Diesel generating set dynamic load cooperative control method based on containerized micro-service

Abstract

The invention provides a diesel generating set dynamic load cooperative control method based on containerized micro-service, and relates to the field of generating set electrical control. When first real-time set fault information indicates a fault state, a first state monitoring container triggers automatic shutdown protection mapping to isolate a first diesel generating set; and after a real-time load deviation value is calculated according to the N-1 local real-time loads and the real-time power supply demand total load, N-1 real-time load compensation gap values are output according to the power supply load task allocation and the N-1 unit performance characteristic compensation deviation values, and load cooperative compensation of the N-1 diesel generating sets is carried out. The technical problems that in the prior art, a diesel generating set is poor in load control stability and reliability, and unit fault recognition isolation response is delayed are solved. The technical effects of intelligent dynamic distribution of load tasks and millisecond-level fault isolation on the basis of quickly and accurately identifying unit faults and guaranteeing continuous stability of power supply are achieved.

Description

Technical Field

[0001] This invention relates to the field of electrical control of generator sets, and in particular to a dynamic load collaborative control method for diesel generator sets based on containerized microservices. Background Technology

[0002] In parallel operation systems of diesel generator sets, the stability and reliability of load control have always been key factors restricting system performance. Traditional control methods often struggle to achieve precise power distribution among multiple units, especially under conditions of sudden load changes or unit failures, which can easily lead to system voltage and frequency fluctuations exceeding allowable ranges, affecting power supply quality.

[0003] On the other hand, due to the lack of a distributed real-time monitoring mechanism, fault detection often relies on periodic polling or threshold alarms, which typically requires a long processing time from fault occurrence to final isolation. This delay not only increases the risk of fault propagation but may also lead to a chain reaction in the system, severely affecting power supply continuity.

[0004] In summary, existing diesel generator sets suffer from weak load control stability and reliability, as well as slow response to fault identification and isolation.

[0005] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0006] This invention provides a dynamic load collaborative control method for diesel generator sets based on containerized microservices, which addresses the technical problems of weak load control stability and reliability of existing diesel generator sets, as well as the sluggish response to generator set fault identification and isolation.

[0007] In view of the above problems, the present invention provides a dynamic load collaborative control method for diesel generator sets based on containerized microservices, the method comprising:

[0008] N status monitoring containers are configured for N diesel generator sets in a parallel-operated motor system to complete the local deployment of a containerized microservice architecture. The N status monitoring containers perform multi-threaded fault monitoring on the N diesel generator sets, outputting N real-time unit fault information. When the first real-time unit fault information indicates a fault state, the first status monitoring container triggers automatic shutdown protection mapping to isolate the first diesel generator set and generates a first fault shutdown warning, which is broadcast to the remaining N-1 status monitoring containers. After receiving the first fault shutdown warning, the N-1 status monitoring containers calculate and output the power supply load task allocation based on short-term power supply load characteristics. After interactively obtaining the total real-time power supply demand load, they calculate the real-time load deviation based on the N-1 local real-time loads and the total real-time power supply demand load. They compensate for the real-time load deviation based on the power supply load task allocation and the performance characteristics of the N-1 units, solving and outputting N-1 real-time load compensation gap values. Based on the N-1 real-time load compensation gap values, N-1 electrical commands are generated and sent to the PLCs of the N-1 diesel generator sets via the CAN bus to perform load collaborative compensation.

[0009] In one implementation, the N status monitoring containers perform multi-threaded fault monitoring on the N diesel generator sets, output N real-time generator set fault information, and also perform the following processing:

[0010] M key monitoring items of the first diesel generator set are predefined; M associated sampling indicators, M associated safety thresholds, and M associated sampling frequencies are configured for the M key monitoring items; M physical channels of M monitoring task sub-threads and M generator set monitoring sensors in the first state monitoring container are constructed based on the M associated sampling indicators; the M associated safety thresholds and M associated sampling frequencies are set as multi-threaded operation constraints for the M monitoring task sub-threads; in the first state monitoring container, the M monitoring task sub-threads receive M real-time associated sampling data transmitted back from the M physical channels at the M associated sampling frequencies, and output the first real-time generator set fault information by mapping and comparing the M real-time associated sampling data with the M associated safety thresholds.

