage (kWh / year), reflecting the power loss caused by power outages.
[0047] Specifically, the investment c1 for improving equipment failure rate includes operating costs, maintenance costs, and failure costs.
[0048] Operating costs refer to the routine maintenance expenses incurred during the service life of equipment. They are calculated based on the equipment's basic annual operating costs and increase linearly with the length of service life. ,in, The basic operating cost is L, the service life is a, and a is the growth coefficient (taken as 0.05).
[0049] Maintenance costs mainly refer to preventative maintenance expenses, including major overhauls and periodic minor repairs. The specific figures depend on the equipment type, maintenance cycle, and extent of maintenance. , where t is the time required for maintenance, and ρi is the degree of maintenance for the i-th maintenance (the value ranges from 0 to 1, representing the thoroughness of the maintenance).
[0050] Failure cost is calculated by multiplying the expected number of failures over the equipment's lifespan by the cost of a single maintenance operation. The calculation of the expected number of failures must distinguish whether the equipment has undergone maintenance. If no maintenance has been performed, the expected number of failures is the integral of the failure rate function over time over the lifespan, i.e., =∫λ(t)dt; If maintenance has been performed, the correction for the failure rate due to service life reduction needs to be considered. That is, the failure rate function is recalculated and integrated based on the equivalent service life after maintenance to obtain the expected number of failures under maintenance conditions. .
[0051] Feeder automation investment C2: includes initial investment and operating costs.
[0052] Initial investment is calculated based on the number of distribution terminals (FTUs): ,in, To increase the coverage of power distribution equipment, The total number of switches, This is the unit price for FTU.
[0053] Operating costs are determined by a certain discount rate to the initial investment, and also take into account linear growth over the years.
[0054] The investment for live-line work (LDP) C3 is also divided into initial investment and operating costs. The initial investment is calculated on a per-feeder basis, based on the line length and the investment cost of LDP equipment per unit length; the operating cost is determined by referring to the investment method for feeder automation, using a certain conversion ratio of the initial investment to determine the basic operating expenses, and taking into account the annual growth factor.
[0055] Furthermore, the outage loss is calculated by converting the system average outage duration (SAIDI) into the expected power shortage (EENS): ,in, Here, N represents the typical power supply area number, and h represents the feeder number. For the first Number of feeder lines in each area For the first Unit user load capacity of a typical power supply area For the first The average outage duration of the h-th feeder in the region. This represents the number of users corresponding to the feeder.
[0056] In this embodiment, the optimization model needs to be solved under the premise of meeting the actual operation and maintenance conditions. The main constraints include the following aspects: Investment budget constraints: The total amount of all investments must not exceed the annual budget limit. At the same time, investments in equipment failure rate improvement, feeder automation, and uninterrupted power supply work must also be controlled within their respective sub-budgets.
[0057] Feeder automation state constraints: Feeder automation transformation is a gradual process. Once a feeder is automated, its state variable is recorded as 1, and it will not revert to the unimplemented state in subsequent planning. The automation state of each feeder can only be 0 or 1, representing unimplemented and implemented, respectively.
[0058] Uninterrupted power supply (UPS) status constraints: Similar to feeder automation, the development of UPS capabilities is also a unidirectional process. Once a feeder meets the conditions for UPS operation, its state variable changes from 0 to 1 and remains unchanged, and it is also a binary variable.
[0059] Equipment failure rate constraints: Through reasonable maintenance planning, it should be ensured that the real-time failure rate of all equipment is always kept below the maximum allowable value. For example, it may be stipulated that the annual failure rate of a certain type of equipment shall not exceed 0.1 times / year.
[0060] Matching constraints between investment and state variables The investment in feeder automation and live-line maintenance (LLM) must match the number of new feeders to be automated or LLM-enabled in the current year. That is, the new investment equals the sum of the unit investment required for each feeder upgrade, where unit investment includes FTU purchase cost, LLM equipment purchase cost, etc., with specific values determined based on factors such as line type, length, and number of switches.