[0011] In one implementation, based on the power supply load task allocation and the performance characteristics of N-1 generating units, the real-time load deviation is compensated, and N-1 real-time load compensation gap values ​​are calculated and output. The following processing is also performed:

[0012] Starting from the timestamp of the real-time total power demand load, historical demand load is backtracked to obtain the total power demand load sequence; a sliding window linear regression is performed on the total power demand load sequence to calculate the short-term load fluctuation threshold; N-1 unit load intervals are obtained by looking up a table based on the performance characteristics of the N-1 units; the real-time load deviation is expanded using the short-term load fluctuation threshold to output the extreme load deviation; the extreme load deviation is compensated according to the power supply load task allocation to output N-1 extreme load compensation gap values; the capacity margin is verified by traversing the N-1 extreme load compensation gap values ​​using the N-1 unit load intervals to output the N-1 real-time load compensation gap values.

[0013] In one implementation, the following processing is also performed:

[0014] If any indicator data in the M real-time associated sampling data deviates from the M associated safety thresholds, then the first real-time fault monitoring item and the first real-time deviation rate are extracted; the first warning level is matched according to the first real-time deviation rate; the first real-time fault monitoring item and the first warning level are packaged and the first fault shutdown warning is output; when the first status monitoring container broadcasts the first fault shutdown warning to the N-1 status monitoring containers through the communication microservice, it simultaneously sends the first fault shutdown warning to the operation and maintenance platform.

[0015] In one implementation, the following processing is also performed:

[0016] If the first real-time deviation rate is greater than the preset risk transmission scale, then after receiving the first fault shutdown warning, the N-1 status monitoring containers will backtrack the N-1 associated sampling sequences within the preset intervention time window based on the first real-time fault monitoring item. After calculating the N-1 fluctuation time series rates of the N-1 associated sampling sequences, the N-1 status monitoring containers will select O diesel generator sets from the N-1 diesel generator sets for collaborative isolation based on the abnormal discrete characteristics of the N-1 fluctuation time series rates, where O≤max(1,N / 5).

[0017] In one implementation, after receiving the first fault shutdown warning, the N-1 status monitoring containers calculate and output power load task allocation based on short-term power load characteristics, and also perform the following processing:

[0018] A short-term supply time window is preset; after receiving the first fault shutdown warning, the N-1 status monitoring containers use the short-term supply time window as a backtracking time constraint to locally backtrack and call N-1 power supply load sequences; and output the power supply load task allocation based on the load fluctuation characteristics of the N-1 power supply load sequences.

[0019] In one implementation, the power supply load task allocation is output based on the load fluctuation characteristics of the N-1 power supply load sequences, and the following processing is also performed:

[0020] The second power supply load sequence is decomposed into time series, and a second trend component, a second periodic component, and a second random component are output. Feature extraction is performed on the second trend component, the second periodic component, and the second random component to output a second trend growth rate, a second periodic amplitude, and a second noise intensity. Fluctuation quantization indicators are obtained by weighted normalization of the second trend growth rate, the second periodic amplitude, and the second noise intensity, and a second load reliability component is output. Similarly, N-1 load reliability components are quantized and calculated based on the N-1 power supply load sequences. The power supply load task allocation is calculated and output based on the N-1 load reliability components.

[0021] The technical solution provided in this invention has at least the following technical effects or advantages:

[0022] The method provided in this invention configures N status monitoring containers for N diesel generator sets in a parallel-operated motor system to complete the local deployment of a containerized microservice architecture. The N status monitoring containers perform multi-threaded fault monitoring on the N diesel generator sets, outputting N real-time unit fault information. When the first real-time unit fault information indicates a fault state, the first status monitoring container triggers automatic shutdown protection mapping to isolate the first diesel generator set and generates a first fault shutdown warning, broadcasting it to the remaining N-1 status monitoring containers. After receiving the first fault shutdown warning, the N-1 status monitoring containers calculate and output power supply load task allocation based on short-term power supply load characteristics. After interactively obtaining the total real-time power supply demand load, they calculate the real-time load deviation based on the N-1 local real-time loads and the total real-time power supply demand load. They compensate for the real-time load deviation based on the power supply load task allocation and the performance characteristics of the N-1 units, solving and outputting N-1 real-time load compensation gap values. Based on the N-1 real-time load compensation gap values, they generate N-1 electrical commands and send them to the PLCs of the N-1 diesel generator sets via the CAN bus to perform load collaborative compensation. It achieves intelligent dynamic allocation of load tasks and millisecond-level fault isolation based on rapid and accurate identification of unit faults, thereby improving the operational reliability of parallel-operated motor systems while ensuring continuous and stable power supply. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This invention illustrates a flowchart of a dynamic load collaborative control method for diesel generator sets based on containerized microservices provided by the present invention.

[0025] Figure 2 The diagram illustrates the process of monitoring and outputting real-time generator set fault information in the dynamic load collaborative control method for diesel generator sets based on containerized microservices provided by the present invention. Detailed Implementation

[0026] This invention provides a dynamic load collaborative control method for diesel generator sets based on containerized microservices, which addresses the technical problems of weak load control stability and reliability of existing diesel generator sets, as well as the sluggish response to generator set fault identification and isolation.

[0027] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings, but it should be understood that the scope of protection of the present invention is not limited to the specific embodiments.

[0028] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0029] Unless otherwise expressly stated, throughout the specification and claims, the term "comprising" or its variations such as "including" or "comprises" shall be understood to include the stated elements or components without excluding other elements or other components.

[0030] Example: A flowchart of the dynamic load collaborative control method for diesel generator sets based on containerized microservices provided in this embodiment of the invention is shown below. Figure 1 The method includes:

[0031] Step A100: Configure N status monitoring containers for N diesel generator sets in the parallel-running motor system to complete the local deployment of the containerized microservice architecture.

[0032] Specifically, this embodiment constructs a distributed control system architecture based on containerized microservices. By configuring N independent status monitoring containers for each of the N diesel generator sets in the parallel-running motor system, the modular deployment of control functions is achieved. Each container encapsulates complete monitoring and control microservice components through lightweight virtualization technology and completes local deployment on the edge computing node close to the generator set.

[0033] The containerized microservice architecture in this embodiment not only ensures the physical isolation of each unit control unit, but also provides underlying support for subsequent multi-threaded fault monitoring (step A200) and rapid isolation of faulty units (step A300). At the same time, the service mesh communication mechanism between containers establishes a high-speed channel for real-time data interaction between status monitoring containers (such as the fault warning broadcast in step A300).

[0034] Step A200: The N status monitoring containers perform multi-threaded fault monitoring on the N diesel generator sets and output N real-time generator set fault information.

[0035] In one implementation, see Figure 2 The N status monitoring containers perform multi-threaded fault monitoring on the N diesel generator sets and output N real-time generator set fault information. The method step A200 provided by this invention includes:

[0036] Step A210: Predefine M key monitoring items for the first diesel generator set.

[0037] Step A220: Configure M associated sampling indicators, M associated safety thresholds, and M associated sampling frequencies for the M key monitoring items.

[0038] Step A230: Construct M physical channels for the M monitoring task sub-threads and M unit monitoring sensors in the first state monitoring container based on the M associated sampling indicators.

[0039] Step A240: Set the M associated security thresholds and M associated sampling frequencies as multi-threaded operation constraints for the M monitoring task sub-threads.

[0040] Step A250: In the first status monitoring container, the M monitoring task sub-threads receive M real-time associated sampling data transmitted back from the M physical channels at the M associated sampling frequencies, and output the first real-time unit fault information by mapping and comparing the M real-time associated sampling data with the M associated safety thresholds.

[0041] It should be understood that in this embodiment, each status monitoring container continuously collects and analyzes the operating status data of the corresponding generator set through a multi-threaded monitoring mechanism that operates in parallel, and generates a set of fault information reflecting the health status of the unit in real time, providing timely status awareness for the parallel-operating motor system.

[0042] This embodiment takes the first state monitoring container for monitoring the operating status faults of the first diesel generator set as an example to illustrate the technical solution in detail.

[0043] Specifically, this embodiment predefines M key monitoring items that the first diesel generator set needs to monitor. The M key monitoring items cover the core indicators that affect the safe operation of the first diesel generator set, such as temperature, pressure, and vibration.

[0044] For the M key monitoring items, M associated sampling indicators, M associated safety thresholds, and M associated sampling frequencies are specifically configured. The associated safety thresholds are the safety limits of the corresponding indicator parameters during safe operation. The associated sampling indicators, associated safety thresholds, and associated sampling frequencies form a complete monitoring specification to ensure the accuracy and timeliness of data collection.