[0061] Taking a typical power supply area as an example, the area contains 10 overhead lines and 5 cable lines. The parameters and basic data of each line are shown in Table 1 below (partial data): Table 1 Set a cap on the total investment budget 10,000 yuan, of which the upper limit for feeder automation investment is [amount missing]. Investment limit for live-line work Discount rate Life cycle Electricity price multiplier b = 25, average electricity price The unit price of an FTU is 15,000 yuan, and the investment per unit length for uninterrupted power supply work is [amount missing]. Operating cost conversion factor .
[0062] A genetic algorithm was used to find the optimal combination of decision variables: feeder automation should be prioritized for overhead lines with a large number of segments and interconnections, while live-line work should be prioritized for longer lines with more users. The optimized TOC value is... Yuan, compared to taking no action ( The amount decreased by 32.8% (in yuan) and all constraints were met.
[0063] Step 5: Generate the optimal scheduling strategy In step 4, the optimal decision variables for each feeder under a given investment budget are obtained using the Total Cost of Ownership (TOC) minimization optimization model: namely, which feeders should implement feeder automation, which feeders should have uninterrupted power supply (UPS) capability, and the allocation of the equipment failure rate improvement investment C1 among operation, maintenance, and failure costs. Then, specific executable scheduling instructions are generated using the optimal scheduling strategy.
[0064] Based on the real-time Health Index (HI) of each device, an equipment maintenance list is generated in the order of priority: severe, warning, attention, and normal. For multiple devices within the same warning level, they are sorted according to their Health Index from smallest to largest (i.e., risk from highest to lowest), with lower Health Indexes having higher maintenance priority.
[0065] The optimization model provides a list of feeders that should be automated, but in actual deployment, the implementation order must be considered to maximize the return on investment of limited resources.
[0066] The optimization model provides a list of feeders that should have live-line working capabilities. Building live-line working capabilities mainly includes equipping them with bypass work vehicles, mobile power supply vehicles, insulated bucket trucks, and training professional workers.
[0067] Substitute the generated maintenance plan, feeder automation deployment scheme, and live-line work implementation plan into the cost model from step 4, and recalculate all investments. The process involves checking whether budget constraints are met. If the budget is exceeded, lower-priority projects are cut in order of priority until the constraints are met. For example, if the investment in feeder automation exceeds... Then, the reduction will begin with the line with the lowest number of segments, until the investment falls back within the limit.
[0068] For example, the optimization results show that: feeder automation should be prioritized for 3 overhead lines with 4 segments and interconnections; FTUs should be installed on 2 ring network cable lines; live-line maintenance should be carried out on 4 overhead lines; and, based on the health index ranking, major overhauls should be scheduled for 2 distribution transformers and 1 circuit breaker. After implementation, the regional SAIDI is expected to decrease by 30%, EENS by 25%, and the incremental investment (4.867 million yuan) can be recovered in approximately 4.8 years through the reduction in power outage losses each year, demonstrating good economic efficiency.
[0069] As another aspect of the present invention, a distribution network reliability optimization scheduling system based on fault monitoring is proposed, comprising: The data acquisition module is used to collect multi-source monitoring data from power distribution network equipment. This multi-source monitoring data includes: real-time operational data from SCADA and distribution automation systems, covering electrical and status quantities such as current, voltage, load, and equipment temperature; environmental data from meteorological monitoring systems, covering meteorological elements such as lightning strikes, wind speed, humidity, and temperature; historical fault records from the production management system, covering fault type, fault time, and repair time; equipment ledger information from the asset management system, covering basic parameters such as equipment type, service life, manufacturer, and model; and maintenance records from the maintenance management system, covering maintenance time, maintenance type, and maintenance content. The data acquisition module has a built-in data preprocessing unit that cleans the collected raw data to remove outliers, normalizes it to eliminate dimensional influences, and aligns it to unify the time scale, ultimately forming a structured equipment status characteristic database. This database is organized and stored according to equipment ID and timestamp, providing data support for subsequent modules.