[0045] Based on this, this embodiment determines that the first diesel generator set can collect the M associated sampling indicators by pre-deploying M monitoring sensor devices, and then constructs M monitoring task sub-threads and M physical channels of the M generator set monitoring sensors in the first state monitoring container. The M monitoring task sub-threads respectively make a one-dimensional judgment on whether the diesel generator is in a fault state based on the data returned by a monitoring sensor device through the physical channel.

[0046] A priority scheduling strategy is adopted to allocate CPU time slices that meet the requirements of the M associated sampling frequencies to the M monitoring task sub-threads, and a memory lock mechanism is used to ensure atomic access to key threshold parameters. The M associated safety thresholds are loaded as fault risk judgment conditions into the M monitoring task sub-threads to complete the precise configuration of multi-threaded operation constraints.

[0047] Based on this, the first state monitoring container will dynamically adjust thread priorities during operation, giving higher scheduling weights to monitoring tasks with high sampling frequency (monitoring task sub-threads), thereby prioritizing the timely collection of key parameters during resource contention.

[0048] This embodiment implements strict and reliable constraint management of the M monitoring task sub-threads based on the M associated safety thresholds and M associated sampling frequencies, providing runtime assurance for subsequent real-time fault diagnosis.

[0049] In the first state monitoring container, the complete workflow of the M monitoring task sub-threads is as follows:

[0050] The M monitoring task sub-threads receive M real-time associated sampling data transmitted back from the M physical channels at the M associated sampling frequencies, and output the first real-time unit fault information by mapping and comparing the M real-time associated sampling data with the M associated safety thresholds.

[0051] Similarly, the N status monitoring containers perform multi-threaded fault monitoring on the N diesel generator sets and output N real-time generator set fault information.

[0052] This embodiment configures N status monitoring containers for N diesel generators in a parallel-operated motor system. Each status monitoring container continuously collects and analyzes the operating status data of the corresponding generator set through a multi-threaded monitoring mechanism that operates in parallel. This achieves the technical effect of generating a set of fault information reflecting the health status of the generator set in real time, providing timely status perception capability for the parallel-operated motor system.

[0053] Step A300: When the fault information of the first real-time unit is indicative of a fault state, the first status monitoring container triggers the automatic shutdown protection mapping to isolate the first diesel generator set and generates a first fault shutdown warning broadcast to the remaining N-1 status monitoring containers.

[0054] This embodiment further elaborates on the technical solution by taking the example of a first diesel generator set being detected and identified as having a fault by a first status monitoring container.

[0055] Specifically, when the first status monitoring container detects a fault in the first diesel generator set, it will immediately execute a preset automatic protection program:

[0056] First, a shutdown command is sent to the first diesel generator set to safely remove it from operation. Simultaneously, a warning message containing fault details is broadcast in real-time via the inter-container communication network to all other normally operating monitoring containers in the parallel-running generator system (the remaining N-1 status monitoring containers). This process achieves rapid isolation of the faulty unit and synchronous updates to the generator status in the parallel-running generator system, ensuring that the remaining units can promptly obtain fault information and initiate the corresponding load redistribution mechanism, thereby maintaining the stable operation of the entire parallel power generation system. The entire fault handling process, from detection and isolation to warning broadcasting, is completed within milliseconds, ensuring timely protection of the faulty equipment and providing the necessary information foundation for subsequent coordinated control.

[0057] Step A400: After receiving the first fault shutdown warning, the N-1 status monitoring containers calculate and output the power supply load task allocation based on the short-term power supply load characteristics.

[0058] In one implementation, after receiving the first fault shutdown warning, the N-1 status monitoring containers calculate and output power load task allocation based on short-term power load characteristics. Step A400 of the method provided by this invention includes:

[0059] Step A410: Preset short-term supply time window.

[0060] Step A420: After receiving the first fault shutdown warning, the N-1 status monitoring containers use the short-term supply time window as the backtracking time constraint to locally backtrack and call the N-1 power supply load sequences.

[0061] Step A430: Analyze the load fluctuation characteristics of the N-1 power supply load sequences and output the power supply load task allocation.

[0062] In one implementation, the power supply load task allocation is output based on the load fluctuation characteristics of the N-1 power supply load sequences, and the provided method step A430 includes:

[0063] Step A431: Decompose the second power supply load sequence into a time series and output the second trend component, the second periodic component, and the second random component.

[0064] Step A432: Extract features from the second trend component, the second periodic component, and the second random component, and output the second trend growth rate, the second periodic amplitude, and the second noise intensity.

[0065] Step A433: Perform fluctuation quantification by weighted normalization of the second trend growth rate, the second period amplitude, and the second noise intensity, and output the second load reliability component.

[0066] Step A434: Similarly, calculate N-1 load reliability components based on the N-1 power supply load sequences.

[0067] Step A435: Calculate and output the power supply load task allocation based on the N-1 load reliability components.

[0068] Specifically, this embodiment pre-sets a fixed short-term supply time window range as the analysis benchmark for the short-term power supply load characteristics of the unit. The duration of this time window is determined comprehensively based on the response speed of the parallel-operated motor system and the load fluctuation cycle characteristics, providing a clear time range constraint for subsequent load sequence backtracking analysis and ensuring the timeliness and consistency of load feature extraction.

[0069] After receiving the first fault shutdown warning signal, the N-1 status monitoring containers that are operating normally use the short-term supply time window as the backtracking time constraint to locally backtrack and call the power supply load records within the preset time window range to obtain the N-1 power supply load sequences.

[0070] In this embodiment, after calculating and outputting the N-1 load reliability components based on the N-1 power supply load sequences, the power supply load task allocation is calculated and output based on the N-1 load reliability components.

[0071] Since the methods for calculating and outputting the N-1 load reliability components based on the N-1 power supply load sequences are consistent, this embodiment takes the calculation and output of the second load reliability component corresponding to the second power supply load sequence as an example to elaborate on the technical solution.

[0072] The second power supply load sequence is decomposed into three components with clear physical meaning using time series analysis. The second trend component reflects the long-term trend of the load, the second periodic component reflects the regular fluctuation characteristics of the load, and the second random component characterizes the unit operating noise and sudden disturbances. This decomposition method lays the foundation for subsequent refined analysis.

[0073] Key feature parameters are extracted from each component obtained from the decomposition. Specifically, the rate of change feature is extracted from the second trend component to obtain the second trend growth rate, the fluctuation amplitude feature is extracted from the second periodic component to obtain the second periodic amplitude, and the intensity level feature is extracted from the second random component to output the second noise intensity.

[0074] It should be understood that the second trend growth rate, the second period amplitude, and the second noise intensity accurately characterize the dynamic characteristics of the second power supply load sequence, providing an objective basis for the subsequent reliability assessment of the first diesel generator set's load transfer.

[0075] The weighting rules are preset. In this embodiment, the weighting rules are not limited. They can be set in practice according to the actual operating capacity of the unit equipment. The second trend growth rate, the second period amplitude, and the second noise intensity are weighted and normalized by the preset weighting rules to quantify the fluctuation index. The second load reliability component, which characterizes the power supply stability and regulation capability of the second diesel generator set, is output.

[0076] Similarly, N-1 load reliability components are quantized and calculated based on the N-1 power supply load sequences, and the data ratio of the N-1 load reliability components is used as the power supply load task allocation output.

[0077] This embodiment uses time-series decomposition and feature extraction techniques to accurately quantify the load reliability level of each generator set in normal operation, and dynamically adjusts the load allocation weight accordingly. This ensures the power supply continuity of the parallel-operating motor system in case of failure, optimizes the overall operating efficiency, and significantly improves the response speed of load redistribution through localized data processing and parallel computing. It provides intelligent decision support for the stable operation of the parallel power generation system under abnormal conditions.

[0078] Step A500: After interactively obtaining the total real-time power demand load, calculate the real-time load deviation based on N-1 local real-time loads and the total real-time power demand load.

[0079] Specifically, in this embodiment, the N-1 local real-time loads that report the actual local load status of the N-1 diesel generator sets that are operating normally are used to obtain the real-time load deviation amount by subtracting the N-1 local real-time loads from the total real-time power supply demand load.

[0080] Step A600: Based on the power supply load task allocation and the performance characteristics of N-1 units, compensate for the real-time load deviation and solve for and output N-1 real-time load compensation gap values.

[0081] In one implementation, the real-time load deviation is compensated based on the power supply load task allocation and the performance characteristics of N-1 generating units, and N-1 real-time load compensation gap values ​​are calculated and output. Step A600 of the method provided by this invention includes:

[0082] Step A610: Starting from the timestamp of the real-time total power demand load, perform historical demand load backtracking to obtain the total power demand load sequence.

[0083] Step A620: Perform a sliding window linear regression on the total power demand load sequence to calculate the short-term load fluctuation threshold.

[0084] Step A630: Obtain the load range of N-1 units by looking up the table based on the performance characteristics of the N-1 units.