[0070] The dynamic failure rate prediction module, connected to the data acquisition module, is used to construct a dynamic failure rate prediction model based on historical failure data and real-time monitoring data in the equipment status characteristic database. This module uses the Weibull distribution to describe the change in equipment failure rate over time and incorporates the Marquardt iterative solution algorithm. By minimizing the sum of squared residuals, it determines the boundary between accidental failure periods and wear-out failure periods, as well as the corresponding Weibull distribution parameters. This module further constructs a multi-source monitoring data correction factor model. Based on real-time operating data and environmental data, it dynamically calculates electrical quantity correction factors, mechanical quantity correction factors, environmental quantity correction factors, and state quantity correction factors. The product of these correction factors is used as a comprehensive correction factor to correct the base failure rate in real time, yielding the real-time equipment failure rate. Simultaneously, this module introduces a service life rollback factor to quantify the impact of maintenance actions on the equipment failure rate. Based on maintenance records, it calculates the equivalent service life after each maintenance and redetermines the post-maintenance failure rate curve from the base failure rate function based on the equivalent service life, achieving dynamic and accurate prediction of the equipment failure rate.
[0071] The health status assessment module, connected to the dynamic failure rate prediction module, is used to assess the health status of equipment based on real-time failure rates. The health status assessment module incorporates a health index calculation model to calculate the equipment health index and provides graded early warnings based on the health index.
[0072] The optimization decision module is used to establish a total cost of ownership (TCO) minimization optimization model with the goal of improving power supply reliability, taking into account equipment failure rate, feeder automation coverage, and uninterrupted power supply (UPS) capability. The module uses the outputs of the dynamic failure rate prediction module and the health status assessment module as inputs, and the investment in equipment failure rate improvement, feeder automation, and UPS as decision variables, with investment budget and unidirectional evolution of equipment status as constraints, to solve for the optimal investment allocation scheme.
[0073] The scheduling strategy generation module is used to transform the investment allocation scheme output by the optimization decision module into specific executable scheduling instructions, generate scheduling strategies that include equipment maintenance plans, feeder automation deployment schemes, and live-line work implementation plans, and verify the investment budget of the generated scheduling strategies.
[0074] Through the coordinated efforts of the aforementioned modules, the entire process of optimized scheduling for distribution network reliability can be automated. Taking a pilot application by a municipal power supply company as an example, after the system was deployed, the accuracy of equipment fault prediction increased by 35%, the number of unplanned power outages decreased by 31%, the regional SAIDI was reduced by 28% within budget, and the incremental investment payback period was approximately 5 years, significantly improving the reliability of distribution network power supply and the economic efficiency of operation and maintenance.
[0075] The technical scope of this invention is not limited to the content described above. Those skilled in the art can make various modifications and variations to the above embodiments without departing from the technical concept of this invention, and all such modifications and variations should fall within the protection scope of this invention.
Claims
1. A power distribution network reliability optimization scheduling method based on fault monitoring, characterized in that, Including the following steps: Step 1: Multi-source monitoring data acquisition and preprocessing, collecting multi-source monitoring data of power distribution network equipment to form an equipment status characteristic database; Step 2: Establish a dynamic model of equipment failure rate based on Weibull distribution, and use Marquardt method to determine the boundary point and corresponding parameters between the random failure period and the wear failure period of the equipment; By integrating real-time monitoring data to dynamically correct the basic failure rate, and introducing a service life rollback factor to quantify the impact of maintenance on the equivalent service life, dynamic prediction of equipment failure rate is achieved. Step 3: Based on the predicted dynamic failure rate, combined with the equipment's operating years, maintenance history, and environmental factors, construct an equipment health index, and classify and issue early warnings for the equipment according to the health index; Step 4: Construct an optimization decision model considering the economics of operation and maintenance. With the goal of improving power supply reliability, a total cost of ownership (TCO) minimization optimization model is established by comprehensively considering equipment failure rate, feeder automation coverage, and uninterrupted power supply (UPS) capability. The optimal investment allocation scheme is solved by taking the output dynamic failure rate and health index as inputs, the investment in equipment failure rate improvement, feeder automation, and UPS as decision variables, and the investment budget and unidirectional evolution of equipment status as constraints. Step 5: Generate the optimal scheduling strategy. Based on the final investment allocation plan, convert it into executable scheduling instructions and generate a scheduling strategy that includes maintenance plans, feeder automation deployment plans, and live-line work implementation plans.