[0085] Step A640: Extend the real-time load deviation using the short-term load fluctuation threshold, and output the ultimate load deviation.

[0086] Step A650: Based on the power supply load task, allocate compensation for the extreme load deviation and output N-1 extreme load compensation gap values.

[0087] Step A660: Use the N-1 unit load ranges to traverse the N-1 extreme load compensation gap values ​​to verify the capacity margin, and output the N-1 real-time load compensation gap values.

[0088] Specifically, this embodiment uses the data collection timestamp of the total real-time power demand load as the starting point, and backtracks to extract historical load data over a certain period of time to obtain the total power demand load sequence. This historical load time series data, the total power demand load sequence, contains trend information on changes in the load demand of electricity users, providing sufficient data support for subsequent fluctuation analysis.

[0089] A sliding window technique is used to analyze the total power demand load sequence in real time. Short-term trend characteristics of load changes are extracted using linear regression. Based on these short-term trend characteristics, a reasonable threshold range for load fluctuations is calculated as the short-term load fluctuation threshold. This threshold reflects the maximum possible load change in the power demand of the electricity user within a short period. This short-term load fluctuation threshold provides an important reference for subsequent load compensation decisions, ensuring that the parallel-operated motor system can cope with load fluctuations under worst-case conditions.

[0090] A pre-built unit performance characteristic database is constructed, which records N sets of load capacity limitations of N diesel generator sets in the parallel-operated motor system under N sample operating performance states.

[0091] In this embodiment, the N-1 unit performance characteristics (real-time operating performance status) are used as search conditions. The optimal load range of the unit under the corresponding operating performance status is obtained by looking up the table in the unit performance characteristic database, and the N-1 unit load ranges are obtained. The N-1 unit load ranges are used to judge the rationality of the additional load allocation of the unit.

[0092] The real-time load deviation is superimposed with the calculated short-term fluctuation threshold to obtain a more conservative ultimate load deviation. This extended ultimate load deviation represents the maximum load shortfall that the parallel-operated motor system may need to cope with in the near future. Compensation allocation based on this value can ensure that the system has sufficient coping capacity and can maintain stable operation even under the most unfavorable conditions.

[0093] The extreme load deviation is weighted and decomposed into N-1 extreme load compensation gap values ​​according to the weight allocation of the power supply load task. The extreme load compensation gap value is the additional load of the diesel generator set under normal operation under the condition of stable load supply.

[0094] To ensure the feasibility of the compensation scheme, the calculated limit compensation gap value is finally verified. Specifically, the capacity margin is verified by traversing the N-1 limit load compensation gap values ​​through the N-1 unit load ranges to confirm that all compensation commands are within the safe operating range of the units.

[0095] The compensation amount exceeding the unit's capacity is dynamically adjusted and redistributed to other units with sufficient capacity. After rigorous verification in this round, the final output of the N-1 real-time load compensation gap values ​​not only meets the external power supply stability requirements of the parallel-operated motor system, but also ensures the safe operation of each unit.

[0096] Step A700: Generate N-1 electrical commands based on the N-1 real-time load compensation gap values, and send them to the N-1 diesel generator set PLCs via the CAN bus to perform load collaborative compensation.

[0097] Specifically, in this embodiment, the calculated N-1 real-time load compensation gap values ​​are converted into N-1 directly executable electrical instructions (electrical control instructions), and these N-1 electrical instructions are sent to the PLC controllers of N-1 diesel generator sets in real time through a highly reliable CAN bus communication network.

[0098] The electrical commands, containing precise voltage regulation setpoints and power output adjustments, are received and executed by the control modules of N-1 diesel generator sets, enabling load coordination and compensation among multiple units. This process ensures that the parallel-operated generator system can quickly and accurately redistribute the power supply load in the event of a fault, maintaining the operational stability of the parallel-operated generator system. Simultaneously, it fully leverages the performance advantages of each unit, completing a fully closed-loop control from fault detection to load rebalancing within milliseconds, providing intelligent dynamic adjustment capabilities for the parallel power generation system.

[0099] This embodiment achieves intelligent dynamic allocation of load tasks and millisecond-level fault isolation based on rapid and accurate identification of unit faults by constructing a containerized microservice architecture. This improves the operational reliability of parallel-running motor systems while ensuring continuous and stable power supply.

[0100] In one implementation, the method further includes:

[0101] Step A3001: If any indicator data in the M real-time associated sampling data deviates from the M associated safety thresholds, then extract the first real-time fault monitoring item and the first real-time deviation rate.

[0102] Step A3002: Match the first warning level according to the first real-time deviation rate.

[0103] Step A3003: Package the first real-time fault monitoring item and the first early warning level, and output the first fault shutdown early warning.

[0104] Step A3004: When the first status monitoring container broadcasts the first fault shutdown warning to the N-1 status monitoring containers via the communication microservice, it simultaneously sends the first fault shutdown warning to the operation and maintenance platform.

[0105] In this embodiment, the specific operation process of the first status monitoring container is as follows:

[0106] If any indicator data in the M real-time associated sampling data deviates (exceeds) the M associated safety thresholds, then the specific abnormal monitoring indicator is locked and the first real-time fault monitoring item and the first real-time deviation rate corresponding to the abnormal indicator are extracted.

[0107] The first real-time deviation rate is used to match the first early warning level in the hierarchical early warning rule base. The first real-time fault monitoring item and the first early warning level are standardized and encapsulated to output the first fault shutdown early warning. The first fault shutdown early warning includes the fault timestamp, unit number, fault monitoring item and early warning level.

[0108] When the first status monitoring container broadcasts the first fault shutdown warning to the N-1 status monitoring containers via a communication microservice, it simultaneously sends the first fault shutdown warning to the remote operation and maintenance platform, achieving dual protection of automated on-site handling and manual back-end monitoring. This two-way communication design ensures both rapid response of the parallel-operated motor system and timely fault notifications for operation and maintenance personnel.

[0109] In one implementation, the method further includes:

[0110] Step A3005: If the first real-time deviation rate is greater than the preset risk transmission scale, then after receiving the first fault shutdown warning, the N-1 status monitoring containers will backtrack the N-1 associated sampling sequences of the preset intervention time window locally according to the first real-time fault monitoring item.

[0111] Step A3006: After the N-1 state monitoring containers calculate the N-1 fluctuation time series rates of the N-1 associated sampling sequences, based on the abnormal discrete characteristics of the N-1 fluctuation time series rates, O diesel generator sets are selected from the N-1 diesel generator sets for collaborative isolation, where O≤max(1,N / 5).

[0112] Specifically, if the first real-time deviation rate is greater than the preset risk transmission scale, the first real-time fault monitoring item may have the same fault propagation risk in the remaining N-1 diesel generator sets. Based on this, in this embodiment, after receiving the first fault shutdown warning, the N-1 status monitoring containers trace back N-1 associated sampling sequences within the preset intervention time window according to the first real-time fault monitoring item.

[0113] After the N-1 status monitoring containers calculate the N-1 fluctuation time series rates of the N-1 associated sampling sequences, based on the abnormal discrete characteristics of the N-1 fluctuation time series rates, they identify and screen the N-1 diesel generator sets and perform collaborative isolation on the O diesel generator sets that have large deviations from other generator sets, where O≤max(1,N / 5), to ensure that the minimum operating capacity of the parallel-operating motor system is maintained while controlling the spread of risks.

[0114] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A dynamic load collaborative control method for diesel generator sets based on containerized microservices, characterized in that, include: Configure N status monitoring containers for N diesel generator sets in a parallel-operating motor system to complete the local deployment of a containerized microservice architecture; The N status monitoring containers perform multi-threaded fault monitoring on the N diesel generator sets and output N real-time unit fault information; When the fault information of the first real-time unit is indicative of a fault state, the first status monitoring container triggers the automatic shutdown protection mapping to isolate the first diesel generator set and generates a first fault shutdown warning broadcast to the remaining N-1 status monitoring containers. After receiving the first fault shutdown warning, the N-1 status monitoring containers calculate and output the power supply load task allocation based on the short-term power supply load characteristics. After obtaining the total real-time power demand load interactively, the real-time load deviation is calculated based on N-1 local real-time loads and the total real-time power demand load. Based on the power supply load task allocation and the performance characteristics of N-1 generating units, the real-time load deviation is compensated, and N-1 real-time load compensation gap values ​​are output. Based on the N-1 real-time load compensation gap values, N-1 electrical commands are generated and sent to the PLCs of N-1 diesel generator sets via the CAN bus to perform load collaborative compensation.