2. The scheduling method of claim 1, wherein, The Marquardt method is used to iteratively solve for the boundary points and corresponding parameters of the two stages. The specific steps are as follows: Collect n sets of failure rate data , arrange in time; get initial estimates of Weibull parameters and by solving two selected data points, as the initial values for the first iteration; Calculate the residual sum of squares for each data point under the current parameters; construct the damping matrix using partial derivatives, update the parameter estimates, and iterate until the residual sum of squares converges; At the same time, for each possible break point k, calculate the following: where is the average failure rate of the random failure rate, is the actual failure rate in the ti year, is the Weibull fitting value of the wear-out failure period; select k that makes Sk minimum as the optimal break point.
3. The scheduling method of claim 1 or 2, characterized in that: Step 2, which involves dynamically correcting the base failure rate by integrating real-time monitoring data, specifically includes: Electrical quantity correction factor, mechanical quantity correction factor, environmental quantity correction factor and state quantity correction factor are constructed. The comprehensive correction factor is obtained by multiplying the correction factors. The real-time failure rate is obtained by multiplying the basic failure rate by the comprehensive correction factor.
4. The scheduling method of claim 3, wherein, The introduction of a service life rollback factor to quantify the impact of maintenance specifically includes: the equivalent service life of the equipment after maintenance is determined by the product of the service life before maintenance and the service life rollback factor, wherein the service life rollback factor decreases exponentially with the number of maintenances; the failure rate after maintenance is obtained by substituting the sum of the equivalent service life and the number of years after maintenance into the basic failure rate function.
5. The scheduling method according to claim 1, characterized in that, The objective function of the total cost of ownership minimization optimization model is: , wherein C is the total investment cost, including the equipment failure rate improvement investment C1, the feeder automation investment C2, and the non-outage operation investment C3; is the discount rate considering the interest rate and inflation; n is the system life cycle; b is the ratio of the unit outage power price to the average power price; d is the average power price; and EENS is the expected power supply shortage.
6. The scheduling method of claim 5, wherein, The investment to improve the equipment failure rate consists of the sum of operating costs, maintenance costs, and failure costs. Operating costs are the sum of basic operating expenses over the years of service. Maintenance costs are the sum of minor and major repair costs and are affected by the degree of maintenance. Failure costs are the product of the expected number of failures and the cost of maintenance for each failure.
7. The scheduling method according to claim 5, characterized in that, The expected power shortage amount is: wherein, is a typical power supply area number, N is a number of typical power supply areas, h is a feeder number, is a number of feeder lines in the th area, is a number of feeder lines in the th area, is a unit user load capacity of the th area, is a system average outage duration of the hth feeder line of the th area, and is a number of users corresponding to the feeder line.
8. The scheduling method according to claim 1, characterized in that, The equipment health index is defined as the degree of deviation between the current operating state of the equipment and the health baseline state. It is determined by the ratio of the real-time failure rate to the maximum allowable failure rate. The higher the index value, the better the equipment condition.
9. The scheduling method according to claim 1, characterized in that, The constraints include: Investment budget constraints, namely, the total investment in improving equipment failure rate, investment in feeder automation and investment in uninterrupted power supply work shall not exceed the upper limit of the annual budget, and each investment shall be controlled within its respective sub-budget. The automation state constraint for each feeder is a binary variable. The uninterrupted power supply operation status constraint means that the uninterrupted power supply operation status of each feeder is a binary variable; Equipment failure rate constraints are implemented by scheduling maintenance operations to ensure that the real-time failure rate of all equipment is always kept below the maximum allowable value.
10. A distribution network reliability optimization dispatching system based on fault monitoring, characterized in that, include: The system includes a memory and a processor, wherein the memory stores a computer program that, when executed by the processor, implements the fault monitoring-based distribution network reliability optimization scheduling method as described in any one of claims 1 to 9.