2. The dynamic load collaborative control method for diesel generator sets based on containerized microservices as described in claim 1, characterized in that, The N status monitoring containers perform multi-threaded fault monitoring on the N diesel generator sets, outputting N real-time generator set fault information, including: Predefine M key monitoring items for the first diesel generator set; For the M key monitoring items, configure M associated sampling indicators, M associated safety thresholds, and M associated sampling frequencies; Based on the M associated sampling indicators, construct M monitoring task sub-threads and M physical channels of the M unit monitoring sensors in the first state monitoring container; The M associated security thresholds and M associated sampling frequencies are set as multi-threaded operation constraints for the M monitoring task sub-threads; In the first status monitoring container, the M monitoring task sub-threads receive M real-time associated sampling data transmitted back from the M physical channels at the M associated sampling frequencies, and output the first real-time unit fault information by mapping and comparing the M real-time associated sampling data with the M associated safety thresholds.

3. The dynamic load collaborative control method for diesel generator sets based on containerized microservices as described in claim 1, characterized in that, Based on the power supply load task allocation and the performance characteristics of N-1 generating units, the real-time load deviation is compensated, and N-1 real-time load compensation gap values ​​are calculated and output, including: Starting from the timestamp of the real-time total power demand load, historical demand load is traced back to obtain the total power demand load sequence. A sliding window linear regression is performed on the total power demand load sequence to calculate the short-term load fluctuation threshold; The load range of N-1 generating units is obtained by looking up the table based on the performance characteristics of the N-1 generating units; The real-time load deviation is extended using the short-term load fluctuation threshold to output the ultimate load deviation. Based on the power supply load task, the extreme load deviation is compensated, and N-1 extreme load compensation gap values ​​are output. The capacity margin is verified by traversing the N-1 limit load compensation gap values ​​through the N-1 unit load ranges, and the N-1 real-time load compensation gap values ​​are output.

4. The dynamic load collaborative control method for diesel generator sets based on containerized microservices as described in claim 2, characterized in that, Also includes: If any indicator data in the M real-time associated sampling data deviates from the M associated safety thresholds, then the first real-time fault monitoring item and the first real-time deviation rate are extracted. The first warning level is matched based on the first real-time deviation rate; Package the first real-time fault monitoring item and the first early warning level, and output the first fault shutdown early warning; When the first status monitoring container broadcasts the first fault downtime warning to the N-1 status monitoring containers via a communication microservice, it simultaneously sends the first fault downtime warning to the operation and maintenance platform.

5. The dynamic load collaborative control method for diesel generator sets based on containerized microservices as described in claim 4, characterized in that, Also includes: If the first real-time deviation rate is greater than the preset risk transmission scale, then after receiving the first fault shutdown warning, the N-1 status monitoring containers will backtrack the N-1 associated sampling sequences of the preset intervention time window locally according to the first real-time fault monitoring item. After the N-1 state monitoring containers calculate the N-1 fluctuation time series rates of the N-1 associated sampling sequences, based on the abnormal discrete characteristics of the N-1 fluctuation time series rates, they select O diesel generator sets from the N-1 diesel generator sets to perform collaborative isolation, where O≤max(1,N / 5).

6. The dynamic load collaborative control method for diesel generator sets based on containerized microservices as described in claim 1, characterized in that, After receiving the first fault shutdown warning, the N-1 status monitoring containers calculate and output power load task allocation based on short-term power load characteristics, including: Pre-set short-term supply window; After receiving the first fault shutdown warning, the N-1 status monitoring containers use the short-term supply time window as the backtracking time constraint to locally backtrack and call the N-1 power supply load sequences. The power supply load task allocation is output based on the load fluctuation characteristics analysis of the N-1 power supply load sequences.

7. The dynamic load collaborative control method for diesel generator sets based on containerized microservices as described in claim 6, characterized in that, The power supply load task allocation is output based on the load fluctuation characteristics analysis of the N-1 power supply load sequences, including: The second power supply load sequence is decomposed into time series, and the second trend component, the second periodic component, and the second random component are output. Feature extraction is performed on the second trend component, the second periodic component, and the second random component to output the second trend growth rate, the second periodic amplitude, and the second noise intensity. The second load reliability component is output by quantifying the fluctuation index by weighted normalization of the second trend growth rate, the second period amplitude, and the second noise intensity. Similarly, N-1 load reliability components are quantized and calculated based on the N-1 power supply load sequences; The power supply load task allocation is calculated and output based on the N-1 load reliability components.

Images